Williams System 3 - 7

From PinWiki
Jump to navigation Jump to search

Click to go back to the Williams solid state repair guides index.

1 Introduction

Williams entered the SS or Solid State era with a conversion of a 1976 Williams EM or Electro Mechanical pinball game called Grand Prix.

System 1 is considered to be the Solid State version of Grand Prix. It is thought that about 5 Grand Prix EM games were converted to SS prototypes using the new System 1 MPU Board and digital displays. The computer was only used for accumulating and displaying the player scores.

IPDB links: Grand Prix SS version          Grand Prix EM version

System 2 was next, with a 10 unit run of another 1976 EM game called Aztec. A working SS version is very rare. Both System 1 & 2 occurred quickly after each other at the end of 1976. Aztec SS was still a hybrid machine retaining the EM chime unit in the cabinet and a credit window with an EM numbered reel behind the backglass). Player scoring was still the only digital function at that time.

IPDB links: Aztec SS version          Aztec EM version

System 3 games were the first Williams SS production games, starting with Hot Tip in Nov. '77 and ending with Disco Fever in Aug '78. There were based on the Motorola 6800 8-bit CPU and using a Motorola 6820 PIA (Peripheral Interface Adapter) to handle the Display I/O from the MPU board. They also had three other 6820 PIAs on the Driver Board reading targets and other switches as inputs and controlling the insert Lamps and solenoids as outputs. Solenoid drives were mainly used for ball handling by firing coils, with a few triggering sound calls or a 'start of game' tune. This was the start of Williams using computers for game rules and settings. At this point they designed an 8x8 Switch Matrix, an 8x8 Lamp Matrix and solenoids driven by Darlington transistors. This design remained virtually unchanged to the end of Williams System 7.

During System 3 (Phoenix and Disco Fever) a memory protect circuit modification was added as to help protect CMOS RAM data during power up and power down of the game. DIP switches were used to program game settings (such as # of balls per game, high score replays).

System 4 games ran from Pokerino in Nov '78 through to Stellar Wars in Mar '79, according to IPDB.com. A notable game which outsold all of the other System 4 games combined with a production run of 19,505 was Flash (Steve Ritchie's first designed game for Williams). During System 4, Williams moved from using DIP switches to change game settings to having the game settings changed from the coin door switches. The settings were still stored in battery protected CMOS RAM. [ed Note: Citation needed] A coin door interlock switch enforced that CMOS memory could not be modified unless the coin door had been opened by the operator. Some of the game audits (coins accepted, total number of games played, etc.) still could not be changed without access to the MPU board behind the backglass.

System 6 games ran from Tri Zone in Jul '79 to Scorpion in Jul '80. Two notable games from this era were from the end of '79 and the beginning of '80 Gorgar and Firepower. Gorgar (14,000 produced) was the first talking pinball, and Firepower (17,410 produced) both talked and introduced the 'Lane Change' and 'Multiball (tm)' features to SS games. Note that there had been Multiball play available in EM games, it just wasn't called Multiball (tm) until Firepower. and this is a common misunderstanding. The features these games introduced became standards for almost all pinball games produced right up until today.

System 6a deserves to be mentioned here as it marked a transition to System 7. The game Alien Poker from Oct '80 used the Syatem 6a MPU board (which was not very different from System 6). But it supported 7 digit scoring displays and a redesigned Master Display Driver board, located behind the backglass on the back of the 'Lamp Board'. It also used a special 4 digit "credit/match" display in the approximate position where the System 6 Master Display Driver had been showing the same information (on a 6 digit display, with 2 of the digits unused). This new 7 digit scoring displays with a 4 digit credit/match display were then used in all the System 7 games (and System 9).

System 7 games ran from Black Knight in Nov '80 through to Star Light in Jun '84.

IPDB link: Complete System 7 Game list

Black Knight (13,075 produced) introduced a two level playing field and Magna-Save (tm) where the ball could be stopped from draining down the sides by pressing a cabinet button that activated an electromagnet. Star Light (100 produced) was a 'boutique' game by Williams' production standards as the focus was on ramping up production for the first System 9 game Space Shuttle (7,000 units). At least one Star Light game was made as the Prototype for System 9.

Note: Further discussion of changes and good pictures of the backbox boards for System 3-7 games can be found here: Tukkan.fliput.net

2 Games

2.1 System 3

Title Date of Release Model # Sound Other Boards Notes
Hot Tip 11-1977 477 Chimes
Lucky Seven 03-1978 480 Chimes
World Cup 05-1978 481 Type 1 Sound
Contact 05-1978 482 Type 1 Sound Widebody
Disco Fever 08-1978 483 Type 1 Sound Used banana flippers

2.2 System 4

Title Date of Release Model # Sound Other Boards Notes
Pokerino 10-1978 488 Type 1 Sound Widebody
Phoenix 11-1978 485 Type 1 Sound
Flash 01-1979 486 Type 1 Sound First game with continuous Background sound/Later games produced on System 6 platform
Stellar Wars 03-1979 490 Type 1 Sound Widebody/Later games produced on System 6 platform

2.3 System 6

Title Date of Release Model # Sound Other Boards Notes
Tri Zone 07-1979 487 Type 1 Sound
Time Warp 09-1979 489 Type 1 Sound Used banana flippers for most of run.
Gorgar 12-1979 496 Type 2 Sound w/ Speech First game with speech
Laser Ball 12-1979 493 Type 1 Sound Widebody
Firepower 02-1980 497 Type 2 Sound w/ Speech First SS Multiball (3 balls). First lane change.
Blackout 06-1980 495 Type 2 Sound w/ Speech First Williams game where a relay was used to turn general illumination on / off for effect.
Scorpion 07-1980 494 Type 1 Sound Widebody

2.4 System 6A

Title Date of Release Model # Sound Other Boards Notes
Algar 09-1980 499 Type 1 Sound 7 digit display board Widebody
Alien Poker 10-1980 501 Type 2 Sound w/ Speech 7 digit display board

2.5 System 7

Title Date of Release Model # Sound Other Boards Notes
Black Knight 11-1980 500 Type 2 Sound w/ Speech First game with magnasave
Jungle Lord 02-1981 503 Type 2 Sound w/ Speech
Pharaoh 05-1981 504 Type 2 Sound w/ Speech
Solar Fire 07-1981 507 Type 2 Sound
Barracora 09-1981 510 Type 2 Sound
Hyperball 12-1981 509 Type 2 Sound Uses different driver board than pinball games Not a pinball - a flipperless game firing tiny pinballs
Cosmic Gunfight 06-1982 502 Type 2 Sound
Varkon 09-1982 512 Type 2 Sound Pinball in stand-up Video cabinet
Warlok 10-1982 516 Type 2 Sound
Defender 12-1982 517 Type 2 Sound
Time Fantasy 03-1983 515 Type 2 Sound
Joust 04-1983 519 Type 2 Sound Two player head-to-head pinball
Firepower II 08-1983 521 Type 2 Sound Uses 50v Flipper Power Supply Board
Laser Cue 02-1984 520 Type 2 Sound
Star Light 06-1984 530 Type 2 Sound

Game date of release and model numbers provided by the Internet Pinball Database - http://www.ipdb.org

3 Documentation

3.1 Manuals & Schematics

Most manuals and ROM files can be found on IPDB.org for each specific game.

3.2 Parts Catalogs

Parts catalogs can also be useful, which include part numbers (helpful for purchasing parts online), exploded views of assemblies (helpful to see how the assemblies are put together), and board layouts & parts lists, and diagrams for controlled lamp, solenoid, and rubber locations.

Online copies of Bally & Williams parts catalogs can be found on Planetary Pinball.

Manufacturer Catalog Version Cover Source Games Notes
Williams 1980 Parts Catalog
Placeholder.jpg
Viewer Hot Tip, Lucky Seven, World Cup, Contact, Disco Fever, Phoenix, Flash, Tri-Zone, Pokerino, Time Warp, Stellar Wars, Laser Ball, Scorpion, Blackout, Gorgar, Firepower, Algar, Black Knight, Alien Poker, Cosmic Gunfight, Jungle Lord, Pharoah, Solar Fire, Hyperball, Barracora This catalog covers all system 3-6 games, and the first third of the system 7 games.


3.3 Service Bulletins

Most service bulletins can be found on IPDB.org for each specific game.

3.4 Instruction Cards

3.5 Fuses

3.5.1 Power Supply

Williams System 3 - 7 Power Supply Fuse Ratings
System F1 F2 F3 F4* F5 F6 F7
3 1/4amp SB 2 1/2amp SB 8amp FB 15amp FB 4amp SB NA NA
4 1/4amp SB 2 1/2amp SB 8amp FB 10amp FB 4amp FB NA NA
6 1/4amp SB 2 1/2amp SB 8amp FB 10amp FB 4amp FB NA NA
7 1/4amp SB 2 1/2amp SB 8amp FB 10amp FB 7amp SB 7amp SB NA


Note: The value of F4 does vary. Check the game manual for the correct rating. It's usually 10A for games with 2 flippers. Later System 7 games with a separate flipper power supply do not use F4.

3.6 ROM Files

Most manuals and ROM files can be found on IPDB.org for each specific game.

Note: The ROMs hosted on Planetary Pinball for World Cup has an issue. World Cup requires a special white flipper ROM specific to World Cup, which can be found on IPDB.

3.6.1 Combining System 3-6 Original ROMs Into a Single ROM at IC14

A System 6 TriZone MPU, featuring combined ROM images at U14.




For all System 3-6 games, the ROMs that might be located at U26, U22, and U21 (left to right on the top ROM row) can be replaced with a single 2716 2kb ROM (or EPROM) at IC14 that contains the combined images. When doing so, jumper J3 should be IN, jumper J4 OUT. These jumpers are located just above the ROMs and to the right of IC15.

3.6.2 Diagnostic ROMs by Pincoder

Free Diagnostic ROM images for Williams System 3-7 games can be found here:

They require the use of EPROM chips and a chip programmer. However, you can choose to skip sourcing EPROMs and the chip programmer by using the available Pincoder Adapter right out of the box. Simply plug it into the MPU board, select a test, and power on the game. Full documentation can be downloaded from the Pincoder website. The ROM images and the adapter are actively supported.

3.6.3 Flipper ROMs

The OS (Operating System) for a Williams pinball game is called the Flipper ROM. Flipper ROMs with the same color label can be considered generic, although there is at least one exception where a 'custom' White Flipper ROM was used "World Cup".

The Game ROM can be considered the 'personality' ROM, it provides the rules and objectives that are specific to that game's playfield layout. It also maps the Lamps, Solenoid and Switch Matrix to their specific purpose for that game and controls how they are sequenced and timed. Examples would be the 'attract mode' lamp sequence or when sound / speech select calls are made.

Because of large game production runs, Williams bought batches of Masked ROMs (fixed and not erasable) for the games. This was cheaper at the time than using Eproms (UV erasable, with a small window) as Eproms were still fairly expensive in the 80's. They used the same method for producing most of their Flipper, Game and Sound ROMs.

It is sometimes a good idea to replace old masked ROMs with EPROMs, since the original masked ROMs are 30+ years old and as the legs blacken and tarnish, they will weaken and fall off. For a similar reason, many of the ROM sockets on the MPU boards will need to be replaced, especially any sockets bearing the words 'Scanbe', which are poor quality. Masked ROMs are very stable as they start life as all 1's and then the information is programmed by "burning" each selected bit open, like blowing a tiny fuse. This process is not reversible, and ROMs will rarely lose their programming over time, unless the remaining whisker-like "fuses" start to fail. The practice and term to "burn" ROMs carried over to other storage media, such as CDs and EPROMs, even though different technologies and processes were used to store the data.

3.6.3.1 Flipper ROM Colors

Williams used standard ROM files for System 3-7. These two Flipper ROMs are located at IC17 and IC20. Systems 3-6 use two 2716 or 2316 Eproms, while system 7 used a 2716 in IC20 and a 2532 in IC17 (a 2732 can be used at position IC17 after modifications are made to the board).

3.6.3.1.1 White Flipper ROMs

White flipper ROMs were used in all System 3 games and two System 4 games...

  • Hot Tip
  • Lucky 7
  • World Cup
  • Contact
  • Disco Fever
  • Pokerino (system 4)
  • Phoenix (system 4)
3.6.3.1.2 White Flipper ROM Exceptions

World Cup uses White Flipper ROMs, but the ROM in IC17 is unique. The MPU will not boot and run with the standard White flipper ROM. However, the World Cup White flipper ROM can be used in all other White Flipper ROM games.

3.6.3.1.3 White Flipper ROM Issues

When downloading ROM software for World Cup, be aware the code available directly from Planetary Pinball has known faults and will not work properly. Code available from IPDB generally is correct and functional. The issues with these games is as follows:

World Cup: Provided with the standard IC17 Flipper Rom, not the unique one required for World Cup.

Lucky 7: When the original 512 byte files were combined into a 2k file for IC14 Game Rom, blank areas were entered incorrectly. This causes the machine to register a failed checksum test when attempting to enter settings via the dip switches. (Note: The copy on Planetary Pinball appears to have been corrected).

Hot Tip: When the original 512 byte files were combined into a 2k file for IC14, they were done out of order. (Note: The copy on Planetary Pinball appears to have been corrected).

3.6.3.2 Yellow Flipper ROMs

The yellow flipper ROMs were used in 2 games...

  • Stellar Wars
  • Flash (later production of Flash used Green flipper ROMs, which are preferred. The game ROM version must match the flipper ROM color)

3.6.3.3 Green Flipper ROMs

The green flipper ROMs were used in System 6 and 6A games from Tri Zone (Jul '79) to Alien Poker (Oct '80)...

  • Tri Zone
  • Time Warp
  • Gorgar
  • Laser Ball
  • FirePower
  • Blackout
  • Scorpion
  • Algar
  • Alien Poker
3.6.3.3.1 Green Flipper ROM Exceptions

Ted Estes developed a "divide by 10" ROM that allows /10 scoring for all 6 digit Green flipper ROM games with the exception of Algar and Alien Poker. These 2 games utilize an extension to the operating system ROM contained wholly in the game ROM which is incompatible with the Ted Estes hack (when enabled - with the hack disabled, the game works fine). Installing the hacked ROMs in these games isn't recommended since there may be other side effects.

The Alien Poker and Algar operating system extensions shift internal scoring before and after the normal routines. i.e. Internally, Alien Poker and Algar are 6 digit games.

3.6.3.4 Blue Flipper ROMs

The blue fipper ROMs were used in all System 7 games, Black Knight (Dec '80) and later...

  • Black Knight
  • Jungle Lord
  • Pharoah
  • Black Knight Limited Addition
  • Solar Fire
  • Barracora
  • Hyperball
  • Thunderball
  • Cosmic Gunfight
  • Varkon
  • Warlock
  • Defender
  • Rat Race
  • Time Fantasy
  • Joust
  • FirePower II
  • Laser Cue
  • Star Light
3.6.3.4.1 Blue Flipper ROM Exceptions

Star Light (June '84, 100 produced) which was the last System 7 game, appears to have non-standard Blue Flipper ROMs. For small production runs, games were supplied from the factory with EPROMs.

3.7 Circuit Board Schematics and Assembly Drawings

3.7.1 Connectors

3.7.2 Displays

3.7.3 Driver Board

Williams System 3-7 driver board layout



3.7.4 MPU

System 3 MPU IC Locator


System 3 MPU IC Identification Map.
The System 4 MPU adds only IC26 on the left side of the board, just above IC22 where the odd looking solder pads are.




3.7.5 Power Supply

3.7.6 Sound Board

During the long run of System 3 through 7 games, Williams used two basic board types, designated Type 1 and Type 2.

3.7.6.1 Type 1 Sound Board

System 3-4 Type 1 Sound Board Assembly Drawing and Schematic

3.7.6.2 Type 2 Sound Board

The System 6-7 Sound & Speech Boards Assembly Drawing, crafted by Phil Butcher is a great debugging resource.

3.8 Other Documentation

3.9 Errors

3.9.1 Power Supply

  • Note that the parts bill of materials for some (perhaps all) system 7 games call out a 2N6087 power transistor for Q5. This is most likely a typo. A 2N6057 or 2N6059 should be used.

4 Technical Info

4.1 Proprietary Numbers on Chips

Some Bally IC chips are marked with proprietary Williams part numbers, and no other markings are present. Below is a list of some of the most common chips found on Williams boards, and the more commonly referred to chip.

  • 5A-8987 - 6800 CPU chip, Motorola part #SC44216P.
  • 5A-9150 - 6808 CPU chip
  • 5A-8972 - 6820/21 PIA chip, Motorola part #SC44067P.
  • 5A-9903 - 6810 RAM chip

4.2 Backbox Overview

The Williams backbox is similar for most of the games in the System 3-7 games. System 3-6 remained fairly similar in their backbox layouts. System 7 introduced a number of changes.

  • During the beginning of System 7 the transformer moved from the backbox to the lower cabinet.
  • The System 7 PSU was redesigned and changed to incorporate a G.I. relay,
  • An upgraded System 7 MPU Board was used, which featured a built-in numeric diagnostic display.

The first two System 3 games (Hot Tip & Lucky Seven) used chimes and had no sound boards in the cabinet backbox. In the remaining System 3 games and all System 4 games, the sound board was located in the body cabinet, rather than the backbox.

Sample games of the first System 6 game, Tri Zone had the sound board located in the lower cabinet, but it was moved into the backbox for production games and all other System 6 & 6A games.

For the System 6A game Alien Poker and all System 7 games, an add-on board sppech board was introduced, which contained additional ROMs for speech.

4.3 MPU Boards

4.3.1 System 3

Williams System 3 MPU Board


System 3 was a major step by Williams to digital pinball games. It ushered in the Solid State machines, but to begin with there were problems with acceptance of the new machines.

Williams had the Driver board firing the solenoids on Chime units on the first two System 3 games, and even had a device called a 'noisemaker' (an EM score reel mounted in the body of the game) to provide the sounds of the score reels clicking over! This gave the player the familiar sounds of EM machines which were still around, The player's scoring, their credits remaining and the ball-in-play were all digital and being displayed on the seven-segment gas plasma displays.

Even once they dropped the chimes and 'noisemaker' (for World Cup in May 1978) and added a digital sound board, the operator could still select the more familiar 'bings & bongs'. A switch was provided on the sound board which allowed selection of simple digital chimes. This option persisted right through on System 3 to System 7 sound boards, but for the most part was not used. Digital sound became more popular as time went on.

The design of System 3 was based on the Motorola 6800 CPU and what has become a standard arcade peripheral I/O device, the 6821 PIA (originally the 6820). The PIA (or Peripheral Interface Adapter) on the MPU boards was used to drive the display I/O. All other peripherals, switches, lamps, solenoids, and sound, are driven by the 3 PIAs on the driver board.

System 3 and 4 boards use a Motorola 6800 microprocessor with a 6875 clock generator chip, while System 6 and 7 boards use either a 6802 or 6808 microprocessor. You cannot use a 6802 or 6808 in a System 3 or 4 board! The chips are not backwards compatible.

4.3.1.1 Upgrading a System 3 board to System 4

Overview of a board converted from System 3 to System 4


A system 3 MPU can be upgraded to use a single 2716 EPROM at what will be IC14 (System 3 schematics do not give this IC position a designation number). This essentially converts the board to a System 4 board. This upgrade is useful to reduce the number of sockets (and failure points). Install a 24 pin socket at the unpopulated area, just to the right of the IC17 socket. Combine the original game ROM code into a single 2716, and install it into the socket of what is now being called IC14. The flipper ROMs are installed in the same locations, ROM 1 into IC20, ROM 2 into IC17.

IC15 (on left), showing necessary cut and jumper


Procedure:

  1. Locate IC15 on the MPU (see picture at left)
  2. Locate the trace that runs from between pins 6 and 7 of IC15 to a via just to the right of the chip
  3. Sever that trace carefully
  4. Using jumper wire, connect the through hole via to pin 1 of IC15.


Doublecheck the work once the jumper is installed. Perform a continuity test between pin 1 and pin 9 of IC15. There should not be continuity. Conversely, there should now be continuity between pin 1 of IC15, and pin 20 of IC14.

Components which must be removed from a stock System 3 MPU board


The System 3 MPU board reset circuit should be updated as follows. This update was recommended by Williams in manual amendment 16P-482-110, and is done to improve CMOS RAM memory content protection during power-on/power-off.

10 Kohm 1/4W resistor added to circuit, upper left, connected to the banded end of the zener diode.

Procedure:

  1. Remove C27, R30, and R40 from the upper left corner of the MPU
  2. Install a 10 Kohm 1/4 watt resistor. One leg of the resistor will be soldered to the cathode of ZR1 (banded side - top). The other leg of the resistor will be soldered to the left through hole where R30 formerly was. The newly installed 10 Kohm resistor is referred to as R96 in System 4 MPU schematics.


Addition of the D19 1N4148 diode installed


Add a 1N4148 diode, if it is not already installed. The location is in the upper left quadrant of the circuit board. The anode (non-banded side) of the diode should be installed in a via, which is just to the right of resistor R23. This via is routed to pin 17 of IC19. The cathode (banded side) of the diode should be tack soldered to the solder pad where the right leg of resistor R23 is installed. Once installed, this 1N4148 diode is referred to as D19 on System 4 MPU schematics.
.

Addition of C68 is shown above, a .1uf 50v ceramic capacitor


Install a .1uf 50v capacitor just below the newly installed IC14 socket. One leg of the capacitor connects to +5VDC, while the other leg is inserted into a via which is routed to the /RESET line.

A factory board without the recommended capacitor and inductor update.
The recommended capacitor and inductor additions have been made to this board.


Add a 120pF capacitor (C66) and a 15uH inductor (L1) in parallel with the R1 resistor located just to the left of the crystal. This modification also applies to System 4 boards which do not have these components installed.

Here, the originally installed capacitor has been replaced with the correct capacitor. This capacitor is located just below the 5101 RAM


Finally, some System 3 MPUs came from the factory with a .01uf capacitor (marked 103) installed at position C22, (located just below IC19, the 5101 RAM). If so, replace that capacitor with a .1uf 50v capacitor. The purpose of this capacitor is to snub the high frequency noise created within the 5101 as silicon gates open and close. This applies to System 4 MPU boards too.

Once the above modifications are performed, a System 3 MPU board will be one-to-one compatible with a System 4 board.

4.3.2 System 4

Williams System 4 MPU Board
Another view of the Williams System 4 MPU Board


The System 4 architecture is little changed from the previous System 3 generation. The only functional changes from System 3 to System 4 was a minor alteration to the reset circuit as well as minor changes to the data and address bus in order to more easily facilitate the use of larger EPROMS for the game specific programming.

System 4 machines began with Pokerino in 1978 and ended with Flash in 1979. Late production Flash machines also utilized the then-new System 6 boardset in place of System 4.
All System 6 software will run on a System 4 board. The exception is Firepower which utilized both the bipolar proms and a 2716 game eprom. Due to the missing memory protection circuit and somewhat unstable clock circuit replacing a System 6 CPU with a System 4 CPU is not recommended.

4.3.2.1 Upgrading a System 4 MPU Board

Changed layout with additional capacitor and inductor to the R1 resistor


The System 4 board has a revised clock circuit which usually does not need to be modified. If there are reset problems, replace the 22ufd tantal capacitor with a 33ufd type. This will lengthen the reset period a little and helps the clock circuit to stabilize. The capacitor modification can also be applied to System 3 boards.

If an incorrect capacitor is installed at position C22, (located just below IC19 - 5101 RAM), it is recommended to replace it. This replacement is only necessary, if a 0.01mfd cap is installed. If so, replace the 0.01mfd capacitor with a .1mfd 50v capacitor. This applies to System 3 MPU boards too. Only early system 4 boards had an incorrect capacitor installed. Also check if the 0.01mfd blocking capacitor for IC14 is missing. If yes stuff it.

4.3.3 System 6

Williams System 6A MPU Board. Note that the .156 headers on this board have been replaced and an LED indicating the blanking status has been added


System 6 was an upgrade to the System 4 design, replacing the 6800 processor with a 6808. They took advantage of the internal clock in the 6808 CPU, removing the need for the 6875 clock generator (a companion chip to the 6800 and now obsolete and impossible to find).

This also changed the way the 'watchdog' circuit functioned. The watchdog monitors the CPU (IRQ signal), and will blank the displays and lamps as well as stopping the solenoids from firing. When something goes wrong (perhaps the MPU board has locked up), 'blanking' should prevent further damage to the game by 'locked on' coils and other output components driven from PIAs on the MPU/Driver boards.

More ROM memory could be addressed, and this was used to hold more sophisticated game code. As an example, Firepower used a 2716 2K Game ROM (which was standard for IC14) plus 3 x Harris bipolar proms (512 bytes each) giving a total game code size of 3,584 bytes. Having just 3K of code space is nothing by today's standards. Programmers had to work hard to get good game code in such a small space.

During early System 6 revisions the transceiver chips 8T28 ICs at IC9 and IC10 were found to not be needed. They were eliminated from the design as they didn't need to amplify signals to and from the data bus with modern ROMs. If the ICs are working, leave them alone. But if not, these buffers are obsolete and can be removed and bypassed, in effect by jumpering the data bus pins of the CPU directly to the data bus. i.e. IC9 and IC10 may be removed, and jumpers installed across pins 2-3, 5-6, 9-10, and 12-13 for both ICs.

Also around this period of game development, 8T97 bus "amplifier" ICs at IC3, IC4, and IC8 were used for address and control signals from the 6808 to peripheral ICs. 8T97s are long obsolete too, but fortunately can be replaced with an 74LS367, which Williams itself used in some revisions of the board. 8T97s can not be removed and jumpered like IC9 and IC10. They must be in place. Testing these chips is easy with a logic probe. The input signal should closely match the output signal of the amplifier. i.e. pins 2/3, 4/5, 6/7, 10/9, 12/11, and 14/13 should all test identical with a logic probe. Note that since the signal is amplified, logic probe indicators might vary somewhat. But, the probe should not measure high on one pin and low on the companion pin, for example.

The 8T97 bus amplifiers communicate the following signals.

  • IC4 - A0 thru A5
  • IC3 - A6 thru A11
  • IC8 - A12 thru A15, VMA (valid memory address), and R/W (read/write)

On Firepower alone, the 'Combo ROM' modification must be done to allow the use of a 2732 EPROM as the Game ROM at IC14. This replaces the Firepower game code found in 4 ROMs as described here, and is an excellent idea. Less chips to worry about, less sockets to replace.

Williams also transitioned to using the updated 6821 PIAs during System 6 because the MC6820 was phased out after 1977.

4.3.3.1 Replacing the 6808 with a 6802

J1 installed for 6808 or 6802 CPU chip
J1 removed, 4.7 Kohm 1/4w resistor (R4) installed for use with 6802 CPU chip


The typically stuffed 6808 microprocessor can be replaced with the more easily obtained 6802 microprocessor. The 6802 uses the same instruction set architecture and pinout, making it "backward compatible" with the 6808. The 6802 incorporates an internal 128 x 8 bit RAM memory. With the proper strapping, the 6810 RAM located at IC13 may be removed when a 6802 is installed. Factory strapping grounds pin 36 of the 6808/6802 socket via jumper J1. Pin 36 must be pulled high by tying pin 36 to 5V via a 4.7K pullup resistor. This enables the 6802's internal RAM. The 6810 at IC13 can then be removed or (if not socketed) simply left in place.


Procedure

  • Replace the 6808 with a 6802
  • Remove jumper J1
  • Install a 4.7K resistor at R4

Officially, the 6808 has no internal RAM. But, internally it's identical to the 6802. Sometimes the 6808 internal RAM works but this is not guaranteed, so when using a 6808, be sure pin 36 is grounded.

4.3.3.2 Factory Jumper on System 6 MPU Board

Williams System 6 MPU board with factory jumper


Williams System 6 and System 3 MPU board comparison

Some System 6 MPU boards have a jumper installed between IJ3 pin 1 and IJ4 pin 1. This is a factory jumper, and should not be removed. Its purpose is due to the addition of the memory protect circuit, which was incorporated with the System 6 MPU. System 3 and 4 MPU boards did not have a memory protect circuit. The method in which this was added is the addition of pin 1 of IJ4. This pin was formerly the keying pin on System 3 and 4 MPU boards. However, Williams failed to run a trace to this pin for the memory protect circuit. Hence, a jumper was installed.

If using a System 6 MPU board in a System 3 or 4 game, it is recommended to remove the keying plug from the IJ4 connector in the game, as opposed to cutting / removing pin 1 or IJ4. This will allow the MPU to be used in any System 6 games in the future.

4.3.4 System 6A

Williams System 6A MPU Board

The System 6A MPU is a System 6 board without buffer chips at IC9 and IC10. In place of these chips were several jumpers in closed frames, which can look similar to a socket at first glance. The jumpers are used to pass the data lines from the CPU chip to the data bus.

System 6A also used four 7-digit slave displays and one 4-digit slave display driven by a centralized display driver board. This separated the displays from any board logic. Prior to this point, a master display driver board (with a 6-digit or 4-digit glass installed directly on the driver board) housed the display logic and drove four 6-digit slave displays.

Some System 6 boards that are not explicitly marked as 6A in the lower right corner were modified to be 6A boards, which can be identified by the lack of the IC9 and IC10 buffer chips. Either several uninsulated wire jumpers will be installed in the solder pads of these two chips, or the same closed frame jumpers as seen on actual 6A MPU boards may be present.

4.3.4.1 Modifying a System 6 or 6A to Use The Firepower 2732 Combo ROM at IC14

Cuts made, jumper run, and 4.7Kohm resistor added
2 1N4148 diodes added where jumpers J3 & J4 previously were


The factory ROMs for Firepower use all 6 ROM socket positions on a System 6 or 6A board. To minimize the amount of chips and sockets used, some modifications have to be performed to the MPU board. In doing these modifications, the board will then be able to use "standard" green flipper ROM chips at IC20 & IC17, and a single 2732 combo ROM at position IC14. Below are instructions pulled from the readme.txt file, which is included with the Firepower L2 ROM code found on www.ipdb.org. Some of the procedure wording has been altered for clarification.

There are similar methods which may or may not work besides this method. Note that the jumper wire can be connected from IC14 pin 21 to either IC30 pin 14 or pin 36 of the 40 pin interconnect. If at all possible, please make sure that the board being modified boots and works with other game code, such as Gorgar, Flash, or Firepower, before attempting to make these modifications for the Firepower combo ROM.

  1. Remove Game ROMs (IC14, IC21, IC22, IC26) from their sockets.
  2. Remove jumpers that may be installed at J3 or J4 (to the right of IC15, a 74LS139).
  3. On the solder side of the board, at IC14 pin 21, sever the trace on both sides of the solder pad. Removing the existing solder on the pad may help. The idea is to sever the +5V connection to the pin, while still leaving the solder pad at the pin itself.
  4. On the solder side of the board, install a jumper from IC14 pin 21 to IC30 pin 14. IC30 is generally not installed. It is a 16-pin chip outline just to the right of IC14 and just above the 40 pin interconnect. Solder the jumper between the hole at pin 14 of the empty IC30 location, and pin 21 of IC14, making sure to not short it to the other side of the trace previously cut at IC14. This connects A11 to IC14.
  5. On the solder side of the board, install a 4.7K ohm, 1/4 Watt resistor between pins 20 and 24 of IC14. With the resistor leads bent at 90 degrees at a comfortable distance from the body of the resistor, the legs should line up at just the right spacing to solder one lead to each of those pins. Again, make sure to avoid shorting the pin 20 connection to the jumper wire at pin 21, or the cut trace edges around pin 21. This provides a pullup for the active-low chip select.
  6. Install two fast switching signal diodes (either 1N914 and 1N4148) at locations J3 and J4. The cathode (banded) end should be oriented toward the TOP of the board for both diodes. This enables both of the two Game ROM address ranges to drive the active-low chip select at IC14.
  7. Program a 2732 with the "FPCOMBO" ROM image (found here) and install at IC14. If the board doesn't already have working Green Flipper ROMs at IC17 and IC20, those should be installed also. Note the orientation of the ROMs with respect to pin 1 of the chip sockets.


4.3.5 System 7

Williams System 7 MPU Board

System 7 was considered a major step change. It had a redesigned MPU board, now supporting a single 7-segment LED display for indicating improved diagnostic information instead of the original 2 LEDs that System 3-6a MPU boards had used. It also added commas to the player scoring displays and moved the sound select support to the MPU board. An extra 6821 PIA supported both the sound/speech selects and the display of commas. An extra 12-pin header at 1J8 was added to provide connections for the new Sound and Commas support. This freed up five solenoid drives at positions #9-13 on the Driver board, which had been sound/speech selects. They were then available to drive extra game Coils or Flash Lamps.

The MPU used two 2114 Static RAMs; these 1024 x 4 bit RAMs replaced the use of 6810 RAMs mentioned above. There was extended memory addressing, support for multiple 2732 ROMs (or EPROMS) as standard and a huge number of jumper selections available. The jumpers support various memory addressing schemes and ROM sizes, making the System 7 board MPU "backwards compatible" and able to emulate any of the previous System 4-6a games. (Provided, of course, that the correct Jumper Settings and EPROMS are installed.)

The Sound and Speech boards were unchanged for System 7; both sound and speech boards remained compatible from their introduction for Gorgar. In some cases the System 7 game had no 0.100" 40-way IDC header for the speech board connection. This was a cost saving measure made by Williams for games produced without speech. This connector is cheap and available today as it is still used for PC IDE hard drives and modern PCB connections. Adding this connector back to the sound board allows it to support a speech 'daughter board' by removing the Jumper at W1.

The Driver Board mates with the MPU board using 40 x 0.156" header pins on the MPU and female sockets on the Driver Board. This is a continual source of repair problems for this era of Williams machines. To solve this, when designing System 9 Williams combined the MPU and Driver Boards (and the Sound Board) onto a single PCB (Printed Circuit Board), and removed the problems associated with the now infamous Williams "40-way" connector. Only the speech board remained separate, as digital speech was considered an optional feature.

4.3.5.1 System 7 MPU Jumper and IC Locator

Positions of the jumpers on a System 7 MPU board
IC Locations on a System 7 MPU board

System7ICLocator.png
The image at left shows the most common jumpers on a System 7 MPU board. There are other jumpers on the right hand side of the board (not shown in the pic). However, these jumpers are rarely added or removed, nor do they pertain to the size of the ROM chips installed on the board.

The image at right is provided to assist in IC location.

The factory release of most System 7 MPUs contained a 2532 PROM at IC17. The more common 2732 is easy to use, with minor board modifications.

Component side of a System 7 MPU board with modifications for a 2732 at position IC17
Solder side of a System 7 MPU board with modifications for a 2732 at position IC17


  1. On the component side of the board, move jumper 22 to jumper 23. This amounts to moving the jumper down one position.
  2. On the solder side of the board, cut the fat trace that grounds pin 21.
  3. Jumper from the via labeled "22" to pin 21 of IC17. Be sure that the jumper doesn't short to the jumper previously moved on the component side.


4.3.5.2 System 7 MPU Test Pad Info

There is an unpopulated connector in the lower right corner of the system 7 MPU board. You can use a meter or logic probe to test the signals that are present on it or possibly add a connector to it (you can use a .100 dual row 20x2 connector).

Unpopulated Connector


gnd 40 39 gnd
d7 38 37 ic36 CA2
d6 36 35 ic37/15
d5 34 33 ic37/14
d4 32 31 ic37/13
d3 30 29 ic37/21
d2 28 27 ic36 pb4
d1 26 25 ic36 cb2
d0 24 23 blanking
reset 22 21 vma
r/w 20 19 bus 02
a14 18 17 a15
a12 16 15 a13
a10 14 13 a11
a8 12 11 a9
a6 10 9 a7
a4 8 7 a5
a2 6 5 a3
a0 4 3 a1
+5vdc 2 1 +5vdc
Bottom of board

4.3.6 MPU Compatibility

Williams 3-7 MPU are pretty flexible. With some modifications, boards can be upward compatible, while others are downward compatible.

Here are the details:

  • A System 3 MPU can be upward compatible to a System 4 MPU, if the upgrades are performed as described above. And if a System 3 MPU is upgraded to a System 4 MPU, it is then upward compatible with a System 6 MPU board. There will not be a memory protect circuit though.
  • System 4 and 4A MPUs are upward compatible to a System 6 MPU, when a socket is installed and used at position IC14. However, there will not be a memory protection circuit. A System 4 MPU is backward compatible with a System 3 MPU with no modifications performed.
  • System 6 and 6A MPUs are backward with System 3, 4, and 4A MPU boards with no modifications performed.
  • A System 7 MPU is backward compatible with System 3, 4, 4A, 6, and 6A MPU boards. However, board jumpers will have to be removed / installed. It not recommended to use a System 7 MPU board in any previous systems, but it can be done.
  • System 3, 4, 4A, 6, and 6A MPU boards are not upward compatible with System 7 MPU boards.

In any instance where a non-native MPU board is used in place of the original MPU board, the correct flipper ROMs and game ROM must be installed. An example would be if an upgraded System 3 MPU were to be used in Gorgar. The GREEN flipper Roms and Gorgar game ROM would have to be installed in postions IC20, IC17, and IC14.

4.3.7 Aftermarket Boards

A Rottendog MPU327, which can be installed in any Williams System 3 - 6 game.
U17 on an MPU327, stuffed correctly with a 74HCT9114.

Rottendog offers aftermarket Williams 3 through 7 combined MPU/driver boards. The Rottendog board (pictured left) contains all the game code for every System 3 through 6 game. The game code is DIP switch selectable. Note that the "sense" of the DIP switch bank is backwards. Follow the on-board silkscreen to the left of the DIP switch.
Note: as of January 2018, Rottendog is no longer claiming support for System 7 games with their MPU327 board.

Some older Rottendog MPU327s were stuffed wrong at the factory. U17, which services the switch columns, should be stuffed with a 74HCT9114. In some boards, that location is stuffed with a 74HCT244, which will not work correctly. Note that there are several revisions of the MPU327 board.

A Rottendog MPU327, installed, and shorted to a switch matrix header pin. Image provided by Tony, PinSider "goingincirclez". Used with permission.
One user's (excellent) correction for the situation. Image provided by Tony, PinSider "goingincirclez". Used with permission.


The RottenDog MPU327 PCB design causes the switch matrix row header (2J3) to short to a mounting bracket as shown at left. This can cause the switch column IC at U17 to fail. A simple correction for this issue is shown at right. See the original Pinside thread for additional details.

Rottendog MPU327 game select dipswitch bank


Note that the dipswitch bank used for selecting a specific game has 8 dipswitches, whereas only the 6 switches on the right are the only ones used. The 2 left dipswitches must be turned "on" (opposite of what is expected of a dipswitch) to "0".

4.4 Driver Boards

4.4.1 System 3-6

Williams System 3-6 Driver Board

The driver board uses three 6821 Peripheral Interface Adapter (PIA) chips to drive:

  • an 8x8 Matrix for the Lamps
  • an 8x8 Matrix for the Switches
  • Solenoids and Sound Selects

More properly, the PIAs were used to drive transistors that in turn drive the lamp matrix, switch matrix, and solenoid circuits. In the case of the solenoids, Darlington transistors are used as "solid state relays", providing a path to ground for coil power, causing them to energize.

The electronic design choices Williams made became the basis for all subsequent Bally/Williams pinball games and even games being made today. The specific CPU and PIA ICs may change, more memory and ROM space is available to the programmers, MOSFETs are used as the switching relays instead of transistors...yet the basic design established for the lamp matrix, switch matrix, and for energizing coils, remains unchanged.

4.4.2 System 7

Williams System 7 Driver Board


The driver board remained almost completely unchanged from System 3 right through to System 7 with the exception of one small change that was made to the driver board during System 7. Eight resistors were changed to zero-ohm jumpers in the switch matrix inputs to increase sensitivity and responsiveness. These zero ohm resistors are located in the upper right corner of the driver board.

4.4.3 System 7 Hyperball Specific Driver Board

Williams System 7 Hyperball Specific Driver Board


Since Hyperball didn't need to drive coils other than the cannon coil and the auger motor, and since it needed to drive many more lamps than a normal pinball game, Williams created a unique, special purpose variant of the driver board specifically for Hyperball. This driver board can only be used in Hyperball, and a regular System 3 through 7 driver board can NOT be used in Hyperball.

4.5 Power Supplies

Early Williams System 3-6 Power Supply


The System 3-6 power supply used in Hot Tip and Lucky Seven have a few additional components that later games did not need, physically located a little above the row of four fuses: two .22mfd mylar capacitors and two 1N4001 diodes. If a later power supply is being installed in Hot Tip or Lucky Seven, these components must be added to the board if they are missing.

Williams continued to populate these components for several following games, but they were not utilized. These components can be safely left populated or simply removed from the board.

Williams System 7 Power Supply (GI input connector modified from stock)


Note that the parts bill of materials for some (perhaps all) system 7 games call out a 2N6087 power transistor for Q5. This is most likely a typo. A 2N6057 or 2N6059 should be used.

4.6 Flipper Power Supply

  • Williams System 7 C-9939 50V Flipper Power Supply Board
  • Front
  • Back

This PCB was first used in Hyperball to control the auger (ball lift) motor. As seen in the pictures above, the PCB layout was designed to accommodate more components than were ever stuffed. This is because this same board, using different components, is used in some Williams video games. In Firepower II, Laser Cue and Star Light the trimmed down version shown here provides the flipper coils with 50VDC power. The AC voltage supplied to the bridge rectifier originates from a second, smaller, transformer mounted on the wooden plank in the bottom rear of the cabinet.

Since the coil voltage this board provides is exactly the same as that of later System 11 games, far stronger Williams coils, such as the FL-11630, can be used in these games. It's advisable when using these coils to install a 2.2µF capacitor across each EOS switch, to reduce spark across the EOS switch, as was done in later Williams System 11 games.

The board suffers from cold solder joints at the header pins (at least). As shown in the picture above, someone did a (poor) job of reflowing the solder on the header pins. A better solution would be to replace the header pins altogether.

The exact same board is used in all System 9 and early System 11 games up until F-14 Tomcat. Data East copied this design. Hence, this 50VDC Power Supply can be used in Data East games from Laser War to Time machine.

4.7 Sound Boards

During the long run of System 3 through 7 games, Williams used two basic board types, designated Type 1 and Type 2

4.7.1 Type 1 Sound Board

Williams Type 1 Sound Board front
Williams Type 1 Sound Board back


The Type 1 sound board is rectangular in shape as pictured at left. Two versions of this board were manufactured. One version placed the 12,000uf 5V filter capacitor vertically. The other version placed the same capacitor horizontally. The boards function identically.
As can be seen on the back of the board pictured at right, the board was modified from the factory to add 3 jumpers (the three longer red wires) to make it work correctly. Every Type 1 sound board should have these jumpers. Note: The factory installed jumpers take advantage of trace "vias" to accomplish these connections.

  1. A long jumper that connects the reset pins on the 6821 PIA (pin 34) and the 6802 (pin 40)
  2. Pin 40 of the 6802 also needs to be jumpered to pin 36 of the 6802 for the reset circuit to be complete.
  3. Jumper pin 35 to pin 8 of the 6802. This ties Vcc1 and Vcc2 together.

Williams documented additional modifications in the "Flash" manual. The board pictured has been updated with all of these modifications. A drawing of these modifications and jumpers can be found here.

The modifications are:

  1. Add a jumper between pins 39 and 40 of the 6821 PIA
  2. Add two 10K, 1/4W resistors
  3. (Possible) replace two 100K resistors with 4.7K, 1/4W resistors. Note that some boards shipped with 4.7K resistors installed.

Modifications 1 and 2 "reduce the susceptibility of the sound board to noise". The "noise" can cause the board to stop execution after a short time, especially when using a 9316B or 2716 ROM. If a sound board makes sound after pressing the diagnostic button, but either makes no other sound, or makes sound for a short time and then quits, these two mods will probably fix this problem. Modification 3 "improves the quality of the sound produced at the speaker".

Theory of Operation

The Type 1 sound board is about as basic as a sound board can be.

  • A 3.58Mhz crystal provides the clock signal across pins 38 and 39 of the 6802.
  • A simple reset generator is configured with Q1, Q2, and CR3 (a 6.8V zener diode). As the unregulated 12VDC rises past 6.8V, the zener breaks down, eventually allowing pin 40 of the 6802 to be pulled up to logic high (5V).
  • A diagnostic pushbutton, upper right, grounds the NMI (non-maskable interrupt) signal to the 6802 processor and causes the board to play a diagnostic tune (in most cases).
  • A two position switch selects between "sounds" (right) or "chimes" (left).
  • A volume adjustment POT is located on the board itself, and is a 5K POT.
  • A 6802 microprocessor controls the board via a 6821 Peripheral Interface Adapter (PIA).
  • The board does not have a RAM IC (like a 6810). Instead, the built-in 128 bytes (yes, 8-bit bytes) of RAM on the 6802 are used.

Power is provided from the transformer secondary at connector 10J1.

  • Pins 1,2 - 9VAC
  • Pins 5,6 - Ground (center tap)
  • Pin 7 - Key
  • Pins 8,9 - 9VAC

Power is fused and rectified via the "in-line" bridge rectifier immediately below the connector. Positive unregulated 12VDC is filtered by a 12,000uf/25V electrolytic capacitor. Negative unregulated -12VDC is filtered by a 1,000uf/25V electrolytic capacitor. The positive unregulated 12VDC is regulated down to 5VDC via a 7805 regulator at IC11. These three voltages, plus ground, define the power requirements for the board.

A speaker is connected to 10J2, positive to pins 1/2, ground to pins 3/4.

Sound selects are presented by the driver board to the sound board at connector 10J3. Selecting a particular sound is a matter of grounding the appropriate sound select pins which are normally held high via 4.7K pullup resistors at R37 through R44. Each signal is buffered via the 4050 non-inverting hex buffers at IC8 and IC9.

Each of the sound select signals is connected to both a 4068 8-input NAND gate and to the (software configured) input port of the 6821 PIA (PB0-PB7). The output of the 4068 is connected to pin 18 (CB1) of the 6821 which is configured by the sound board software to create an interrupt to the 6802 processor. When the 6821 PIA interrupts the processor, the interrupt servicing routine reads the input port (PB0-PB7) of the PIA to determine which sound was commanded. The processor pulls sound data from the sound ROM and pushes the data to the 6821 PIA (software configured) output ports (PA0-PB7) one 8 bit byte at a time. The PIA output port signals drive a simple 1408 DAC (digital to analog converter) which produces the analog signal that is stabilized and eventually amplified by a TDA2002 audio amplifier.

What Goes Wrong

The Type 1 sound board shares most of the common faults associated with the Williams System 3-7 board set. Chief among them are:

  1. ScanBe sockets. These are well documented to be long past their useful life and should always be replaced.
  2. Fractured solder joints at the header pins. The Williams manufacturing process trimmed the .156 header pins too short, damaging the solder meniscus. Over time, these solder joints develop fractures. The header pin eventually begins to work itself out of the joint which causes reliability issues.
  3. Aged electrolytic capacitors. At the very least, the 12,000uf/25V 5VDC filter capacitor should be replaced. Replacing all of the electrolytic caps is advised. Following is the capacitor parts list.

Electrolytic Capacitors Used

  • C10 - 25uF, 25 Volt Axial (can use 22uf)
  • C13 - 500uF, 16 Volt Axial (can use 470uf)
  • C14 - 800uF, 16 Volt Axial (can use 1000uf)
  • C27 - 100uF, 10 Volt Axial
  • C29 - 12000uF, 16 Volt Axial
  • C30 - 1000uF, 16 Volt Axial
Williams Type 1 Sound Board Version 2 with the horizontal cap, front
Williams Type 1 Sound Board Version 2 with the horizontal cap, back


These images show a Type 1 sound board that was introduced later in production. The 5VDC filter capacitor was turned horizontally. Also, an extra connector was added to enable a remote volume POT. Note that this sound board has been "recapped".

This revision of the sound board picked up the jumper modifications of the earlier board in the PCB design, with one exception. Pins 39 and 40 of the 6821 need to be bridged (or jumped) to avoid the board locking up after a short (and random) period of time. A solder bridge can be seen at pins 39 and 40 of the sound board pictured at right.

Williams Type 1 Sound Board Version 2, closeup of jumper to enable onboard sound POT


Later production Type 1 sound boards added a connector to enable a remote volume POT. If the remote POT is not used and the onboard POT is used instead, a jumper must be installed between pins 1 and 2 of this connector. The pictured jumper is quite Gucci. A simple wire jumper on the back of the PCB will suffice.

4.7.1.1 Type 1 Sound Board Jumpers

Williams Type 1 Sound Board

The Type 1 Sound board can be jumpered to use a 7641 masked ROM, a 9316B masked ROM, or a 2716 (or doubled 2732). The board pictured at left is jumpered to accept either at 9316B or a 2716. To use a 7641 masked ROM, remove the four jumpers to the right and just below the sound ROM, and install the opposite set of four jumpers.

4.7.2 Type 2 Sound Board

  • Williams System 6-7 Type 2 Sound and Speech Boards
  • Williams System 6-7 Type 2 Sound Board
  • Williams System 6-7 Type 2 Sound Board with Speech Board Connection
  • Williams System 6-7 Type 2 Sound Board with Speech Board. Note that this -3 revision board features a tantalum capacitor soldered directly between test point 1 (+12VDC) and test point 3 (ground). In later board revisions, this part was added to the PCB design.

4.8 Speakers

With the exception of Hot Tip and Lucky 7, all Williams System 3-7 games use a speaker. And, nearly all of them use a single 6-1/2" speaker located on the cabinet bottom (later games, like Firepower II, had an oblong speaker below the smaller backglass instead). Speakers used during games such as Firepower were 8 ohm 5 watt.

4.9 Displays

Williams System 3-6 Master Display Board (Using UDN7180A - Common)
Williams System 3-6 Master Display Board (Using NE584 - Less Common)
Williams System 3-6 Master Display Board w/ Discrete Components
Williams 6-Digit Slave Displays
Williams System 6A-7 Display Driver Board
Williams 7-Digit Slave Displays
Williams 6-Digit Puck Bowler Display Glass
Williams 6-Digit Puck Bowler Display Edge Connector


4.10 Other Boards

4.11 Flippers

4.12 Drop Target Banks

4.12.1 First Generation Drop Target Bank

Typical first generation Williams solid state drop target bank (Time Warp)

Starting with the first Williams solid state game, Hot Tip, Williams introduced their version of the solid state drop target bank. Drop target banks were a rather new design for Williams. They were the last major pinball manufacturer to introduce the drop target bank earlier in 1977 with the electromechanical game, Big Deal. And, only 3 games prior to the solid state version of Hot Tip, including the electromechanical version of Hot Tip, were manufactured with drop target banks.

The reason for mentioning the background regarding Williams drop target banks is because of the actual design implemented. There is one distinct difference between Williams drop target banks and any other manufacturer's drop target bank. In addition to using a scoring switch for each drop target in a bank, Williams used a secondary set of switches paired in series to detect when all of the drop targets were physically dropped. The drop target bank frame, reset coil or coils, appearance of the drop target faces on the playfield, and operation of the bank was virtually the same between Williams electromechanical games and solid state games. Therefore, it is believed that Williams designed their banks this way, because they had not been designing drop target banks that long, and decided upon a legacy approach. In other words, the Williams solid state drop target banks used hardware instead of software to detect when all the drop targets were down.

Backside of Williams first generation drop targets


The design of the drop targets are very different from any other manufacturer's design. Unlike Bally, Stern, or Gottlieb, the Williams drop target is a meaty, reinforced plastic which is fairly consistent in thickness and width throughout the target. Williams does not use a face with a long neck stemmed down in size. The benefit to this design is that breakage is minimized. The downside to this design is extra added weight. There are several other design differences too, such as the compression spring used to pull the target down. It is physically screwed to the front of the drop target plastic, as opposed to being attached via a molded "hook" like other manufacturers. This too is a plus, because the molded hooks on other drop target systems are prone to breakage. However, probably the most recognizable difference is the drop target switch design. On the back side of a Williams drop target, a switch triggering mechanism in the form of a copper horseshoe with two contacts is physically screwed to the plastic. The combination of the horseshoe contacts and discrete circuit boards for each target mounted to the bank of the bank are what is used to detect the three states of the drop target position.

Simulation of the three different states of a Williams drop target

A simulation of the operation of a Williams drop target switch can be viewed at the animation located to the left. As mentioned before, the horseshoe contact is attached to the backside of the drop target plastic. The horseshoe contact makes physical contact with the thick copper or tinned traces of the circuit board, which is secured to the backside of the drop target bank.

An explanation of the three different states of the drop target are:

  1. When the horseshoe contact is located at the uppermost set of traces, the first frame in the animation, the drop target is in the upright position (fully reset).
  2. Once the drop target face is struck by the ball, the drop target plastic is pushed from the metal ledge of the metal drop target frame. As the drop target falls, the horseshoe contacts slide over a second set of traces on the circuit board, as seen in the second frame of the animation. These contacts are tied to the switch matrix via a switch strobe and return line respectively. The switch strobe line is common to all drop targets in the bank, while each drop target in a bank will have its own switch return line. Each switch pair is used for scoring purposes.
  3. Finally, the drop target will fully drop, and the contacts come to rest on a third set of traces on the circuit board. This set of traces is tied in series from one drop target circuit board to the other. Because the pairs of traces for each target are tied together in series, a switch, secondary to the individual scoring switches is used. Once all of the targets in the bank are dropped, the series switch circuit is complete. The switch matrix detects this as a single switch closure. In turn, the drop target bank is reset.


Switch solder points identified on a drop target bank




With the maze of similar-colored wires on the back of a Williams first generation drop target bank, it can be somewhat intimidating as to what wire goes where. The switch return lines (white wires with colored traces) attach to the solder tabs of the eyelets. Likewise, the non-banded side of the diode attaches to the same solder tab of the eyelet. The switch strobe line (green wire with a colored trace) can be attached to any of the points circled in green. On the bank in the adjacent pic, the switch strobe wire was attached to the green circled area on the far right. The black and gray wires are sometimes substituted with different colors like yellow.

4.12.2 Second Generation Drop Target Bank

Typical second generation Williams drop target bank




Starting with Black Knight, Williams slightly changed the design of the drop target bank. Instead of using horseshoe contacts and circuit boards for switch detection on each drop target, the horseshoe contacts were replaced with a copper leaf spring and the discrete circuit boards were replaced with leaf switch stacks. Detecting when all the drop targets down was now handled via software. There are subtle physical differences in the drop target plastic. The most obvious difference is the thickness of the bottom of the drop target. The thickness was increased to compensate for the use of bottom mounted leaf switches.

4.12.3 Third Generation Drop Target Bank

+++ Need pic of drop target bank with microswitches +++

The third and final generation of System 3-7 drop target banks are essentially the same as the second generation. The main difference is that microswitches were used to detect when a drop target drops versus leaf switches.

4.13 Additional Diagnostics - Pincoder test ROMs

A comprehensive set of free, third party diagnostic/test ROMs can be found here: http://pincoder.ca

These ROMs can be used in-game to test individual components of your pinball machine. Additionally, the RAM and CMOS tests thoroughly check all addressable locations within each chip, resulting in extremely accurate results.

Support for these ROMs can be found on Pinside here: https://pinside.com/pinball/forum/topic/new-williams-system-6-in-game-test-roms

5 Recommended Modifications

5.1 Additional Fuses

Fuse holders with 8 amp slo-blo fuses added to one of the AC inputs to each of the bridge rectifiers. (Firepower)
Inline fuse holders installed with solderless crimp butt connectors


Williams failed to include fuses in the AC connection between the transformer secondary and the solenoid and controlled lamp bridge rectifiers. In theory, the primary power fuse should blow if one of these bridges short. However, this is not always the case. Sometimes, the bridges short in such a manner that the wiring from the transformer secondary to the bridge becomes the fuse, and the wiring insulation melts until something gives.

This problem can be addressed by adding 2 fuse blocks and two 8 amp slo-blo fuses. For each bridge, afuse is added between either transformer secondary wire and the bridge. The AC inputs to the bridge are both marked with a 'tilde' on the BR which looks like an 'S' on it's side: ~. The wire colors are typically red for solenoids, and blue for the controlled lamps. However, it is best to consult the game schematics to confirm.

Although the preferred method is to put fuse holders on the backbox wall, installing fuses in any configuration is paramount. One way is to use inline fuse holders, and solderless crimp butt connectors. This is a quick and easy alternative, even though it may not be the prettiest solution.

The Firepower schematics, updated to include two 8ASB fuses between the transformer secondary and both the solenoid and controlled lamp bridge rectifiers.


Williams incorporated an 8A normal-blo fuse in later games (Big Guns and newer), while Data East chose to install 8A slo-blo fuses.

A nice solution for this problem is the Inkochnito Bridge Board that can be found at this link: http://www.inkochnito.nl. Click on the Bridge Board image for more information.

System 3 and System 4 games did not incorporate a fuse for the 9.3VAC supply lines. If either diode D7 or D8 shorts, something in the circuit will act as a fuse. Sometimes, a board trace acts as a fuse under these conditions. It is much, much better to add two 4 amp slo-blo fuses directly after the transformer, just as Williams did when System 6 games were introduced.

5.2 MPU Batteries & Memory

5.2.1 Relocating the batteries away from the MPU board

Leaking Battery

Relocating the 3xAA batteries from the MPU board is always a good idea. Leaky alkaline batteries are the #1 killer of pinball boards. Sometimes the battery terminals don't look corroded, but the metal rivet which contacts the battery are actually missing.

If "04 00" is displayed in the credit/match display, rather than the game going into attract mode, then the game is in audits mode:

  • The batteries have failed and need replacing
  • The battery voltage is not reaching the 5101 CMOS RAM (check Test Point TP7, which should measure 4.3v with the power on and 3.9v (or so) with the power off.
  • Blocking Diode D18 (1N5817) has shorted, and the batteries are trying to run the MPU board when the game is off.
  • There is another problem holding settings, such as a faulty 5101 CMOS RAM.

Simply removing the batteries is not really an option as the game will not boot directly into "attract mode" when switched on. It also will not retain the settings such as the number of balls per game, the free play setting (usually obtained by setting maximum credits to 0), or high scores It's possible on System 3-7 games to switch the game off, and then quickly back on to go from the '04 00 audit mode' to attract mode. Credits would still need to be added by the coin switches.

5.2.2 Remote Battery Pack

One option is to remotely locate the battery holder somewhere below all the other boards. This ensures that even if the remotely located batteries leak, they won't leak onto (or even drip onto) any circuit board (see driver board pic below). Use good quality alkaline batteries, mark the date of replacement with a Sharpie, and replace the batteries annually.

Adding a connector between the battery pack and the MPU board is a good idea, so the pack can easily remove the battery pack from the board. Plus, it the batteries are forgotten and do leak, the MPU board will not have to be removed to add another battery pack. Another solution is to buy a battery holder with the 'transistor radio' type connector which can be snapped on and off. Just be very careful if using this style of connector. Adding a tag noting that a 9 volt battery should never be connected is a very good idea. A 3 x AA battery holder is the typical recommended replacement. If only a 4 x AA battery holder is a available, a diode or jumper can be soldered in the first position. A diode is the preferred approach, as this will prevent the batteries from being charged and 'cooked' by the game if blocking diode D17 on the MPU board fails. Keep in mind that an added secondary diode to this circuit will decrease the voltage passing to the 5101, if D17 is still good. Locate a 1N4001 or 1N4004 diode in the position closest to the last + terminal (where the Red Wire exits). The banded side of the diode must be pointing in the direction of current flow, which is towards the Red wire and towards the (+) terminal marking on the MPU board.

The game takes 3 x AA Batteries, so it uses at least 4.5v total. However, it will still store settings down to about 3v or so. After that an AA battery's life will taper off very quickly. A fresh set of batteries should last for at least a year, or the is some other issue. If the 5101 CMOS RAM is socketed, it's possible that it isn't the low power version. Or, possibly the D17 diode has shorted.

The images below illustrate how to install the wiring for battery packs on the different MPU Boards.


Williams System 4 MPU Board


On the System 4 MPU, solder the battery cables: Ground (Black Wire) to the Bottom Left pad and Positive (Red Wire) to the Top Right.

Williams System 6 MPU Board


On the System 6(a) MPU, solder the battery cables: Ground (Black Wire) to the Top Right pad and Positive (Red Wire) to the Bottom Left.

Williams System 7 MPU Board


On the System 7 MPU, solder the battery cables: Ground (Black Wire) to the Bottom Left pad and Positive (Red Wire) to the Top Right.

After adding a remote battery pack, and while the board is still out of the game, it is a good practice to measure the battery pack's voltage at the (+) and (-) pads of the MPU board. All battery packs are pretty cheaply made, and failures "out of the box" can be somewhat common. Checking to make certain the battery pack is functioning before reinstalling the MPU board in the game will save some headaches.

Another good practice, while the MPU board is already out, is to check the D17 blocking diode. An open blocking diode will not allow the battery pack voltage to pass through to the 5101 non-volatile memory, and the newly installed battery pack will be ineffective. Conversely, a shorted blocking diode will allow the board's +5vdc logic power bus to pass through to the battery pack. This in turn, will charge the batteries, while the game is turned on. If Alkaline batteries are charged, they will be damaged and fail, leak, or (in extreme cases) explode. Testing the D17 diode is quick and easy, and worth the trouble checking it out. When in doubt, replace the D17 diode with a 1N4148, or add a secondary 1N4004 to the battery pack. Once again, if a secondary diode is added, it will decrease the voltage passing to the 5101, if D17 is still good.

Alkaline Damaged Williams Driver Board


It is important to remotely relocate the batteries from the MPU board, unless using a lithium replacement. The potential is there for not only the MPU to become damaged, but other associated boards in the backbox. The pic to the left is a prime example of extended damage. The damage occurred in this instance, because the shuffle bowler this board came from was stored with the batteries still on the MPU board for an extended period of time (roughly 5-8 years). It's a costly shame, because this was an otherwise decent, clean, working driver board. The amount of repairs which have to be performed now are going to be extensive.

5.2.3 Installing a Lithium Coin Cell Battery Backup

An alternate solution to a remotely located battery pack is to use a single CR2032 battery and holder mounted on the MPU. Lithium button batteries do leak, although not nearly as commonly or as severe as alkaline batteries. And, since the battery will be contained in a holder mounted above the PCB, there is little risk of corrosion attacking the board.

The holder and replacement batteries are inexpensive.
The following photo gallery shows the installation procedure.
This method can be adapted to work with all Williams System 3-7 MPU board.

  • Williams System 4 MPU Coin Cell Installation
  • Small hole drilled at intersection of holder pin "arc" and trace on board solder side.
  • Battery holder installed, front side. Note the "zero ohm jumper" soldered between the top left and top right battery holder holes.
  • Battery holder installed, rear side. Note the negative connection marked lower right and the path of the positive connection.


Williams System 6 MPU Coin Cell Installation


Coin cell installation on a System 6 MPU board is slightly different physically but exactly the same electrically.

5.2.4 Installing a Memory Capacitor (SuperCap) Instead of Batteries

Another alternative is to install a memory capacitor. In essence, a memory capacitor is similar to a rechargeable battery, although, the likelihood of a memory capacitor leaking is minimal compared to a rechargeable battery. When the game is turned on, it is charging the capacitor. When the game is turned off, the memory capacitor slowly loses its charge over time. Therefore, it is imperative that the game periodically be turned on to allow the capacitor to charge up to its full capacity again. If a game will not be turned on for long lengths of time, a memory capacitor may not be the best solution.

5.2.4.1 System 3/4

5.5V 1F Memory Capacitor Added to a System 3 MPU Board.


When installing a SuperCap, the polarity of the cap needs to match the battery pad polarity. Connect the negative lead of the cap to ground. In the picture at left, the old top center battery holder connection was leveraged as an attach point. A jumper (not shown) on the back of the board connects from there to the edge of board ground.

Connect the positive lead to the the upper right battery holder connection through hole.

Remove D18 and replace with a 100ohm 1/4W resistor.

Once the capacitor is installed, and the board installed back in the game, leave the game powered on to allow the game to charge the capacitor fully. After that, turning the game on monthly for about 10 minutes allows the capacitor to recharge.


5.2.4.2 System 6

5.5V 1.5F Memory Capacitor Added to a System 6 MPU Board. Labels added to help locate parts.


When installing a SuperCap, the polarity of the cap needs to match the battery pad polarity. Connect the negative lead of the cap to ground. Connect the positive lead to the banded side of diode D18 (1N5817). The picture to the left shows a jumper added to connect the negative lead of the cap to ground on a System 6 MPU board.

As an alternative to running the positive lead of the cap to diode D18, the positive cap lead is installed in the positive battery terminal. The diode at D17 (1N4148) is removed, and a 100ohm resistor (or similar value) replaces the diode. Adding a resistor in place of D17 allows the game to charge the capacitor. Once the capacitor is installed, and the board installed back in the game, leave the game powered on to allow the game to charge the capacitor fully. After that, turning the game on monthly for about 10 minutes allows the capacitor to recharge. If the cap is losing its charge rapidly, diode D18 may be faulty, and should be replaced.

5.2.4.3 System 7

Follow these steps to add a SuperCap to a System 7 MPU.

  1. Drill a 1/16" hole NorthWest of the upper right positive battery terminal solder pad. The distance from the positive battery terminal hole should be the width of the SuperCap leads. The hole will be pierce a large ground trace on the solder side of the board (right hand image).
  2. Solder the SuperCap positive lead through the battery terminal solder pad.
  3. Solder the SuperCap negative lead to the ground trace on the solder side of the board.
  4. Remove diode D17, a 1N4148.
  5. Install a 100ohm (or thereabouts) 1/4W resistor in place of D17. Polarity is not important as resistors have no polarity.

The following images demonstrate the procedure.

  • Williams System 7 MPU SuperCap Installation
  • 1/16" hole drilled for the negative "leg" of the SuperCap.
  • The SuperCap and 100ohm resistor installed.
  • The solder side of the board showing SuperCap connections.


5.2.5 Installing NVRAM

NVRAM (non-volatile RAM) can be used in all Williams System 3 through 7 MPUs. By installing NVRAM, batteries will never, ever, be required to retain game settings and high scores. The possibility of spewing alkaline onto the board is eliminated.

For System 3-7 games, the 5101 RAM chip is always at IC19, although the actual physical position of IC19 on the MPU differs between the systems. Unfortunately, the 5101 RAM is always soldered directly onto the board. It must be desoldered and replaced with a 22-pin 0.3" width socket to accommodate the NVRAM.

NVRAM are discussed here.

5.3 Improving the Early System 3 Flipper Response

For the early solid state games, Williams re-used the flipper mechanism from their electromechanical games. This design, composed of multiple pieces screwed to the bottom of the playfield, is functional but somewhat clunky in operation. There are those that advocate scrapping the mechanism and replacing it with modern Williams baseplate design flippers. While this certainly works, and provides a much nicer flipper response, it also costs money for new parts when the old parts can be made to work reasonably well.

First, rebuild the flipper assembly with fresh parts. This includes the coil stops, plunger and links, coil sleeves, EOS switches, cabinet switches, and pawls (if needed). Pinball Resource has all these parts.

This leaves a functional, reasonably strong, but 'clunky' flipper. It will work, but it will not seem as 'snappy' and 'fast' as a modern game. While the modern (WPC) baseplate style flipper mechanism is nice, the most important change is actually not the baseplate, it is the modern flipper bat itself.

There is a slight difference in mass between the original bat at 1.5 oz and the new bat at a little under 1 oz, but what accounts for the difference is the way the mass is distributed. On the old flipper design, a heavy metal base is welded to the shaft, and the plastic bat is screwed to the base. On the new design, the plastic bat is simply molded to the shaft, eliminating the base entirely.

  • Williams Flipper Bat Comparison
  • Williams System 3 flipper bat assembly (two-piece)
  • Williams molded flipper bat assembly

When the flipper coil is energized, it begins accelerating the plunger, pulling it in to the centre of the coil. This acceleration is transmitted to the shaft via the crank, accelerating the flipper bat. In this rotational acceleration of the bat, the further the mass is from the centre of the shaft, the more work required to accelerate it, and the slower the rate of acceleration. (Incidentally, this is why magnesium wheels first became popular in drag racing; they are lighter than steel wheels and increase the car's rate of acceleration. It does not change the car's top speed, but it does reduce the time it takes the car to reach top speed, making it quicker.)

At the end of the coil's power stroke, the plunger slams in to the coil stop, rapidly decelerating. In turn, then, the link, crank, and flipper bat also slam in to the previous part and rapidly decelerate. In this part of the action, the higher mass of the old flipper bat has more momentum, and due to the location of that mass, has more leverage to apply the force of the momentum. This leads to a hard clunk as all of that mass slams to a stop, and often results in a flipper bat that 'bounces' at the end of stroke.

The solution to both of these is to reduce the mass of the flipper bat, especially reducing it as far from the centre line of the flipper shaft as possible. Looking at the options, there are only three pieces to the flipper bat. The plastic bat is not really easily modified to reduce any serious mass. The screw that holds the plastic bat to the base is not easy to reduce in mass, and being located close to the centre line of the flipper shaft, it has very little effect on the operation. That leaves the flipper shaft and its welded on base.

The shaft itself has significant mass, but the mass of the shaft is located very close to the centre line, so it can be ignored as well. That, then, leaves the base. Here there is quite a lot of mass, and much of it is located a good distance from the centre line of the flipper shaft.

First remove the plastic shoe from the bat (there is a single screw underneath the bat). Then screw the metal bat to a piece of scrap wood, and drill the metal bat baseplate to remove metal. This is easiest to do neatly with a drill press. Drill two 5/16" hole near the shaft, and progressively smaller holes toward the tip. A total of seven holes can be drilled: 5/16", 5/16", 1/4", 7/32", 3/16", 5/32", 1/8". These holes remove about 1/2 oz from the flipper bat assembly without significantly affecting its strength or durability.

  • System 3 Flipper Bat Modification
  • Drilling the flipper bat shoe
  • Flipper bat shoe after being drilled

Reassemble the flipper bat, install, and enjoy. The flippers should feel snappier.

6 Problems and Fixes

6.1 Cross Connecting Black and White Backbox Connectors

The infamous System 7/9 connectors that can be cross connected. Yikes!
These two FirePower connectors are clearly misconnected as can be seen by different wire colors entering and leaving the connector pair. The connector colors should not be trusted. Image courtesy of Kevin Scholbe.


In one of the biggest mistakes in pinball manufacturing, Williams used two identical 24 pin .062 connectors to connect cabinet and backbox wiring. The connectors are identical with the exception that one is black and one is white. During manufacturing, the wrong color connector was sometimes used. Bottom line: do not simply mate black to black and white to white. Instead, ensure that the wire colors going into one side of the connector are the same as the wire colors coming out the other side. This is the only foolproof method of avoiding the catastrophe of cross connecting circuits, causing certain damage to the MPU and/or driver board.

6.2 Advanced Diagnostic Tool by Pincoder

An advanced diagnostic tool with full (downloadable) documentation and active support is available here: Pincoder Adapter. The Adapter simply plugs into the MPU board right out of the box and can be used in-game for all tests.

6.3 Built In Diagnostics

Diagnostic switch location on a WMS System 3 MPU
The internals of the diagnostic switches. These are single pole, double throw switches spring loaded to close one side of the switch

The built in diagnostics of the Williams System 3 - 7 board set are extremely limited. The board set either powers up or it doesn't.

There are two diagnostic LEDs on the MPU (system 7 MPUs feature a 7 segment display). However, the usefulness of these LEDs is at best limited and generally just confusing. A properly booting System 3-6 game will flash both LEDs once, and then the LEDs will remain off. This flash is a bit longer than a flicker, staying on for about 1/2 second. A properly booting System 7 game will flash a zero on the 7 segment display. The 7 segment display will then go blank.

Pressing the Diagnostic button on a properly functioning System 3 - 6 MPU will cause both LEDs to flash twice in unison and then go off. If the bottom LED remains on, the game software is indicating a ROM failure. If the top LED remains on, the game software is indicating a RAM failure (6810). And if both LEDs remain on, the game software is indicating a CMOS RAM failure (5101). However, the accuracy of this test is questionable since the game software must be functioning properly for the test report to be valid and the game software requires ROM and RAM to function in the first place. The short advice is: don't rely on this diagnostic test result.

System 3 games require the diagnostic button (lower button) to be pressed to start diagnostics. Once that button is pressed, navigating through the diagnostic menu is identical to later game systems (see below).

Pressing the Diagnostic button on a properly functioning System 7 MPU will cause the 7 segment display to flash a zero and then go blank. System 7 diagnostics are a bit more useful than prior Williams game systems but are still not 100% trustworthy. If the MPU built-in-test software detects a problem with the game hardware, the 7 segment display will remain on with an error code displayed. Possible error codes are:

  • 0 - and then goes out...all tests passed
  • 1 - IC13 RAM or IC20 ROM failed
  • 2 - IC16 RAM failure
  • 3 - IC17 ROM failure
  • 4 - IC17 ROM failure (duplicated test...yes)
  • 5 - IC20 ROM failure or the coin door was closed or memory protect fault
  • 6 - IC14 Game ROM failure
  • 7 - IC26 Game ROM failure
  • 8 - IC19 (5101) RAM failure or the coin door was closed or memory protect fault
  • 9 - IC19 (5101 again) RAM failure or the coin door was closed or memory protect fault

To test displays, lamps, coils and switches for System 3 and 4 games, follow this procedure first, then proceed to the System 6-7 procedure below .

Entering the familiar diagnostic mode of later game systems is a bit less intuitive with a System 3 or 4 game. To enter diagnostics, press the diagnostic button on the MPU once. This is the lower of the two buttons as pictured at left. Both LEDs should flash and all displays will go blank. At this point, the game is ready to accept new game settings like "balls per play" via the obtuse and ridiculous to use dip switches on the MPU. Skipping over the process that essentially amounts to "memory pokes" (who remembers that), the diagnostics can now be controlled from the front door switches. If the AUTO/MANUAL switch is in the MANUAL position, pressing the ADVANCE button will cause each score display to show all zeros. The next press will show all ones, etc. If the AUTO/MANUAL switch is in the AUTO position, pressing the ADVANCE button will cause the score displays to begin cycling, starting with all zeros, then all ones, etc.

YT.png A video demonstrating System 3 and 4 game diagnostic entry can be found here.

Williams System 3 - 7 Coin Door Switches

For System 6-7 games, follow this procedure to test displays, lamps, coils and switches.

With the AUTO/MANUAL switch in the MANUAL position, press the ADVANCE button. This will cause each score display to show all zeros. Pressing the ADVANCE button again will display all ones, etc. If the AUTO/MANUAL switch is moved to the AUTO position, the score displays will advance automatically through each of the digits (zero through 9).

With the AUTO/MANUAL switch still in the AUTO position, press the ADVANCE button to advance to the next test. The match portion of the credit/match display will show "01" and all playfield lamps will begin to cycle on/off.

Pressing ADVANCE again will display "02" in the match window, and the game will begin an ordered activation of each coil in sequence (including buzzers, flashers, etc).

Pressing ADVANCE again will display "03" in the match window. This is "switch test" mode. If no switches are closed, the credit half of the credit/match window will be blank. Closing a switch will result in the switch number being displayed in the credit window. If more than one switch is closed when switch test begins, the number of each switch will be displayed once in the credit window. After that point, only closing a switch will cause it's number to be displayed. i.e. switch test does not cycle through all closed switches but instead displays ONLY the last switch closure detected (at least in System 3 games).

Pressing the ADVANCE button one more time displays "04" in the match window, and game status information is displayed in sequence, automatically for system 3 and 4 games and by pressing the advance button for system 6 and 7 games.

6.4 Quick Start: Repairing a Williams System 3-7 MPU

The procedure for repairing a Williams MPU is outlined below. More detail can be found later in this document.

  1. Replace the power header at 1J2, which is located on the left side of the MPU. The cornerstone of a well running MPU is clean, consistent power. Tarnished connections, worn connections that have exceeded the mate cycle spec, and cracked header pin solder joints all negatively impact power.
  2. If the OEM chip sockets are still on the board, replace them with quality twin leaf or machine pin sockets. Good chip sockets will ensure rock solid connections between the ICs and the various busses. Always replace Scanbe and white Molex sockets - no exceptions!
    3. The ROM on the right in this picture has obviously been heat stressed.
  3. Examine the 40-pin male connector closely. If it exhibits any of the following characteristics, replace it.
    • Tarnished pins
    • Corroded pins (due to batteries on the MPU leaking corrosive alkaline)
    • Fractured solder joints
    • Solder joint “craters”, i.e. an attempt to “reflow” the solder joint that doesn’t adhere to the pin very well. The solder surrounds the pins, but does not create a solid electrical connection to the pins.
  4. Verify the ROM contents. The OEM ROMs at least 30 years old, and have been subjected to thousands of hours of heat, consequently reducing their reliability. Replacing them with known good ROMs is a better option if possible.
  5. Measure power coming onto the board at the filter capacitor. About 5VDC should be present, with very little AC component (measure AC too).
  6. Test socketed processors and PIAs in a known working board. AMI branded chips seem to experience a much higher failure rate than any other brand and are always suspect.
  7. Use a logic probe to test the processor reset signal at pin 40. The signal should start low at power on, then go high a split second later.
  8. With a schematic of the board handy, “sanity check” each pin on the processor to ensure each leg is doing something that makes sense. i.e. data lines D0-D7 should be pulsing.
  9. Make certain the blanking signal goes high and remains high almost immediately after successful MPU boot. The blanking signal is easily tested at pin 37 of the 40-pin interconnect board.
  10. Install Leon’s test ROM to verify processor, PIA, and memory operation.

6.4.1 Further Recommended Reliability Improvements

  1. Replace all headers. Reflowing solder does not create a solid mechanical or electrical connection.
  2. Replace masked ROMs with new ROMs.
  3. Update the MPU to use fewer sockets. i.e. use a 2716 (or double the image into a 2732) at IC14.
  4. Move the backup battery to a remote location or install NVRAM. Batteries should be removed from the MPU board as soon as possible.

6.4.2 Adding System 3-6-like Diagnostic LEDs to a System 7 MPU

Diagram depicting addition of 2 LEDs and 2 resistors that provide the same function as System 3-6 MPU boards
A System 7 MPU with factory installed diagnostic LEDs


System 7 MPUs replaced the two diagnostic LEDs of past systems with a 7-Segment LED display to provide diagnostic indications. At times, the 7447 driver IC or the 7-segment LED display itself can fail. Or perhaps running system 3-6 software in the board, while retaining the LED diagnostic capability, is the objective.

Rather than replacing the 7-Segment display or 7447, adding System 3-7-like LEDs to the MPU is easy. Procure 2 standard LEDs and two 1/4 watt resistors of any value between 100 ohms and 560 ohms. Incrementally higher ohm resistors will dim the LEDs incrementally, but can still be used.

Add two resistors on the right-most pads as pictured at left.
Add two LEDs to the top two available pads as pictured at left, with the "flat side" (the short leg or anode) oriented toward the 7-segment display. It's as easy as that.

6.5 Quick Start: Repairing a Williams System 3-7 Driver Board

The detailed procedures for repairing a Williams driver board can be found later in this document.

  1. Replace the 40-pin female interconnect. Due to fatigue and/or battery damage (alkaline corrosion), this connector can and will become unreliable over time. Replace it.
  2. Remove and reflow the solder on *all* of the header pin connectors. Another choice is to replace the connectors with new ones.
  3. Replace all suspect chip sockets. PIAs on the driver board were the only chips socketed from the factory on this board. However, the factory used crappy Scanbe sockets for these chips. If Scanbe sockets are present, replace them.
  4. Inspect the PIA chips. If AMI chips were used, suspect eventual, premature failure. Replace if possible.
  5. If the factory 3 watt current limiting resistors in the lamp matrix appear heat damaged, replace them with 27 ohm 5 watt resistors. "Sand" resistors are a good replacement choice.
  6. If heavy heat damage is present in the vicinity of the lamp matrix power resistors, suspect potential failure of the lamp matrix ICs (IC13 & IC14) located just above the resistors.


6.6 Using a PC Power Supply For Bench Testing

Sometimes it is more beneficial to do testing / troubleshooting while the MPU and / or driver boards are outside of the game. It's unfortunate, but the System 3-7 boards were designed in such a way that both the MPU and driver board have to be attached to one another for the MPU to effectively boot and function. Since the driver board has no intelligence located on it, it needs to be connected to the MPU to be of any use.

WMS System 6 MPU and Driver Board Powered by a PC Power Supply. Power connections are attached to the readily available test points on the MPU board.
WMS System 7 MPU and Driver Board Powered by a PC Power Supply. Power connections are attached to IJ2 on the MPU board.

The voltages needed to power a System 3-7 board pair on the bench are +5v, +12v, and ground. Connections can be made to the appropriate pins on 1J2 (only header connection on left side of MPU board). The connections are as follows:

  • Pins 1-3 (ground)
  • Pins 4- 6 (+5 volts)
  • Pin 9 (+12 volts)

If connecting a System 6 or 7 board pair, it is easier to connect to the test points. The connections are as follows:

  • Test point 1 (+12 volts)
  • Test point 9 (+5 volts)
  • Test point 10 (ground)

Note that test points 9 and 10 have different locations between a System 6 and 7 MPU board.

6.7 Connecting a Logic Probe to the Williams System 3-7 MPU

Connecting a logic probe to a System 3-7 MPU. This picture is a System 3 MPU.
An alternate method to connect a logic probe on a System 6/6A MPU.


The easiest way to power the logic probe is across the input power's electrolytic filter cap, as shown on this Williams System 3 board at left. All MPUs, including Systems 3, 4, 6, 6A, and 7, feature this cap.

Since the MPU and driver board logic are electrically connected and powered by the same source, connecting the probe to the MPU allows probing driver board logic also.

6.8 Connectors, connectors, connectors

Cracked solder joints on the header pins of a Williams driver board


With this era of game, most of them 30+ years old now, and they will have problems with connectors. One of the most common issues is cracked solder joints on header pin connections to the PCBs. Some cracks are more obvious than others, while some can only be seen clearly under proper magnification.

Solder removed from the header pins of a Williams driver board


Although it can be somewhat time consuming and a tedious chore, it is best to remove the old solder from a header pin joint, and then add new, fresh solder. Performing this procedure is a necessity on these games to ensure reliability. The other option is removing the header pin connection completely, and replace with a new one.

If the game starts to reset for no reason during play, or when both flippers are engaged, it may be worth first re-pinning the connectors at 1J2 on the MPU Board and at the outputs from the PSU. This is always a good step even if it does not fix the immediate issue, as it avoids power problems later on.

This will involves replacing the header pins on the PSU board and the female connectors on the cables. The pins are .156" and square pins are preferred, although round are still available. Use phosphor bronze. On the female side, crimp pins of the trifuricon type are best.

Second prize goes to reflowing solder to the back of the header pins, and just replacing the female IDC connectors with new ones. In that case rather than a crimper, the .156" 'punch-down' tool must be used for this replacement to work correctly. If the correct tool is not used, the IDC housing and pins will be damaged, especially if a screwdriver or some other makeshift tool is used. After that, the wires just don't seat correctly and will not maintain proper contact. Always cut back to clean wire and punch down correctly to a new IDC connector, reusing an old one rarely works.

Again, the connector is a .156" 9-way connector at 1J2. There is also a 'key' pin where the hole is blanked off on the connectors and the pin cut off flush with the board. This prevents plugging the power in to say the switch matrix or swapping two connectors over. It's easiest to just buy a bag of these for blanking off the 'key' positions, rather than try to re-use old ones.

6.9 Wiring Connector Issues

Socketed and replaced ICs blown by incorrect connector hookup.


Williams made a poor wiring harness design decision with this series of games that allows the cabinet harness to be connected to the head harness (36 pin inline connectors). Mating these connectors incorrectly will short the 28V solenoid circuit to the logic voltage level circuit, destroying ICs on the MPU and perhaps the driver board. Always be sure to double check connector pin counts and colors, but most importantly, wire colors going into and out of both sides of the connector. Connector color is not a 100% reliable indicator as Williams sometimes used white connectors, sometimes used black connectors for both, and sometimes may even have used one black and one white connector. Wire colors should always match up.

A real world example on a System 7 Jungle Lord, pictured at left, resulted in MPU IC7 (7404), IC32 (7400), and IC12 (7408) being blown. The 5101 wasn't destroyed. On the driver board, IC16 (7406) was also damaged. The amount of damage can vary, however, other instances suggest a similar path of destruction. The best course of action should this occur is to spend a lot of time with a DMM in diode test, checking each individual IC. Socketed ICs, like the 5101 and the 6808/2 can be easily tested in a known working system.


6.10 System 7 MPU Board Issues

Diagnostics, and the 'System 7 MPU Board won't boot'.

Forget about getting anything at all on the Player Displays without having working ROMs in place and a good MPU board. Blanking has to go 'high' before the displays will work. So coin door game diagnostics are useless until the MPU board runs correctly.

But help is at hand on the System 7 MPU, Williams added a single 7 segment LED display to the board. To get anything at all on the 7 segment display LED on the MPU, these conditions must be met:

  • The 74LS47 Display Driver IC34 to be working.
  • The 6821 PIA (at IC18) on the MPU responsible for driving the score displays to be working.
  • The logic +5v to be good.

Normally when a properly functioning board boots, the 7 segment LED display will flash "0" briefly and then the display is turned off, and remains off. If within a game, the game would then be in attract mode if everything was perfect. Pressing diagnostics should show a "0" and then return to attract mode with the LED display blank. [ed Note: Correct me if that is wrong]

Once the above-mentioned chips are installed and working and the sockets have been replaced or tested good:

  • With no ROMs installed, a board with a fault or a ROM fault if they are installed, the "top two" LEDs (if there) would be lit and the LED display would show "0". The MPU board is 'locked up', in that condition.
  • Display Driver PIA IC PA4 - PA7 will be high. Pressing the * Diagnostic switch will not change things.

Then an OS is needed, which is the 'Flipper ROMs' in pinball. When they 'Boot' (provide a set of instructions to the CPU)- even with the Game ROM removed, the fist thing they do is to "turn off" the LEDs 1+2 and so the onboard LED display (7447) would then go blank.

In fact, "0" should flash once and go blank. That means the board is not locked and at least booted the OS.

Installing a good Game ROM after that may provide access to the diagnostics if everything is working up to this point. Pressing the diagnostic switch should then provide a (hopefully) valid indication of what component is stopping the (pinball or whatever) from running. Here are the key indications (for System 7, Sys 8&9 may be similar):

These results can appear with error conditions by using the Flipper and Game ROMs, or the 2532 WMS Test ROM in IC17 (on Sys 7 boards): If all the support chips are good, and one "0" flash appears and then goes blank, then the on-board error display can be trusted.

Press the diagnostic button, the numbers are:
System 7 LED Diagnostics
Number Meaning
0 Test Passed (only if display goes blank)
1 IC13 RAM Faulty
2 IC16 RAM Faulty
3 IC17 ROM 2 Faulty
4 IC17 ROM 2 Faulty
5 IC20 ROM I Faulty
6 IC14 Game ROM 1 Faulty
7 IC26 Game ROM 0 Faulty
8 IC19 CMOS RAM or memory protect circuit faulty
9 Coin-door closed, memory protect circuit faulty, or IC19 CMOS RAM Faulty.


Getting a 0 and remaining on could potentially be that the CPU / board has not successfully reset. In other words, no instructions have been passed to and from the CPU chip. This problem could be due to a faulty reset circuit. The lack of +12v on the MPU board can cause this too.

Getting an 8 or 9 is *good* indication - it means the board is almost there!

Here are some further tips about those conditions:

8 - MPU board may be good. Is it looking for the Driver Board? (Like the missing 7th flash on a Bally MPU).

Make sure the Driver Board is connected and the "40-way" interconnect has continuity on all pins. From System 8 on there is no interconnect... so by System 9 MPU and Driverboard (and Sound!) are all one PCB.

Then suspect that IC19 RAM is faulty or finally a memory protect fault.

9 - First check for coin door closed (or pin 1J4-1 or 1J3-1 is being grounded), then is IC7 faulty? Finally is IC19 RAM faulty?

If a "9" is displayed, it may be tempting to install it in a game, open the coin door, and try and boot up. But this is wrong, as there still is a problem on the board. Either the IC19 5101 CMOS Ram is faulty, or another memory protect component is faulty.

Remember to do the "switch on, off and back on again quick" trick to see attract mode appears when reinstalling in a game after taking an MPU out. Many times, that does the trick. Also, if removing the Driver Board and reinserting it on the MPU board 'fixes' the problems and the boards will then boot correctly, then it points to the "40-way" connector needing to be replaced. Both header pins on the MPU and sockets on the Driver Board.

Parts:

The 7-Segment display can be replaced with a KINGBRIGHT SA03-12HDB LED 0.3" RED DISPLAY.
5101 CMOS RAM. 5101-1 The low power version is needed, as it needs battery backup to hold RAM contents when the game is powered off.
A 6821 PIA is a standard part. MC6821 MC68B21 are common. The 6820 designation is much earlier but also works. Other designations:

xx6821, xx68A21, xx68B21, where xx can be MC (Motorola) or HD (Hitachi Data) will all work. The 'A' in 68A21 means the PIA will run at up to a 1.5Mhz Clock , the 'B' in 68B21 means up to a 2Mhz clock, without any letter is up to 1MHz. Any final letter is Package type (P,S or L) P=Plastic S=Cerdip and L=Ceramic. Any of the above speeds or package types will work for this era of pinball, so the letters can be ignored and any 6820 or 6821 can be used.

6.11 Chip Sockets

Reliable socket connections are a requirement for any printed circuit board to work as designed. Old sockets, as discussed below, should be replaced. Use extreme care in desoldering the old sockets, the traces and pads on circuit boards are easily lifted, especially if any battery corrosion is in the area.

The chip sockets on old Williams boards (also most any board of this era) are long past their reliable lifetimes. They may work, but they may also cause intermittent connections that will have you chasing your tail tracking down odd problems with the game. Like the 40-pin interconnect used in Williams System 3-7 games, these sockets should always be replaced.

Perhaps the most maligned socket brand, and rightfully so, is the Scanbe brand. In the picture below, it can be seen why. These 30+ year old sockets passed the point of reliability many years ago. Included in the Scanbe socket picture below are two pins pulled from a Scanbe socket. The pins were designed to grip the SIDES of the IC legs, unlike the design of modern sockets that grip the front and back faces of the IC leg. Get rid of them now.

Later, WMS switched to sockets made by Augat or Raychem (all Tyco now). They occasionally go bad, but they can usually be left alone. But there is another type made by Molex which should also be replaced if it is found on a PCB. When looking at the picture below, it can be seen why. Curvy isn't it? And the contacts break easily too. Not as bad as the Scanbe but not much better either.

  • Poor quality and old sockets found on many early pinball machine circuit boards
  • It's hard to believe that these sockets ever worked.
  • A Scanbe socket being removed from circuit board by gently prying it up.
  • Closeup of the infamous "Scanbe" brand socket. "Scanbe" is molded into the top of the socket face.
  • Closeup of "Scanbe" socket pins. Surface contact area was small from the factory.
  • Another bad socket type usually found in system 7 pcbs. This one was made by Molex.


6.12 Voltages

The MPU board needs two voltages to boot: +12v which only the MPU uses for the reset circuit, and +5v (logic power) for the MPU and Driverboard IC chips.

If the game goes dead, but the sound board continues to work then there may be a fault with the PSU, power connections to the MPU, blanking or the reset circuit.

Check for sounds by pressing the sound board diagnostic test. The game will need to bed switched off & back on to cancel the looping sound test. The sound board has its own PSU. It takes 18v AC directly from the transformer and makes +12v (reset) +5v ( logic) and also -12v (Speech and mixers). It's probably OK to assume the +5v and +12v going to the MPU (and Driver Board) are OK from the PSU if the sound board boots and runs. It's like a mini MPU, with the same CPU and one PIA chip.

The main PSU board only needs to make the first two, +12v and +5v (no -12v) for the MPU and Driver board. But it also generates DC power for the flippers and the solenoids (+28v DC or so).

6.13 Power Supply Issues

The key thing to remember is that the PSU board is made up of separate output circuits combined on one board. They don't interact, and are in order:

  • +18v for Lamps,
  • +28v Solenoids (and Flippers),
  • Regulated +5v for logic B+ and an unregulated +12v for CPU reset,
  • +/- 100v HV supply for the Displays.

We're not counting the G.I. (General Illumination) circuit as it is AC only and is fused on the fuse card below the PSU. It only passes through the PSU board to be fused and controlled with a relay starting from System 7 games.

Input AC power gets to the PSU board directly from the transformer for the +5v/+12v and the display voltages. However, in the case of the Lamp and Solenoid voltages AC is 'full wave' rectified (to DC) before reaching the PSU board. This is done using two Bridge Rectifiers BR1 for Lamps (to the Right of the backbox) and BR2 for Solenoids and Flippers (to the Left of the backbox).

Lamp Voltage Supplies:

  • The +18v DC Lamp voltage passes through a very large 'blue can' filter capacitor between the BR and the PSU board. This 'smooths out the 'bumps' (called AC ripple ) in the DC waveform.
    • Because the MPU Controlled Lamps (for the inserts) are strobed by the Lamp Matrix, the resultant voltage to the bulbs is "averaged down" to about 6v DC.
      • Lamps are rated at 6.3v and 250mA for #44 Bulbs (so 1/4 Amp) or 150mA for #47 Bulbs (about 1/7 of an Amp).
      • Because they use less power, #47 bulbs will generate less heat.
      • Replace any shiny 'silvered' bulbs and any that say 0.3A as they burn way too hot and will ruin inserts and cause paint to flake on your backglass!
    • Note: This has nothing to do with the G.I. (General Illumination circuit) which is fused on a fuse card and reaches to G.I. bulbs as about 6.3v AC.
      • As they (almost) always stay lit, the G.I. bulbs tend to last longer running on an AC voltage.
      • The AC is 'Tee-d' off to the cabinet, playfield and the insert board (for the backglass lighting).
      • Wire colors for G.I. (and must be measured as AC) are Yellow and Yellow/White.

Flipper and Solenoid Voltage Supplies:

  • The +28v DC Solenoid voltage passes straight through to the PSU where it is fused before leaving on the way to the playfield.
    • There are some additional components (a varistor and some filter capacitors) on the PSU board for the Solenoid/Flipper power.

Connectors on the PSU Board

  • All AC voltages come into the PSU board in the large 12-way connector 3J1, while the black ground wires are on the 6-way 3J2.
    • The fused AC and DC voltages exit on the larger of the 'D' shaped connectors, while the black Lamp and solenoid ground wires exit on the smaller of the 'D' connectors.

+5v Logic and +12v Unregulated Supplies:

  • In the case of the +12v (and +5v a logic power), the AC comes into the PSU board at 3J1-10 11&12 directly from the transformer windings.
    • You can measure about 10v AC in at the bottom of the diodes D7 and D8 with respect to the board's ground.
      • The total AC input is 18.6v, so each diode 'sees' 9.3v because the transformer winding 'center tap' at 3J1-12 is grounded.
      • D7 connects to 3J1-10, and D8 to 3J1-11 and you would trace connections from there back to the transformer.
      • If you measure AC across the bottom of these two diodes you should see the full 18.6v AC.
      • Do not worry if this AC value is higher without a load, it can (and will) vary with changes in line voltages.
      • Values over 20v AC are considered normal without a load. It's a 'schoolboy error' to chase the voltages of unregulated supplies trying to get them to be exact. This includes the AC input and +12v DC unregulated supply.
    • The two large diodes at D7/D8 rectify the voltage to around +12v DC, again this is unregulated.
      • It is filtered by a large capacitor (12,000 uF or higher) to smooth out the AC ripple.
      • It is fused at 4A Slo-Blow, before exiting at 3J6-6 as the +12v DC for the reset circuit on the MPU Board.
  • To generate the regulated +5v logic supply
    • The +12v input is filtered more by capacitor C16 and then passed through X3 which is a +5v voltage regulator.
    • This exits at 3J5-6 as exactly +5v for the Master Display Board ICs and at 3J6-7 to 3J6-10 for the other +5v supplies (MPU and Driver Boards).

Please note that the Sound Board generates its own +12v and regulated +5v supplies for both Sound and Speech Boards. This is discussed in the Sound & Speech Board Repair Section.

6.13.1 General Illumination Connection Design Flaws

Starting with System 7 games, the general illumination passed through the power supply. This was done to turn the general illumination off (for effect) via a newly added relay on the power supply board. Unfortunately, Williams did not spec connections which could properly handle the current used for the general illumination. The end result was connections, whether it be housings or header pins, which became hot and / or burnt. As the connections tarnish, resistance increases, and the connections become hotter. The cycle continues until an inline fuse blows, or in most cases, discontinuity in the circuit occurs.

The only option is to purchase the proper Molex housing and pins and replace both the male and female connections. Replacing one side but not the other is futile. The issue will reoccur at a much faster rate than replacing both sides. Another preventative measure is to reduce the amount of current draw by replacing the lamps with LEDs. LEDs draw less current than 44, 47, and 555 bulbs, and the GI connections become less of an issue in the future when LEDs are installed.

6.13.2 +/-100v Display HV Section of PSU

Here is a document for troubleshooting and repair of the Williams System 3-6 PSU Board Display Power.

An attempt was made to create step by step tests so that reading a schematic is not required. Includes replacement parts for the HV section repairs.

If the 1/4A (0.25A) fuse at F1 always blows on power up, you may have a shorted UDN IC chip on the Master Display Driver board. See this section on Sys 3-6 Master Display Drivers for testing all the UDN chips on the display board. You can use these tests for both UDN7180 and UDN6118 (or UDN6184) IC's.

Try removing the master display power connectors. If that stops the F1 fuse from blowing, suspect the Master Display Driver needs repairs before you rebuild the HV section of the PSU. Test the HV outputs of the PSU without any displays connected and see if the proper voltages are being generated from the PSU without a load. You need +5v logic, +100v and -100v supplies to run the Master Display board correctly. This will let you figure out if you need to rebuild the Master Display, the HV section on the PSU Board or both.

From experience, if components on the PSU Board HV section are stressed, it is best to go ahead and rebuild the entire HV section replacing all components with new parts from a $10 kit. It will save you time!

6.13.3 PSU Parts Suppliers

Check with:
Ed at GPE has HV rebuild kits and most parts.
Rob Anthony's LockWhenLit (Borygard from RGP) may also have some PSU parts.
Bob's PSU parts list shows the Mouser part numbers which is useful if you live in the USA, or just want to order a few items for your PSU repair.

Also check with your favorite electronics supplier. Please be aware that NTE replacement parts may have wide tolerances so they will be compatible with a greater number of original parts. So get the correct replacement parts wherever possible.

6.13.4 PSU Diagrams and Resources

Please refer to these files while using this section to diagnose and repair your PSU Board:

  • Williams 3-6 PSU Logic Diagram. A PSU Board Schematic for System 3-6 games, containing test points with approximate voltages and help with fault finding your game's power.

Other Resources to visit
Except for the possible use of different transistors for Q1 and Q3 the later System 9-11 HV section is very similar in design. So it will be worth reading through the Sys 9-11 Master Display Driver Problems, for more detail before starting your repair.

Missing PSU Upgrade Sections to add

  • Testing the AC in and DC out for all sections, including the Solenoid and Lamp BRs. Simple to do, but should it be spelled out?
  • How to fit the replacement transistors with the 'twisted legs' as the SDS-201 and SDS-202 are obsolete. Although a diagram is on the GPE PDF for Ed's HV kit, some shrink wrap tubing could be added to the 'lowest leg' to reduce failures.
  • How to fit the 2 x 1n4763a 91v zenerDiodes, at locations Z2 & Z4 to reduce the Display out put to about+/-90v under load and extend display life.
  • Repin of Headers at 3J1 and the used pins at 3J2, theory for using crimped trifuricon connectors and new housings which replace the existing female IDC connectors.

...etc.

6.14 The MPU Reset Circuit

The +12v for reset goes into a circuit that waits (for about a 1 sec. delay) for the +5v to stabilize before it lets the CPU boot. If the +12v (or output from the reset circuit) drops, the reset halts the CPU, PIA(s) and the 5101 CMOS RAM chips before the +5v to them dies. A shutdown of the CPU will also pull blanking low and halt solenoids, lamps and displays to protect them from further damage (coils firing, memory glitching) during the power down. So spikes or drops to the +12v line may halt the CPU at the wrong time.

An MPU unable to boot may mean that its reset circuit is faulty. A +12v line comes in from the PSU board to the MPU on pin 9 of 1J2. That's the top pin on the top left (and only connector on the left) of the MPU board. +12v can also be measured at TP1 on a System 6, 6a, or 7 MPU. Don’t get hung up on it being exactly +12v, as it's not regulated. However, it can’t drop too much lower than +11.5v, or the reset circuit may have a problem generating enough voltage to keep reset “up” to the CPU. But +11.5v to +14v is not unusual and will still work. A fluctuating reset can cause the game to reboot during a game.

The state of the reset (high or low) can be observed at pin 40 of the 6808 or 6802 CPU chip. Likewise, the reset can be observed at TP8 on System 6, 6a, and 7 MPU boards. TP8 is located just to the right of IC19 (5101 memory) on a System 6 or 6a board, and the far right test point on the bottom of the MPU just above the 40-way connector on a System 7 MPU board. It will be marked as TP8 on the PCB mask. For System 3, 4, 4a MPU boards, the reset will have to be measured at pin 40 of the 6800 CPU board. When booting the game and watching reset, it stays low for a heartbeat after power on and then rises to +5v. +4.75v or around that is fine. Don't get hung up too much on the voltages being exact for the reset circuit either.

If the +12v appears OK when measured, and the sound board is always booting and working, it's safe to assume that +12v at the PSU is stable.

A good trick to 'inject' a reset, is to try connecting +5v for a second or two to the reset test point at TP8 or the leg of the CPU chip (pin 40). If the CPU boots (reset must be a logic high now on pin 40 of IC1), then the reset section will need to be rebuilt..

So that involves running a jumper lead from TP8 and briefly touching TP9 (+5v). TP9 is located just to the bottom right of the battery holder on a System 6 or 6a MPU board, while it is at the top right the battery holder on a System 7 board. The battery holder should have been removed, so the test point is located in the vicinity of where the original holder used to be.

6.14.1 Repairing the Reset Circuit 'Divide and Conquer'

Assuming that the game boots and works if the CPU reset pin is held high, how can it be determined what is wrong? One solution is the shotgun method: just start changing transistors and components until it is fixed. However, this wastes time and parts. A better way is to narrow down the fault location using the 'injection' trick.

The reset circuit is easily seen as two halves. For a System 7 MPU board, +5v can be applied directly to the top of resistor R12. If the board boots and stays stable, then that indicates the problem is on the 'right hand' side of the circuit. (On System 6, the same thing can be done, but using R32. This method helps narrow down the faulty components by a half.)

Assuming that the +12v getting to MPU board is good, if the above injection of 5v doesn't work, measure where the voltage disappears on the left hand side of the reset circuit. Then replace parts on that side, checking for a good reset when the board boots each time. Remember that for System 6 and 7, there are just 12 active components that make up the reset circuit, and 8 of those are transistors. Those are usually the source of the problem.

6.14.2 Step One

The Right Hand Side. Start by replacing transistors Q6 – Q9 (all 2n4401 NPN) get 10 or more of them, they’re cheap to buy and they are commonly used. They are used as pre-drivers for the solenoid power transistors on the driver boards for System 3-7).

Test the board for the previous booting problem. If they are still there, replace the two zener diodes: ZR1 is a 6.8v Zener (1n5996) and ZR2 is a 3.9v Zener (1n5990) - you can't replace these components with anything else to my knowledge. These are critical values. And put the orientation of the band of the diode the same way as the ones that are removed. Taking a photo of how the circuit in the top left of the board looked before making any changes is a good idea. Or, only replace components one at a time. A test boot could be done on the bench each time something is replaced, but that may slow things down. On the other hand, this may result in avoiding replacing more components than necessary.

It should be booting now, if not attack the other half.

6.14.3 Step Two

The Left Hand Side. Start by replacing transistor Q2. Q2 is a 2n4403 PNP transistor, cheap as well but you won’t need as many for pinball repairs. then replace Q3-Q5 with the 2n4401 's like before. I would replace D19 (a 1n4148 fast switching diode) you could test it with the diode setting of your DMM, but I don't trust testing diodes in circuit, and if I’m soldering in that section, I just cut it out and replace it. Cost will be pennies, and most pinball people should have new ones on hand. Again, be sure and notice the direction of the band on the old diode, and put the new one back the same way. Same thing goes for the orientation of the flat side of the transistors. Don’t get them wrong when replacing. The last component to replace would be the capacitor at C27 (10 uF @10v is the original part –look on the side of the cap). Replace it with a tantalum 10 uF @ 16v, if you can find it. You can go up in voltage even to 20v or more, but don’t change the value of the capacitance. Note that some caps have markings showing a (-) side, usually running down the side of the capacitor. If it shows a polarity, the (-) side of C27 points down on the PCB, towards the 40-way interconnect.

6.14.4 Step Three, Hopefully Reset is Back Up

That’s it. Usually fixed at step one, and down to one (or more) of those transistors. I have seen a zener or the 1n4148 diode at D19 be the problem before. But for me, the cap C27 has always been good. If all else fails, test the capacitor for a dead short with the power off and replace it with a 10uF electrolytic if it is shorted to ground or you aren't sure it's working.

6.14.5 An Alternative to Rebuilding the Reset Circuit (System 6 and 6a)

Williams System 6 MPU board stock reset circuit
Williams System 6a MPU board reset circuit with an MCP130-460GI/TO reset generator installed. Note that ZR1 has not yet been removed in this image


There is an alternative to the stock reset circuit on System 6 and 6a MPU boards. A reset generator can be installed. To do so, some components will need to be removed and two jumpers installed. The following work will have to be done.

  1. Remove the following transistors - Q4 and Q6
  2. Remove the following resistors - (from left to right as installed on the board) R43, R38, R39, R41, and R42
  3. Remove zener diode ZR1. This will keep any of the existing components in the reset section from being powered by the unregulated 12 volts.
  4. Add reset generator MCP130-460GI/TO where Q4 was (green highlighted area in image). Orient the reset generator in the same manner Q4 was oriented (pin 1 of reset generator will install where the "E" emitter was for Q4). In other words, the "flat" side of the reset generator will face to the left.
  5. Add a jumper between the top pad and bottom pad of where R39 was (yellow highlighted area in pic). Essentially, replace R39 with a zero ohm jumper.
  6. Add a jumper between the top pad of R42 and the "C" collector pad (top pad) where Q6 was installed (orange highlighted area in pic).


This a very simple, and in some cases, cheaper alternative to repairing the existing reset circuit. After this modification is performed, the MPU board will no longer need the +12v to successfully boot. Likewise, any of the other components in the reset circuit can be removed if desired. Please consult the System 6 MPU board BOM and schematics to determine exactly what components can be removed. The MCP120-460GI/TO reset generator can be purchased from Great Plains Electronics.

Williams System 6A MPU board reset circuit with an MCP120-460GI/TO reset generator installed with a pull-up resistor


Alternately, an MCP120-460GI/TO reset generator can be used, except a 4.7K 1/4 watt pull-up resistor will have to be added to the reset signal line.

  1. Remove the following transistors - Q4 and Q6
  2. Remove the following resistors - (from left to right as installed on the board) R43, R38, R39, R41, and R42
  3. Remove zener diode ZR1. This will keep any of the existing components in the reset section from being powered by the unregulated 12 volts.
  4. Add reset generator MCP120-460GI/TO where Q4 was. Orient the reset generator in the same manner Q4 was oriented (pin 1 of reset generator will install where the "E" emitter was for Q4). In other words, the "flat" side of the reset generator will face to the left.
  5. Add a jumper between the bottom pad of where R38 and R39 formerly were.
  6. Add a 47k 1/4w resistor between the top pad of where R38 was and the bottom pad of where R41 was.


6.14.6 An Alternative to Rebuilding the Reset Circuit (System 7)

Components to remove from a Williams System 7 MPU board reset circuit
Williams System 7 MPU board reset circuit with a reset generator installed


Like the System 6 and 6a MPU boards, a reset generator can be installed on a System 7 board. Some components will need to be removed, a single jumper installed, along with a 4.7k ohm resistor installed.

Procedure

  1. Remove Q2 transistor
  2. Remove the following resistors - (from left to right as installed on the board) R27, R24, R31, R25, R29, and R30 (resistor located to the right of Q2)
  3. Remove zener diode ZR1. This will keep any of the existing components in the reset section from being powered by the unregulated 12 volts.
  4. Add reset generator MCP120-460GI/TO where Q2 was (this is the same reset generator as used on Data East MPUs). Orient the reset generator in the same manner Q2 was oriented (pin 1 of reset generator will install where the "E" emitter was for Q2). In other words, the "flat" side of the reset generator will face to the left.
  5. Add a jumper between the bottom pads where R27 and R24 formerly were.
  6. Add a 4.7K ohm 1/4 watt resistor in place of R25.


All components of the reset section have been scraped except those required to use the MCP120-460GI/TO reset generator (marked with gold Sharpie). In this picture, the 74125 stuffed at the factory has been removed as it's original function was to read the DIP switches. System 7 games do not read the DIP switches.


This a very simple, and in some cases, cheaper alternative to repairing the existing reset circuit. After this modification is performed, the MPU will no longer need the +12VDC to successfully boot. The remaining components in the reset circuit can be removed if desired as pictured at left. The MCP120-460GI/TO reset generator can be purchased from Great Plains Electronics.

Note: if using the MCP130 variety of reset generators, the 4.7K ohm resistor does not need to be installed as they incorporate the resistor internally.

6.14.7 Reset Circuit Summary

Connectors were mentioned above for completeness. While I agree a good 40-way connector is essential for correct operation of this era of games, changing the Bridge Rectifiers in the backbox or rebuilding connectors probably won't help fix the reset circuit .

That because the BR's are only for the Lamps and Solenoids and won't fix a problem with the MPU power (or reset) as that is sourced directly from the PSU. The +5v logic power comes from a regulator on the PSU board, and the +5v for the reset pins of the MPU ICs is created on the MPU from the +12v unregulated supply.

If you are looking here because game resets (reboots to attract mode or locks up) when flipping the flippers, it makes sense to check and rebuild the solenoid power circuit as well. That is the flipper power on games prior to System 7 when 50v flippers and a Flipper Power Board was introduced. You should also read the WPC section, while the games only similar- the approach for resets caused by power fluctuations is sound. It takes you from the 'AC mains' wall plug right through to the power needed on the boards.

Topics such as electrolytic capacitors aging and drying out, testing the Bridge Rectifiers and checking and tweaking the LM323K 5V regulator on the Power Baord will all be the same or very similar.

6.15 The Blanking Circuit

The blanking circuit is used by the MPU to prevent damage to the game when the MPU can not determine that the game is in a valid state. A low blanking signal means that the MPU isn't operating properly and the blanking circuitry is "blanking" (turning off) the lamps and displays. A high blanking signal means that the MPU is operating normally, allowing lamps and displays to function normally.

On the driver board, the blanking signal is logically ANDed, by 7408s, with the enable signal for all 8 of the lamp strobes and the first 16 solenoids. The 6 special solenoid enable signals are not ANDed with the blanking signal.

On the MPU, the blanking signal for the displays leaves the board via connector IJ3 pin 4 (top left connector on board). For the driver board blanking signal, it leaves the MPU board at pin 37 of the 40-pin interconnect.

A special note: the blanking signal will only go high when the MPU and driver board are connected. It will not go high if testing with a stand alone MPU board.

6.15.1 Adding a Blanking Indication LED to the Driver Board

A "blanking" LED added to a Williams System 3-7 Driver Board.


Adding an LED as shown in the image will let you easily monitor the Blanking state while troubleshooting. Note that the "flat side" of the LED faces away from the 40-way connector and is on the same side as the shorter lead. The “current limiting” 150 ohm resistor is soldered to the ground trace.

6.15.2 Adding a Blanking Indication LED to the System 6/6A MPU Board

A blanking signal indication can be added to a System 6 MPU quite easily.


  1. Remove the TP4 test stud. This is the blanking stud.
  2. Install the "non flat side" of the LED into the open hole.
  3. The "flat side" of the LED is oriented up, or toward the top of the board.
  4. Solder a 150 ohm (or so) resistor between the flat side lead and ground trace that has a zero ohm resistor soldered to it.
  5. All done.


6.15.3 Adding a Blanking Indication LED to the System 7 MPU Board

A blanking signal indication can be added to a System 7 MPU can be done in at least two very clean ways.
When the LED is lit, the blanking signal is high, indicating a properly functioning MPU.

6.15.3.1 Option 1

Implementing this modification assumes that the battery holder trace is no longer in use!
Procedure

  1. Remove the Blanking test point stud (numbered "4"). Since we are adding a blanking indication LED, the stud is no longer required.
  2. Install an LED, with the flat side pointing toward the former location of the battery holder and the other leg through the hole created by removal of the blanking TP stud.
  3. Bend the LED lead through the former battery holder hole and solder.
  4. Form a 150 ohm 1/4W resistor between the appropriate battery holder hole and the fat ground trace along the perimeter of the board.
  5. Scrape the solder mask from the board, enabling the resistor to be soldered to the ground trace.
  6. Solder both ends of the resistor.
  7. Add a label to indicate the meaning of the LED; lit means blanking is on and the MPU is working properly; unlit means blanking is off and there is a problem with the MPU.


  • Option 1: Adding a Blanking Signal Indication to a System 7 MPU
  • An LED replaced the Blanking test point stud, flat side pointing toward the battery holder hole.
  • The back of the System 7 MPU, showing connections of the LED, repurposed battery holder trace, and 150 ohm resistor connection.


6.15.3.2 Option 2

This method does not require the battery holder to be removed.
Procedure

  1. Clean the original solder from the lower of the 4 "paired" solder pads and the "via" pictured below.
  2. Install an LED, with the flat side oriented as indicated by the silkscreen
  3. On the solder side of the board, connect a 150 ohm 1/4W resistor to the LED lead on the "flat side" of the LED.
  4. Solder the other end of the resistor to the large ground trace.
  5. Bend the "non flat side" lead of the LED so that it can pass through the via hole and solder securely. This connects the LED to the blanking signal.
  6. On the solder side of the board, sever the trace shown in the image below that runs NorthEast from the LED (trace severed at gold dot using a Dremel tool and ball cutter bit).
  7. Add a label to indicate the meaning of the LED; lit means blanking is on and the MPU is working properly; unlit means blanking is off and there is a problem with the MPU.


  • Option 2: Adding a Blanking Signal Indication to a System 7 MPU
  • An LED added to available solder pads, flat side pointing in the same direction as indicated on the silkscreen.
  • The back of the System 7 MPU, showing connections between the LED, a 150 ohm resistor connected to ground, and the "non-flat-side" lead of the LED connected to the blanking signal.


6.15.4 Blanking Circuit Stuck Low

The blanking circuitry that appears on Williams 3-7 MPU boards remained the same across all generations of Williams MPUs, all the way through the System 11 board set. Data East copied the same blanking circuitry too!

The blanking circuit explained

  1. A properly executing CPU will send data to the display 6821 PIA at IC18 at regular intervals.
  2. This action causes PA2 at pin 4 of the 6821 at IC18 to periodically pulse.
  3. The signal is inverted by the 7404 at IC7, in at pin 5, out at pin 6.
  4. The coupling capacitor at C32 can be ignored. These ceramic caps rarely fail.
  5. The signal is applied to the base (pin 2) of a 2N4403 at Q5 and to pin 8 of the 556 timer at IC23.
  6. The 556 timer is configured as a "missing pulse detector" meaning that if it does not detect the pulse that originated upstream at the 6821 PIA at IC18, the output of the 556 timer at pin 9 (and 13) will go low, which invokes blanking and protection of the system circuitry.

Upon a successful boot of the MPU board, the blanking will go from a low signal to a high signal, and remain high (no pulsing) as long as there are no issues. If the blanking signal does not go high, the system has not and will not boot.

One of the most common issues for a blanking signal to remain low is discontinuity between the MPU board and the driver board. The blanking signal passes via the 40-pin interconnect at pin 37 (4th pin in from the left side). Poor connections due to fatigue or seeping battery damage can cause a break in this connection. It is best to replace both the male and female sides of the 40-pin interconnect to ensure that all is well. Remember, the blanking will not go high, if the MPU is not connected to the driver board. A flaky connection with discontinuity between boards is the same as if they're not connected.

A System 6 MPU board with the upper right battery holder through-hole circled, along with the ground trace and the 74LS02 labeled.


A special note about System 6 & 6A MPU boards. The upper right through-hole for the battery holder passes the ground from the solder side of the board to the IC5 (74LS02) chip. This through-hole can become compromised from battery corrosion or the careless removal of the battery holder. If there is a break between the two sides of the board, ground will not be connected to IC5, and the chip will not function. If this occurs, the blanking signal will never go high, and the board will not boot. Just check for continuity between pin7 of IC24, the top right leg of chip shown and TP10, the bottom post on the right of picture. You may have to add a ‘stitch wire’ soldered to the top & bottom of the board going through the hole indicated.

System 7 MPU board blanking signal trace

A special note about System 7 MPU boards. The trace for the blanking signal is very fragile. If the batteries have leaked on the board, or leached under the 40-pin header pins, make certain that the trace which carries the blanking signal has not been broken. The break most commonly occurs at the junction of the trace and the component side of the through-hole pad for the header pin on the MPU board. A continuity test can be performed to determine if the trace has been compromised. Place one lead of the ohmmeter on TP4 and the other on pin 37 of the 40-pin header. Also, keep the lead on TP4, and place the other on the 12th pad up on the far right of the board. The header pin connector will have to be removed to inspect this junction, should there be a break.

6.15.5 Blanking Circuit Strobing

There is a peculiar instance where the blanking circuit will be strobing. The end result is erratic behavior from the displays and lamps (both the lamps and displays will be very dim if at all visible during attract mode). The prime source of this issue is due to the C31 capacitor (1.0uF 25v axial tantalum) falling out of spec.

6.16 Driver Board Issues

6.16.1 The "Once Dreaded" 40-pin Connector

The driver board mates with the MPU board using 40 x 0.156" extra long header pins on the MPU and female 'bottom entry' Molex sockets on the driver board. Common belief is that these connectors are good for about 25 reliable connects and removals. While there may be disagreement with this number, the fact remains that the connector is good for a limited number of uses before the connection degrades. At that time, the connector must be replaced.

The connectors between System 3-7 MPU and driver boards were/are a continual source of problems for this era of Williams machines. However, the issue was eliminated with the release of System 8, (and the more common System 9 platform), when the MPU board and driver board were merged together into a single board.

If the header pins on the MPU board aren't tarnished, burned, or previously sanded, you may be able to get away with re-flowing the solder on these pins following these steps.

  • Heat up the pad, add a little new solder
  • Remove all the solder you can (the old and new) with a solder sucker
  • Add fresh solder back to get a nice flow around each pin
The 40 pin female connector on the driver board

Then, replace all the female connectors on the driver board. Even if the driver board side pins look good, replace them. Simple cleaning of the pins with a contact cleaner or burnishing them are not viable solutions. The bottom feed pins loose tensile strength over time, which degrades the connection between the MPU and driver board as they age. The picture at left illustrates the 40 pin female bottom feed connectors on the driver board. It should be noted that the connector in the picture is an extreme example of poor condition connections. Missing and alkaline damaged pins, connectors from position 17-40 on the left, are not the norm. Connections from position 1-16 on the far right are more commonly seen.

Installing a new 40 pin female connector on the driver board
A new 40 pin female connector installed on the driver board

If using four 10 pin female connectors, (the originals on the driver board were five 8 pin), here's a little trick to keep them lined up nicely when soldering them. Take the old 8 pin male header pin connectors, which were installed on the MPU board. Install three of the 8 pin headers on the back side of the driver board and into the new female connectors (see picture to the left). If any of the older header pin sections have battery corrosion on them, do not use use them for this. Allow enough space between the connectors so some of the female connector leads can be soldered. Solder all of the exposed leads on the female connector. Once they are soldered in place, remove the male headers, and solder the remaining leads on the female connectors.

Fresh male headers installed on this System 6 MPU. Nice. Note the new headers are 10 pins long while the original headers were 8 pins long.


Parts for the MPU are:

  • Square .156" header pins. These can be found at Great Plains Electronics here. Alternatively, Molex brand longer pins, part number 26-48-1080, can be used. Both of these options meet the length requirement for the pins to be at least 0.629" or 16mm length.

Parts for the Driver board are:

  • 10-way 0.156" Bottom Entry Female Molex Part Number: 09-52-3102 (4 needed per board). The original part, 5A-9066 8-way .156" Female Molex Part Number: 09-52-3082 (5 needed per board). Alternates are Molex 09-62-6104 and 09-52-3102 which are hard to obtain.

Check with: Ed at GPE or your favorite supplier for these.

Female housings being used to align the male headers.


A trick similar to the method aligning the female housings can be used to align the male headers. Use the old (or new one housings) to align the male headers as pictured at left.

There is no substitute for doing this correctly, other than buying reproduction boards. A solution is a board made by Rottendog Amusements, where the MPU and driver board are combined into one board. Cost is anywhere from $350 - $400. Another option is purchasing both the MPU and driver board from Kohout Enterprises for $425. When both boards are purchased, they can be tied together via an auxiliary ribbon cable. Of course they can be connected via the 40-pin interconnect too.

If doing it yourself, repairs on the MPU and driver board connectors can cost up to $30 or less in parts (header pins, sockets and solder), along with several hours to replace the connectors and pins correctly. This of course depends on your level of soldering skills and the value you put on your time. It's a fairly long job but it's likely that it only needs to be done once.

6.17 Display Issues

6.17.1 Every-Other Digit Shown On All Displays

Half of the digits, or every-other digit, are missing on all displays. Unintuitively, the root cause of this failure is the lamp matrix 6821 PIA at IC10 on the driver board.


This failure seems to occur quite frequently and perplexes most techs. While the 6821 in the MPU display interface circuitry would seem to be the logical culprit, it is the lamp matrix PIA on the driver board that is the actual villain. The issue shown at left was remedied by replacing PIA III (also known as IC10), the lamp matrix drive PIA, which is the middle PIA on the driver board.

YT.png A YouTube video demonstrating this failure, and the correction, can be found here.

6.17.2 Every-Other Digit Dim On All Displays

Half of the digits, or every-other digit, are dimly displayed. This behavior was caused by a failed 4020 IC25.


This rare failure is caused by the 4020 at IC25 (on a System 7 MPU) failing such that either pin 12, 13, or 14 is "dead". In this case, display interrupts occur too rapidly, and the 6800/6802 can't service those interrupts quickly enough to properly render display data.

YT.png A YouTube video demonstrating this failure, and the correction, can be found here.

6.17.3 Repairing the Master Display Driver

Note that the detail here is for System 3 - 6a with 6-digit displays only. So it does not include Alien Poker or Algar, which both used the System 7 type display setup. System 7-9 used a smaller Master Display Board located on the back of the lamp board, and it had IDC type ribbon cables that ran to the 4 player 7-digit displays. They added a 4-digit Credit/Match display which also used a ribbon cable. The harness connecting the backbox (PSU and MPU) signals to the Master board used the same 'edge-connector' as System 3-6).

6.17.4 First, a High Voltage ***Warning***

With the game on you are dealing with +100 and -100v HV DC (High Voltage) going to the displays. That's a potential 200v difference and you only need to feel that once to know it. Wear tennis shoes (trainers or any rubber soled shoe) when working on displays with the backbox open. This is a shock hazard! If you are not capable or happy with measuring these voltages with a steady hand, then get someone else to help you and be there to 'spot' you. Always a good idea and more fun than working on your own.

Another useful tip is to remove Fuse F1 from the PSU (Power Supply Board). This is the HV fuse and is 1/4A Slo-Blow (250mA 'T' fuse in Europe). This will allow the logic (BCD binary decoders and Buffer ICs) to be troubleshooted with a DMM or Logic Probe without the danger of shocks or shorting them out when measuring TTL (logic) voltages. It may also save the logic probe from a crispy fate.

6.17.5 Master Display Wiring Harness

Sometimes displays just go blank and it is a fault in the wiring harness from backbox to the Master Display Driver. If you can swap a working Master Display Driver board from a System 3-6a game with stable displays, then do that as a first step. This way you quickly determine if the fault is in the backbox OR the Master Display Driver PCB or displays on the front of the light board. If it's the towards the backbox then suspect display harness cabing Driver to MPU, or the display PIA or decoding on the MPU. If you don't have the ability to swap in a spare Master Display board, then turn off the game and test the harness for continuity between the backbox (MPU board) and the cables at the left and right side of the Master using this Master Display board harness wiring diagram.

6.17.6 Segment failures

Segment failures on a single display could be that display failing. Same issue if just one display is blank. Swap that display with another (power off first) and see if the problem moves with that display. If it moves, it's that display glass or its board. If it doesn't move and stays at the same player location, see below for the possible suspects.

Segment failures on multiple displays and probable cause:

Players 1 & 2 Segments and master display Segments are out:

  • BCD inputs are A1 B1 C1 D1 from the MPU board
  • MC14543 BCD to 7-Seg. Decoder IC5
  • UDN7180 Display Driver IC9
  • Resistors R1-R7

Players 3 & 4 Segments

  • BCD inputs are A2 B2 C2 D2 from the MPU board
  • MC14543 BCD to 7-Seg. Decoder IC8
  • UDN7180 Display Driver IC10
  • Resistors R8-R14

Start out by measuring all the resistors on the master display board with the power off. R1-R14 should be around 10K ohms. Any that are not within about 10% (say a range between 9.6K to 10.4K ohms) need to be replaced. Also look and see if any look "toasted".

Those resistors get cooked on the Master display boards and usually will then cause single segments to fail on both the player 1&2 or 3&4 displays together, as they are linked. Te resistors need to be replaced with the same value: 10K but at 1/2 Watt. Some of the modern 'metal film' resistors are rated at 0.6W which is perfect and they fit nicely in that location. Older 0.5W (1/2W) resistors are larger, but are fine. Mount them slightly off the PCB so they get good airflow all around the resistor body.

Do the above steps anyway, no matter what the display problem is although it's probably not going to be the whole story. Replacement resistors are cheap and will prolong the life of your displays. Reducing the voltage going to the displays when rebuilding the HV section of the PSU is a good step to take and will help as well. You only need to replace two Zener diodes to achieve this.

6.17.7 Digit Failures

Strobe inputs for the display digits come in to three 14069 hex inverters on the driver board. IC1, IC2 and IC3.

  • IC1 is dedicated to the Master Display 'Credit/Match'. Strobes are: 15,16,7,8 left to right, for the 4 digits it uses.
  • IC2 and IC3 strobe inputs are more complicated, to strobe distribution between these IC chips is shown in the display diagram below.

Strobe (or strobe input) failures are likely to show up in the player display pairings 1&3 or 2&4 because:

  • Strobes 1-6 are shared for players 1 & 3
  • Strobes 9-14 are shared for players 2 & 4

6.17.8 All Displays are Blank

This could be that the +100v or the -100v HV are missing from the PSU board. Both have to be there, so check that the output from the PSU and that the voltages (HV) are good first, and getting to the Master display driver to be relayed to the player displays.

There are also 5 x 3 Mega ohm resistors, at R15-R19, These are for the cathode "keep alives" and again should be near that value. If the 3 Meg resistors don't look cooked and are within spec, check that you can see an "orange glow" in the displays when the lights in the room are dim or off. If you see a faint glow (some describe it as an 'orange neon dot'), then look elsewhere for the fault.

If you don't see the display glow, check the wiring to the connectors carefully looking for a burnt wire at pins 4J7-1 -2 and -6 on the master display. Do this with the power off, as you are dealing with 100v and -100v DC. If you find a cooked wire, sometimes just cutting the wire back a bit to expose clean metal and then reinserting it firmly in the IDC connector will repair the problem with a bad connection for one of the HV lines. Remove and reseat all the edge connectors on the master display board and especially examine the ones that go to the backbox. You can clean the copper contacts on the edge connectors gently with an eraser to shine up the copper if it's dull or has "dead spots" worn on it. Check any inline connectors as well.

I also recommend disconnecting all the player displays 1-4 at the Master display. Get it working with the just the credit/match and then with one other display attached, like player 1. Then add back the player 2-4 displays one at a time (you need to power off each time you add or swap a display), testing for correct function each time. You can also swap the player displays as a diagnostic step and carefully observe if the fault(s) stay with the Display in question or move to the new location. Use the "display test" on the diagnostics for this. After the above is checked, perhaps you do have to suspect the IC chips (or for the discrete version the transistors). Depending on whether the Segments or Digits are out, it will point you at a specific IC (or transistor array). Knowing which displays are out helps reduce the fault domain down to one chip. You need be like a detective, following the clues. Having a Master Display Assembly drawing and a schematic on hand will help with this process.

Here's hoping that it isn't one of the UDN6118A-1 (aka UDN6184-5) IC chips that's faulty. Hard to remove from another board in one piece, without a de-soldering station and becoming obsolete.

Also, check that the display blanking connector is in place. This will be a 4 pin connector with a single wire. If this connection has failed or has been unplugged, the game will boot, go into attract and seem to be working, but the displays will remain blank.

6.17.9 Displays only showing even number

This is caused by a display signal connector issue. This shows up in pairs; player one and two or player three and four. Likely to be header pins on MPU board connector J5, J6, J7

6.17.10 Scrolling Digits on Display

The rather unusual issue of "scrolling digits" on the displays is caused by failed/failing filter capacitors on the power supply board. Over about 40 years of operation, these capacitors begin to fail and no longer effectively filter the high voltages. The OEM capacitors were rated at 100uf/150V. These capacitors can be replaced with 220uf/160V axial capacitors to correct this behavior.

A video of this behavior can be found here.

6.17.11 System 3-6 Displays

6.17.11.1 System 3-6 Master Display Driver Boards

Master Display Driver Boards came in two versions, discrete and IC based. Williams designed the discrete version when the UDN7180 / UDN6184 Gas Plasma Display Driver ICs became scarce (and they were expensive even then). There are some logic chips that decode the BCD data for the 7 display segments (MC15453 or MC15458) and to buffer and invert the display strobe lines for the digits (MC14069). These buffers and decoders are common to both the D8000 and D8168 versions.

6.17.11.1.1 Discrete version Display Driver
Williams System 3-6 Master Display Board (Discrete Components)


The D8168 board uses MPSA92 (PNP in a TO-92 package) and MPSA42 (NPN also TO-92) transistors to drive the Master (Credit/ Match) and player scoring displays. These gas plasma display driver transistors are available today and still very cheap to buy. Unfortunately, this style of board is less common to find in the wild, but actually easier to maintain. At least no one can try to charge you $20+ for an obsolete driver IC.


6.17.11.1.2 IC version Display Driver
Williams System 3-6 Master Display Board



The D8000 IC version board is composed of ICs that implement the transistor arrays. Usual factory chips are the UDN-7180 for the Segment Driver and the UDN6118a-1 for the Digit Driver. You may find the Digit Driver IC is a UDN6184-5 as well from Williams.

The UDN-7180 IC chips are still fairly easy to find. Cost should be reasonable, if you shop / ask around. (Sprague) or uPA6118c may still be available for $3-5 but have a lower breakdown voltage (~85v DC) than original part. The displays run at +/-100v DC as standard and need the higher rated original part which is the the UDN6118A-1 (note we have -1 at the end). The UDN6118a-1 is rare and also becoming very expensive, as much as $20+ for an IC chip.

If you do find the lower rated UDN6118a (or uPA6118c), then you will need to replace two zener diodes on the PSU to lower the display voltage down from 100v. It's a good idea anyway on a working System 3-7 game. Not overdriving display glass and less stress on the expensive display driver chips will lengthen the life of your game. Arcades sometimes had bright lights or windows, while in your home you won't notice any difference running at a lower voltage.

I usually drop the voltage down leaving the Power Supply to about +/- 90v with 2 x 1n4763a 91v zeners, at locations Z2 & Z4. This seems to work well, but you are still probably driving near the limits of the replacement UDN6118a. Unlike the Bally games of the same era, there is no fine adjustment on the voltage output of the HV section of the PSU.

See the Power Supply repair section for more information about parts to use and other upgrades.

Remember to compare the costs and trouble of repairs with buying a replacement Master Display Driver board:

  • If all that is wrong with the Master is that the Credit/Match display is out, but it drives all player displays correctly, in my experience the repair is easier than on the player display parts of the driver board:
  • Carefully placing a new 6-digit glass against the pins of the non-working display will confirm if it is the display glass or an IC chip that has failed. Be Careful of the HV (High Voltages) when doing this, and only hold it by the display glass as it is a natural insulator! It takes a steady hand and some practice, but this test will work on other player displays as well. This usually proves the old display has a digit/segment missing, is 'out gassed' or is just prone to flickering. This isn't foolproof, it works works provided the existing display doesn't have a short. If both displays exhibit the fault, cutting the correct 'leg' or legs in the middle on the old display may help reveal this, and you can bend it back and solder back together if it's not shorted.
    • You replace the Master display glass with a standard 6-digit gas plasma display, you can even use an old (but working) player display.
    • You can also replace the Master display with a player display that has a digit fault and would otherwise be unusable! It's possible because the '100,000s' and '100s' digits are not used on the Master [x00x00]. So you can take two faulty components (a non-working Master and non-working player score display) and make a working Master Display Driver Board. Pinball repair nirvana.

Once you feel the display glass is good, you can look to the ICs which are dedicated to driving the Master (or Credit/Match) Display:

  • Start by testing IC1, a 14069 Hex Buffer/Inverter with your logic probe. Run the display digits diagnostic test, with the Auto/Manual switch in the Auto/Up position. The digits on all displays should be counting 0-9. Test the strobe lines are being inverted. The strobe lines and associated pins are in the diagram given below.

When reading this diagram remember that Player 1&3 Score Displays share Strobes 1-6, while Player 2&4 Score Displays share Strobes 9-14. One strobe for each digit.

The Master Display has 4 digits only: Strobes 7,8,15,16. So if the output pin is not the inverse of the input pin or a signal is missing, then replace IC1. It's available as a 4049U or MC14049UB. Cost is maybe 50 cents, certainly < $1 even in low numbers. Useful chips to buy and have around, as the 4049 is also used on the System 3-7 Driver Board for the switch matrix and on other later games.

Sys3-7-Displ-Digit-Driver.jpg
Digit failures will show up in these pairings if they are related to the strobes (so digit drivers):

  • Strobes 1-6: shared for players 1 & 3
  • Strobes 9-14: shared for players 2 & 4
  • The Diagram below is the same for all games, as are the Master Displays which makes them interchangable:
  • The Table / Diagram is listed in the most obvious order: Master, Player 1,2,3,4 displays.

This is a clearer Table from the above diagram:

MASTER DISPLAY
DIGIT (UDN)6184 DIGIT DRIVER STROBE 4049 INVERTER BUFFER
(Left) 10,000 IC4 pin 1 / 18 15 IC1 pin 11 /10
1,000 IC4 pin 3 / 16 16 IC1 pin 1 / 2
10 IC4 pin 5 / 14 7 IC1 pin 3 / 4
(Right) Units IC4 pin 7 / 12 8 IC1 pin 5 / 6
PLAYER 1 DISPLAY PLAYER 2 DISPLAY
DIGIT (UDN)6184 DIGIT STROBE 4049 INVERTER DIGIT (UDN)6184 DIGIT STROBE 4049 INVERTER
100,000 IC11 pin 1 / 18 1 IC2 pin 5 / 6 100,000 IC11 pin 7 / 12 9 IC3 pin 9 / 8
10,000 IC11 pin 2 / 17 2 IC2 pin 3 / 4 10,000 IC11 pin 8 / 11 10 IC3 pin 5 / 6
1,000 IC11 pin 3 / 16 3 IC2 pin 1 / 2 1,000 IC12 pin 5 / 14 11 IC3 pin 3 / 4
100 IC11 pin 4 / 15 4 IC2 pin 9 / 8 100 IC12 pin 6 / 13 12 IC3 pin 13 /12
10 IC11 pin 5 / 14 5 IC2 pin 11 /10 10 IC12 pin 7 / 12 13 IC3 pin 11 /10
Units IC11 pin 6 / 13 6 IC2 pin 13 /12 Units IC12 pin 8 / 11 14 IC3 pin 1 / 2
PLAYER 3 DISPLAY PLAYER 4 DISPLAY
DIGIT (UDN)6184 DIGIT STROBE 4049 INVERTER DIGIT (UDN)6184 DIGIT STROBE 4049 INVERTER
100,000 IC12 pin 1 / 18 1 IC2 pin 5 / 6 100,000 IC13 pin 3 / 16 9 IC3 pin 9 / 8
10,000 IC12 pin 2 / 17 2 IC3 pin 3 / 4 10,000 IC13 pin 4 / 15 10 IC3 pin 5 / 6
1,000 IC12 pin 3 / 16 3 IC3 pin 1 / 2 1,000 IC13 pin 5 / 14 11 IC3 pin 3 / 4
100 IC12 pin 4 / 15 4 IC3 pin 9 / 8 100 IC13 pin 6 / 13 12 IC3 pin 13 /12
10 IC13 pin 1 / 18 5 IC3 pin 11 /10 10 IC13 pin 7 / 12 13 IC3 pin 11 /10
Units IC13 pin 2 / 17 6 IC3 pin 13 /12 Units IC13 pin 8 / 11 14 IC3 pin 1 / 2

If that doesn't fix it, next move on to the Segment Drivers:

  • Look at IC4, the UDN6118A, follow the steps to do a DMM test of the UDN chip. A trick here can be that all the pins are not used (because of the missing digits, so it is possible to make use of the original UDN6118-A-1 (or UDN6184-5) that has a previous fault and one I/O pair is faulty.

This works for the Master Display, so I would keeep a faulty UDN chips (or a faulty board) for spares. You never know (e.g. there is an extra unused pair on every UDN7180 type chip pins 1+18 when we get to Segment driver tricks!!! This does not work when you get to for System 7-9 Masters, as all the pins are used.

Pairs that are needed for IC4 are: (1,18) (3,16) (5,14) (7,12). So you can get away without pairs: (2,17) (4,15) (6,13) (8,11) so you can have up to 1/2 the chip faulty, but it needs to be the right positions without using wire jumpers.

There is a way to test the UDN6118's with the power off and your DMM (multimeter).

  • With the game OFF:
    • Remove the power in to the Master Display board, connector 4J7.
    • DMM goes on the Diode test setting, usually a symbol like this: ->|
    • Red lead clipped to ground, I use the ground braid in the backbox, or on the display board connected by a jump lead (alligator clips each end)
    • Touch black lead to the UDNxxxx pins 2 through 8
    • You should get .5 to .7 for each pin
    • Then touch Black lead on each UDNxxxx pins 17 through 11.
    • An open reading (no reading) is the correct result.
    • A shorted display glass can show up during the UDNxxxx 11 through 17 test.
UDNReading.png


Shorts Are Found in UDN Chip Test
If a short reading is found in the tested pins (don't test the pairs 1,18 and 9,10), the UDN chip should carefully be desoldered and removed from the board. Take care to preserve this chip, as they are nearly impossible to find and expensive to replace. Now install an IC socket in its place. Repeat the test with no chip installed. If the short is gone, then the UDN chip needs to be replaced. If the short remains, then the display glass needs to be replaced.

Notes on different board revisions: Williams also used the DI-0512 (Dionics-512) for the Digit Driver IC. I have owned the discrete transistor version, but the Dionics-512 version is even harder to find, and I've not seen one in 7 years of collecting. That's not to say they are not out there. All the troubleshooting instructions above are intended for the UDN type boards, but the Dionics boards work in a similar way, so consult the schematics. You will also need to make sense of the warnings below.

Warnings: These are factory changes, but if you replace or swap chips, you should be aware of the following

  • The Display board must run similar chips, so for example all UDN-6118 / UDN-6184 types for the Digit Drivers OR all DI-0512 Digit Drivers (which are longer, meaning they have more pins). They have different power requirements! The two Segment Driver ICs need to match, too. So two MC14548 ICs at IC6 & 7, OR two MC14543 ICs at IC5 & 8. Mix & Match or having all four installed is mental (crazy)! You should have empty pads on the board for the optional IC chips.
  • With the DI-0512 at ICs on these boards, the '10K' resistors R1-R14 are all 15K. Adjust the repair instructions below accordingly.
  • If IC4 and IC11-14 are UDN-6118 chips, then a +100v trace is cut connecting pin 2 of connector J7 and ZR1 (1N4740A Zener diode) is added.
  • With the DI-0512's this trace is left, and there is no Z1 used.
Williams System 3-6 Master Display Board



Sometimes a HV arc can occur between two pads on the Master Display board, and if this happens it will blow out the HV section of the PSU (power Supply Unit). If you have the UD7180 version of the System 3-6 Master Display board, it may be worthwhile taking this easy step to prevent a problem:

There is an unused track on the front of the board near to the bottom right corner of the Master display glass. The top round pad in this picture is unused on the UDN7180 version but goes to the +100v supply at Pin 2 of 4J7. Click to expand the picture at left for a better view.
It appears that the pad shorted out to the ground track that goes around both of the round pads. Cutting the track to the top round pad as shown in the picture (with the board removed from the game) will cause no problems for the displays on your game. And it could prevent disaster. A short could damage the HV section of the PSU, even the Master Display board itself. I have seen it happen, so better safe than sorry. BTW- This track should not be cut if the Dionics Digit Driver ICs are installed. The writer has owned many games and has never seen a board with the Dionics ICs.

6.18 Sound / Speech Board Issues

Sounds and speech are one of the most important parts of a pinball game for me.  Few things are more depressing for me than the sound of a pinball rolling around the wood with no background noises and without sounds of points being scored.

This section deals with the square sound board Type 2 ( System 6 & 7, with or without the speech "daughter card"). The Williams part number is 1C-2001-146-x, although much of the information applies to the Type 1 boards (System 3 & 4).

A description of the sound board Type 2:

  • The sound board is actually a "mini-MPU" board running a 6808 CPU and a single 6821 PIA. So read the MPU board sections about the replacing sockets and the reset circuit as it will work in a very similar way.
  • It has an external 6810 RAM for the CPU registers / 'stack'. This can be replaced by internal RAM in the 6802 CPU if that is used.
  • It has it's own PSU (Power Supply Unit) and takes ~18v AC voltage directly from the games mains transformer and rectifies it to DC voltages that provide the +5v logic to the ICs and +12v for the CPU reset circuit.
  • It provides -12v DC only used on the speech card for IC2 & IC3 the two 1458 OP Amps. They act as mixers for the analogue sounds and digital speech.
  • A single 2K 2716 Eprom (ROM) which contains program instructions that 'boot' the CPU and code to create the sounds. Different ROM versions are containing specific sounds are numbered: Williams' Sound ROM 1, Sound ROM 2, Sound ROM 3...

The last item deserves some explanation. The Sound ROM does not store sounds as samples, but as mathematical 'strings' which describe parameters such as attack, frequency, decay and echo. So the board functions as a synthesizer, rather than playing a WAV file or compressed sound like an MP3 player today. In fact a 2K sample at 22K Hz sample rate could last about 1/10 of a second.

The 4K 2532 Eproms on the Speech Board can hold sampled sound but at very low quality. Black Knight used 4 Speech ROMs for a total of 16K and could speak about 20 words along with the famous laughter. Earlier games having 3 Speech ROMs Like Gorgar or Firepower could speak only about 11 words using 3 x 2532 ROMs.

Why did they provide a different PSU for the Sound card? This early Williams game design could be for ground isolation, perhaps to provide a cleaner sound by avoiding interference from the other components which created RFI. Notice that I used "early Williams game design" and "cleaner sound" were just used together in the same sentence. I have also seen Sound Boards where the BR1 was missing and replaced by 4 x diodes, which also works. Two were mounted on the front and two on the back, with the bands on the diodes all pointing in the direction of TP1 (+12v). Start at 1N5401, rated 100v @ 3A.

Understanding this background should help pinpoint the source of problems with Sounds and Speech on games of the early 80's.

6.18.1 Useful Sound Board Repair Links

Leon Borre's work is a good starting guide for Williams sound board repair. He was a clever guy in Belgium, who developed special test ROMs for lots of pinball systems, including Williams System 3 through 7. I can't thank him enough here! He developed a test chip that can be downloaded and burned to a 2716 Sound EPROM, which will start the CPU, "pulse" the PIA and also test the memory. It's harder to do any advanced fault repairs without this test ROM. So it's a good place to start if the sound board doesn't boot at all. Leon's technical article pertaining to Williams System 3 through 7 sound can now be found here.

Dave Langley's site robotron-2084.co.uk has a clear diagram of jumper settings for most games on Williams Type 2 boards. Check jumpers first, W1 must be present if no speech board is attached.

6.18.2 Testing for a Sound Board fault

Before starting to repair a Williams sound card, first determine that there is actually have a sound board problem! Do the sounds work correctly from the sound board 'self test' by pressing the diagnostic button? Have you followed the diagnostic procedures in the game manual and checked the 'Solid State Flipper Maintenance Manual' or equivalent for the specific game? There will probably be a fault on the sound board if the diagnostic test on the sound board produces nothing (or the wrong sounds) and grounding the input pins won't work. See the 'sound selects' section for specific games below. Check out the Basic Sound Troubleshooting section first. After reviewing it, here are are some specific things to eliminate first:

  • Bypass the volume control in the cabinet by removing the connector at IOJ4 and jumpering pins 1 & 2. It will be really loud (full volume in fact) so a couple K ohm resistor (3K?) looped around those two pins will substitute for the volume pot and save your ears.
  • Take the sound board out, leave all the connectors still plugged in and the game running.
    • The speaker and volume control can be set as above.
    • Touch the back of the sound amp at IC1 with a finger - it's just under the Volume control connector J2 which was jumpered and has 5 legs. Not hearing this hum could be a missing +12v or a bad amp IC.
    • A loud loud hum should be heard from the speakers, this proves the TDA2002 sound amp is working and also that the speakers are capable of producing sound.
Loss of +12 or -12VDC could be caused by a failed bridge rectifier

Treat the sound board as a mini-MPU board (but with an on-board PSU) and all the usual fixes apply:

  • Test there is around 18v AC coming in to the sound card. It's directly from the transformer to connector 10J1 and then fused with 2 x 8A fuses.
    • Are both fuses good? Take them out and test them.
    • Is the voltage getting further than the fuse clips? Are the fuses hot to the touch after the power is switched off? This may indicate high resistance at the fuse clips.
    • Pin 5 of 10J1 is the 'center tap' or reference for the AC. There should be about 9.6 V AC between pins 1 and 5, and the same voltage between pins 9 and 5.
  • Use the test points to measure DC +5v +12v and -12v. A short-cut (and much easier) is to touch the Red test lead to the left side leads of the 3 big capacitors, with the black lead on Ground which is the metal trace on the edge of the board.
    • Remember -12v is only needed for the speech board, and even around -10v will normally be fine.
    • Don't worry if the voltages aren't perfect (especially the +/-12v DC which is unregulated). Only the +5v needs to be in a narrower range (4.8-5.2v).
    • No +5v with +12v present suggests the 1 amp '7805 5v regulator' at IC8 is faulty. It looks like the TIP-xxx transistor as it is in a TO-220 case, with a heatsink. Measuring +12v input on the top pin and no +5v output on the bottom pin will confirm this diagnosis. Obviously the CPU (and other ICs) can't run without the +5v logic voltage.
  • If there are Scanbe sockets, replace them first, including the ROM socket. The PIA can be left alone for now if it's soldered directly to the board and not socketed.

Read the MPU board repair and troubleshooting if the board still won't 'boot' (the CPU, working RAM and reset circuit are very similar).

    • Because there are no LEDs to show the board is 'locked up', you will need to look at pins 9 - 20 & 22-33 on the CPU for any activity with a logic probe. If the CPU has started, there should be some activity on these address and data bus lines. If they are 'dead', then the CPU has not started.
    • If you look at the CPU bus lines with Leon's sound board test ROM (instead of the sound ROM) you can see which lines are missing, and also will see the PIA lines 'dance' as he calls it. The PIA I/O lines strobe Hi/Lo/Hi/Lo if they are working as they should. The test ROM will need to be burned on a 2716 Eprom.
    • Review Leon Borré's Williams sound board article for advanced troubleshooting. You can fix boards more easily with his free test ROMs.
    • Note: The the test chip will not start on a Type 2 board with jumper W1 missing, unless the speech card is also connected. Mine didn't start "pulsing" without the speech card. Because of this, here's a key tip: Jumper W1, and set DS1: SW1 and SW2 to OFF and run without a speech board until the sound is working 100%. Do this now if you haven't removed the speech board already, it may not be obvious at first, but you have to get back to a basic setup and work forward from there.

See also Advanced Sound/Speech Board Repair on the Bench

6.18.3 Back to Basics

Temporary W1 jumper installed
Bypassing the volume control


The next step, once you have verified voltages is to remove the speech board (if one is fitted) and jumper W1. Verify W1 is present for any board you are working on, as no sound gets to the audio amp without a speech board or jumper W1 present.

To bypass the volume control, a 470Kohm resistor can be installed across pins 1 & 2 of connector 10J4. Alligator clips can be used, but if the intentions are to test sound boards in the future, a resistor can be crimped and installed in a .156" housing. Although a 470Kohm resistor is used in this example, other values can be used to obtain the desired volume. 470Kohms sets the volume to medium loudness.

The .156" set up can be used on all Williams sound boards from System 3-7 games and 9-11 games too.

6.18.4 Modification for Noisy Type 1 Sound Boards

On the Type 1 sound boards, Williams did not use a small value bypass capacitor on the input to the 7805 +5V regulator as recommended by the manufacturer. This bypass capacitor is used to keep noise off of the +12V voltage rail, which is used by the power amplifier IC. Installing a .22uF ceramic capacitor across the input pin (pin 1 - left pin from front) and ground (pin 2 - center pin) of the 7805 regulator can significantly reduce the amount of noise and hum on these sound cards. Important note -- the Williams Type 1 Sound Board schematic diagram has an error. Pin 1 is the input, pin *2* is center pin and connected to ground and pin *3* is the output.

6.18.5 System 4-6 Sound Selects

Press the diagnostic switch on the Sound Board. When sounds do work on the sound board self test, but you get missing or incorrect sounds during a game, then a simple test for all models of sound card is to ground the input pins. This gives you an indication that the sound board is working, as each pin should produce a distinct sound. For Type 2 boards (square ones) that's 10J3 pins 2-4 and 7, pin 1 is on the right, It may also be the same pin locations on the Type 1 boards (Flash, and other System 3 & 4 games).

Type 1 Sound board connections and grounding sound select header pins. Here, one end of a jumper wire is clipped to the ground plain of the board itself.


Grounding those sound select pins is exactly what the Driver Board does during a game solenoid diagnostic test. It fires solenoids 9-13 to ground the select pins and this triggers 'sound calls' on System 3-6 games. If grounding the sound board pins works, then suspect cabling between the Sound Board and Driver Board or faulty transistors (or pre-drivers) on the Driver Board for the System 3-6a games. You can ground the transistor's metal tabs from the chart below, if this does not produce sounds then the cabling between Driver and Sound Boards is then suspected.

Please note that grounding the center tab of a transistor does not test the transistor(s), it only proves that the cabling from that tab out to the sound board select pins has continuity. It does not prove the transistor(s) work as a switch under the CPU's control.

This is true for any testing of solenoids (or flashers) by grounding transistor tabs. It only proves that the cabling from the Driver Board out to the device is intact, and that the device has power and can function when grounded.

So the Solenoid portion of the diagnostic test will help you figure it out for a System 3-6 game. It can perhaps help for System 7 game too, although remember that sound calls are triggered from the MPU board instead.

On System 6 (and earlier) you will have a fault on the Driver Board, and specifically the solenoids in the range 9-13 if the following are true:

  • The self test on the sound board works as expected,
  • Grounding tabs on the Driver Board produces the same sounds,
  • But the game's diagnostic tests do not trigger sounds when testing solenoids number 9-13.

If it fires the device, the problem is the transistors or the logic on the driverboard and then the PIA upstream from them. You should first replace both the associated TIP 120/102 and the matching 2n4401 pre-driver transistors on the Driver Board (a fairly easy fix) and re-test. Then follow Driver Board testing and repairs.

If a transistor gets shorted 'on' (or a 7408 IC is faulty) and one of the sound selects is constantly grounded from the Driver Board, you will certainly get fewer and usually wrong (or no) sounds played. This type of problem can be chased around for hours, if you don't work logically through the problem.

Driver Board Positions
2J9 10 9 6 7 8
2J9 11 12 13 14 15

Blue shows the 5 Sound Select Transistors on the bottom two rows of a System 3-6(a) Driver Board.

If diagnostics don't work as you expect, as explained, ground the metal tabs of the TIP122 transistors (briefly - they may be real coils or flash lamps!) on the bottom 2 rows of the driver board. You should hear 5 different sounds. The first 3 sound select transistors are at the bottom left of driver board. The first two the next row up are the remaining sound select transistors (solenoids 9-13 as shown on the chart above).

You can of course also turn off the game and measure between the center tab and the input pins on the sound card with an DMM (meter set on continuity or Ohms). If you get high resistance readings, check for cold solder joints on the input pins of the sound board at 10J3. Then check the output pins on the Driver board at 2J9. Reflow the solder on all these pins. Do the other pins on the sound board at the same time.

Another tip is that you should see the center tab of the Blue "sound transistors" running at about +5v DC on the driver board with the game running. If a tab is at zero volts, and others are at +5v then that grounded select points to the problem. Other TIP122 transistors could be at the coil potential, so have your DMM set to voltage and +50v or above for this test if it does not 'auto-range'.

If you get this far and can get 5 distinct sounds, then the problem is not the sound board, nor the connection to the driver board. With System 3-6, this could still be on the driver board and specifically:

  • a power transistor TIP122,
  • a matching 2n4401 pre-driver transistor ,
  • a 7408 logic IC
and finally
  • a 6821 PIA for Solenoids at IC5. As this is a 40 pin IC, it should be replaced as a last step.

At this point just follow the driver board testing and repair guide to get further.

Complete list of Driverboard 2J9 pins and the associated transistors (for Sys3-6) are:

System 6 Driver Board 2J9 Sound Outputs
Pin Sol. Function Transistors Wire Colour
2J9 P9 # 9 Sound Select 0 Q31 / Q30 Brown / Black
2J9 P8  (*) Key N/C
2J9 P7 #10 Sound Select 1 Q33 / Q32 Brown / Red
2J9 P6 #16 Coin Lockout* Q45 / Q44 Brown / Grey
2J9 P5 #15 Depends on game* Q43 / Q42 Brown / Violet
2J9 P4 #14 Credit Knocker* Q41 / Q40 Brown / Blue
2J9 P3 #13 Sound Select 4 Q39 / Q38 Brown / Green
2J9 P2 #12 Sound Select 3 Q37 / Q36 Brown / Yellow
2J9 P1 #11 Sound Select 2 Q35 / Q34 Brown / Orange
  • Typical functions, but depends on the game.

List of J3 sound select pins on the System 6 or 7 sound board are:

System 6/7 Sound Board Inputs
10J3 Pin Sol. Function Wire Colour
J3 P3 # 9 Sound Select 0 Brown / Black
J3 P2 #10 Sound Select 1 Brown / Red
J3 P5 #11 Sound Select 2 Brown / Orange
J3 P4 #12 Sound Select 3 Brown / Yellow
J3 P7 #13 Sound Select 4 Brown / Green
J3 P1  (*) Key N/C

On System 6, using only 5 solenoids for sound limits the number sound and speech calls. Makes 25 = 32 so only 31 usable combinations, a call of all 0's being null. But (using Firepower as an example) you then can have all the combinations of maybe 20 sounds plus the 11 speech phrases it "knows". So it's enough given the small amount of sound and speech memory they had to work with at the time. In fact it's amazing! For more background, see: An interview with " Eugene Jarvis, the sound engineer and programmer for Williams at the time.

Changing the sound board won't cure the problem of a missing solenoid signal, and a major clue is that the sound board tests good, and is also good when grounding the 5 input pins at (Board #10) J3 input connector on the sound board.  While the "in game" sounds or game diagnostics will play incorrectly or have missing sounds. For example, an email from someone said that their Firepower game could say "Power" but not "Fire"! A fairly easy diagnosis if you know the game, it pointed to transistor Q35 (a TIP122, a better replacement is the TIP102) which was completely missing from the Driver board. Transistors on the bottom row often get bent back and forth and snap clean off - hard to spot if you aren't familiar with inspecting the Driver Board.

That transistor's center tab connects to 2J9 pin1 and solenoid 11 then fires sound select #2, according to the table above. Without that sound select being triggered, only half the game sounds could be produced (24-1> = 15). "Fire" was one of the missing sounds, there were others but they may have not been so obvious to the owner. They would not be listed in the manual as a sound triggered by an achievement (making to rollovers F-I-R-E in this case). The point is to use all the information available to help determine the root cause of your fault.

6.18.6 System 7 Sound Selects

By System 7, Williams had stopped using solenoids to fire sounds and had added a dedicated Sound/Comma PIA on the CPU board to trigger sound/speech. This meant more playfield coils and even flash lamps were now possible, as the Driver Board wasn't used to trigger the Sound Board.

On a Black Knight and later games, the fault could still be on the MPU board. Perhaps PIA 5 (IC36 a 6821 PIA) is suspect as this drives the sound board (and display commas) only. Again you can trigger sound selects by grounding pins on the sound board, and then move back to the header pins of the MPU at 1J8. The pins are12,11,10,9,8 in that order for the 5 sound selects 0-4

The outputs of the IC36 PIA at PA0-PA6 connects to 1J8 starting at pin 12 (1J8 P12) and run backwards numbered as the Sound Selects in the table below:

SYS 7 1J8 Sound & Comma Outputs
Pin Function Wire Colour
1J8 P1 Comma 3 & 4 Brown / White
1J8 P2 Comma 1 & 2 Violet
1J8 P3 Key (*) N/C
1J8 P4-7 No Connection White
1J8 P8 Sound Select 4 Yellow
1J8 P9 Sound Select 3 Green
1J8 P10 Sound Select 2 Blue
1J8 P11 Sound Select 1 Red
1J8 P12 Sound Select 0 Red / Yellow


In theory Williams had the ability to trigger Sound Selects 0-6 from the System 7 MPU board, and could have programmed 127 sound / speech calls. But the sound boards at the time never used that many, to my knowledge. So you only need to worry about the same 5 sound selects 0-4. Be sure to first check the output pins at 1J8 on the MPU Board for cold solder joints, before replacing the 40 pin PIA.

6.18.7 Put the Sound/Speech Board on the Bench

By this point, it is likely that there is a sound board fault. If you don't feel comfortable troubleshooting or don't want to get involved with low level board repairs, there are repair services available--check pinball communities/forums for recommendations. Reproduction sound and speech boards are also available if all else fails.

WMS System 6/7 Sound Board Pinouts




If none of this helped so far, read the rest of the section below. The boards will likely need to be bench tested in order to proceed with testing. Any PC power supply has the +12v and -12v DC that will be necessary for bench testing, which are normally located on the motherboard connectors. The older the PC power supply, the better, as they will have a power switch, but there are two pins that can be jumpered (or connected to a switch) to enable the voltage outputs.

WMS System 6/7 Sound Board Connected to an AT Power Supply
WMS System 6/7 Sound Board Connections for an AT Power Supply (Four Sound Board Test Points)


Search for an "AT Power supply" on the web for the connector locations or measure them with the DMM. The wires on one example were colored Yellow (+12v) and Blue (-12v). Ground was black as usual. Sliding the existing pins from the motherboard connector housing, and they will potentially fit right in a .156" socket to connect to the sound board. The +12 and -12 goes to the outsides of a 9 pin connector (pins 1 & 9) and the ground to the center pin (pin 5). It doesn't matter which way around the connector is installed--if the Bridge Rectifier (BR) is working it will sort it out by conducting through the correct diodes.

Another method is to connect the AT power supply connections with alligator test leads directly to the four test points on the sound board. However, this method will not ensure that the bridge rectifer on the sound board is working or not.

At this point refer to Leon Borre's sound board test link. He has an excellent approach to a sound board that is dead. A copy of his sound test ROM will need to be burned to an EPROM. A logic probe (under $20) will also be needed, but Leon describes how to make an LED test probe with some wire, an LED, and a resistor which costs a few cents. The logic probe will be more versatile and easier to use. As mentioned before, the Test ROM can check the CPU, memory and also exercises the PIA inputs/outputs. An added bonus is that the test ROM will allow the board to boot, even though there are faults on the board.

6.18.8 Checking for Sounds Prior to the Sound Amplifier

An unamplified speaker used to probe for sounds before the sound amplifier

A method to test for sounds prior to the amplifier is to check the output of the D/A converter. This can be accomplished by building a speaker set up. It starts with a low wattage speaker, which does not need amplification to produce sounds. A speaker from an old computer case is a great source. After acquiring a speaker, solder two wire leads to it. Then, solder an alligator test clip to one of the leads, and a small piece of solid, copper, house wire (14 or 12 gauge romex works well) or a small finishing nail to the other lead. The second lead will be used to probe the sound output before amplification. This speaker test set up can be used on any other brand sound board too.

After the speaker test set up is constructed, clip the alligator test lead to ground. Then, place the probe lead on the "C" collector leg / solder pad of the Q2 transistor, as shown in the picture. Once the probe is in place, press the sound board test button. The speaker will have to be placed close to one's ear to hear the sounds, if any sounds are being output. If the sounds are being heard, either the amp (IC1 - TDA2002) has failed, has lost connection (cold solder joints), or is not being powered properly.

Normally the probe would be used to check the sound output right on the leg of the D/A converter (IC13 - 1408). However, pin 4 of IC13 (the output) is adjacent to pin 3, which inputs the -12vdc power to the D/A converter. It is much safer to use the collector leg of Q2 as a probing point to avoid shorting pins 3 and 4 together.

Note that the if the 1408 DAC has failed, it can be subbed with the more readily available DAC0808.

6.18.9 Getting Sound and Speech Working Together

The very first thing is to make sure you have the correct jumper settings and ROM types.  You can NOT "mix and match" sound and speech ROMs as you like.

My experience is on Williams System 4-7 pinball where the Sound ROMs are 2716 Eproms and the Speech are always 2532 Eproms. The exceptions are video games (and a few pinballs), such as Defender, Sinistar, Robotron and Joust which use a 2532 on the sound board ROM and each has special jumper settings.

You need the correct Sound ROM type for your game's Speech ROMs.

Firepower and Alien Poker both use "WMS Sound ROM 3" no matter what the Alien Poker manuals (or ROM download archives) say. It's the only way I ever got speech to work on my Alien Poker. If you have Blackout, you will need a 2716 Sound ROM 2 to go with Blackout's 2532 Speech ROMs.

This is mainly a warning if you will be plugging together different boards on a workbench or when troubleshooting a problem, don't trust that all Sound ROMs work with all speech boards !

Here is a diagram on placement of the Speech ICs, as the layout of the ROMs aren't very logical. Check they are correct. This diagram works for any of this type of Speech Daughter board. Replace your game number nnn in ST-nnn-ROM# format. See examples for Firepower (497), Alien Poker (501) and Black Knight (500), which are among my favourites. The information is provided for clarity, it is recommended that you replace Original Speech Mask Roms with newer 2532 Eproms.

Note: unless a System 9 speech board is available with jumpers, 2732 Eproms cannot be used for the speech board, unless the board is modified.

Speech ROM IC socket locations: Speech-ICs.jpg

Also distrust the switch settings and be aware that you could have a faulty DIP switch on the Type 2 board. It can happen! Test by removing the 2-way switch, or jumpering across the back of the switch. If DS1-2 is not connecting, you will hear sounds- but no speech.

The clue here is that if you press the sound board diagnostic switch, you hear BOTH sounds and speech.

System 6/7 Speech Board with Scanbe Sockets
Identifying Scanbe Sockets



Speech ROMs and Speech Board sockets should eliminated, you can measure continuity on all pins from the Speech ROM to the bottom of the socket or back of board before replacing Sockets. If they say SCANBE then replace them as they will be faulty.
Gaps in speech usually point to a faulty Speech ROM, look in the manuals for which ROMs hold the missing words.
Suspect the logic selects on the sound board or the 40-way IDC connection between the speech board to the sound board. Watching for 'chip select' pins 20 on each Speech ROM with a logic probe will help you see the chips being selected by the sound board. Again knowing which ROMs hold the missing phrases will help. No speech selects can point to the interboard cables, or missing outputs from IC2 (a 7442 decoder) on the Sound Board.
Bear in mind that the Speech card is an 'expansion board' to hold ROM space for the speech and the digital/analogue mixer. The address and data buses extend across the '40-way' IDC connector in a similar way to the connector between the MPU and Driver Board. Except the connector solution is more robust and IDC was still used until recently to connect IDE PC hard drives.


Other problems that could cause sounds but missing speech include:

  • W1 is jumpered,
  • W9 is NOT jumpered (W4 is probably in place),
  • IC7 (inverter) faulty,
  • IC10 (6821 PIA) is faulty (use Leon's Sound Board test chip)
  • Faulty speech board (no speech) or speech ROMs are installed incorrectly

Switch 2 MUST be ON for speech. OFF for no speech.

Switch 1 selects between Tones and Synth Sounds. ON for musical tones (more like bings and bongs), OFF for Synthesized Sounds.

SW 2 ON
SW 1 OFF

    DS1 Set correctly for Sounds and Speech.


SW 2 OFF
SW 1 OFF

     DS1: Set for Sounds only.

     You will also have to jumper W1 if this is not in place.

6.19 Speech Board Issues

6.19.1 Common failure parts

Speech board sockets. These are Molex, but "they gotta go".
Nice twin wipe sockets, which are preferred over machine pin sockets.


The first and most important action when debugging a Speech board is to eliminate all of the common failure issues. With these issues out of the way, it is much easier to focus on the real issues preventing proper operation. And, eliminating these typical trouble areas almost always brings the board back to full operation.

The most frequent reason for failure of the speech board are the OEM sockets. ScanBe sockets, or white sockets pictured here, should be replaced. If a ROM can be easily removed using only your fingers, the socket should be replaced.

Also check the ribbon cable to the sound card. It breaks frequently at pin 1 which carries the 5V Analog voltage. If you replace the cable make sure that you do not damage the J1 connector. The part is either NLA or so obscure no one stocks it. Fortunately you can gently pull off the cover and reuse it with a new cable. The counterpart is a standard ribbon connector and must be replaced.

These days one or two of the MC1458 chips on the speech board go bad too frequently. This results usually in either no sound and/or missing speech. These should be replaced in any case while you are on it. Use good quality machine pin sockets when replacing.

If your speech is garbled and you replaced the MC1458 chips the 55516 CVSD at location IC1 might be bad. This is somewhat rare but not unheard of. The part is long since obsolete and hard to source. It can be replaced with a 55532, 55536 or 55564. Only the 55564, which was used up to the Williams WPC era, seems to be somewhat available. Check your favorite pinball shop for it.

As with all electronics other parts can fail but the above parts are responsible for around 95% of all speech card failures.

6.19.2 Using the Sound Board Test ROM

Leon's test ROM and instructions will be very useful. The ROM and instructions are available here.

Note: The the test chip will not start on a Type 2 board with jumper W1 missing, unless the speech card is also connected. Mine didn't start "pulsing" without the speech card. You also need to put the Test code on a 2716 Eprom.

Here's a key tip: Jumper W1, and set DS1: SW1 and SW2 to OFF and run without a speech board until the sound is working 100%. Do this now if you haven't removed the speech board already.

This may not be obvious at first, but you have to get back to a basic setup and work forward from there.

The following is an attempt to list notable Williams games that used these boards.

There are conflicting lists, Hyperball could be a 2532 ROM #8. as the file hypsnd12.532 has a checksum of 0x25A8 - which doesn't match other sound ROM images. So the list is not perfect. Some corrections have been as of Feb 2011 after more research. If you have updates or can confirm Williams sound ROM numbers (and compared checksums) then please go ahead and edit this table.

The games with speech in the first table are right, I'm less happy with the Video games and Hyperball.

Williams Pinball Games which use a Type 2 Sound Board and a Speech Board:

NAME SYS GAME WMS SOUND ROM
Blackout 6 495 Sound ROM 2 
Gorgar 6 496 Sound ROM 2
Firepower (r) 6 497 Sound ROM 3
Black Knight(r) 7 500 Sound ROM 5
Alien Poker 6a 501 Sound ROM 3 confirmed
Jungle Lord 7 503 Sound ROM 3
Pharoah 7 504 Sound ROM 6


(r) Steve Ritchie Game Design

Video Games with Sound and Speech Board: Sinistar (cockpit front) & Upright, Type 2,VIDEO ROM 9

Pinball Games with Sound Boards Only:

NAME SYS GAME WMS BOARD, SOUND ROM
Flash (r) 4 486 Type 1 Board, ROM 1*
Time Warp 6 489 Type 1 Board, ROM 1
Stellar Wars (r) 4 490 Type 1 Board, ROM 1
Laser Ball 6 493 Type 1 Board, ROM 2
Scorpion 6 494 Type 1 Board, ROM 1
Algar 6a 499 Type 1 Board, ROM 4
Cosmic Gunfight 7 502 Type 2 Board, ROM 12
Solar Fire 7 507 Type 2 Board, ROM 7
Hyperball

(n)

7 509 Type 2 Board,

SPECIAL ROM 8 ?

Barracora 7 510 Type 2 Board, ROM 4
Time Fantasy 7 515 Type 2 Board, ROM 13
Warlock (pin) 7 516 Type 2 Board, ROM 14
Defender (pin) 7 517 Type 2 Board, ROM 15
Joust (2 player pin) 7 519 Type 2 Board, 

SPECIAL Snd_IC12.532 ROM

Laser Cue 7 520 Type 2 Board, ROM 3
Firepower II (m) 7 521 Type 2 Board, ROM 13
Star Light ** 7 530 Type 2 Board, ROM 3

ROM 4 is on my game!


(r) Steve Ritchie Game Design

(m)  Mark Ritchie Game Design

(n)  Not a Pinball, a gun game that shoots tiny pinballs.

* Sound ROM 1 was used from World Cup #481 to Scorpion #494, but not Laser Ball from my notes.

** Quite a rare and beautiful pinball. Only 100 were produced, the last Williams System 7.

Here is some useful pinball sound file checksum information. At Tom Callahan's pinlogic.com

NAME WMS BOARD VIDEO ROM
Bubbles Type 2 Board, VIDEO ROM 5
Defender (Video) Type 2 Board, VIDEO ROM 1
Joust (Video) Type 2 Board, "VIDEO" ROM 4

(Checksum matches ROM 4 from Algar and Barracora)

Robotron 2048 Type 2 Board, VIDEO ROM 3
Sinistar (cockpit rear) Type 2 Board, VIDEO ROM 10


Please Note: VIDEO ROMS and those for Joust (a 2 player head to head pin) and Hyperball are 2532 Eproms

An annoying fact is that from around System 7, most of the WMS sound ROM files are all named something like SND_IC12.716 , which makes them harder to identify.

6.19.3 Running a Type 2 Sound Board with a 6802 CPU

You might not have an 6808 CPU, or you might upgrade to a 6802 because then there is one less IC (the 6810 RAM) to go wrong. You can leave the 6810 (as it's not socketed) if the card functions with it in place.

You can get the Type 2 board to run on a 6802 CPU, without a working 6810 RAM in IC11. To do this you cut a track called W14. It isn't a 0 ohm resistor (or a real jumper), it's a track located under R30, the 3rd resistor in from Left on the Bottom Left of the board. W14 grounds pin 36 of the processor for a 6808 CPU.

If you want to use a 6802, you cut the track W14. Pin 36 will go high through the 4.7K "pull up" resistor at R30. You could then remove the 6810 RAM as the 6802 has internal RAM. My advice is again to just leave it in place if the card functions with it there. I had to rework tracks damaged by removing a 6810 on one sound board being repaired. Avoid making extra work for yourself, and spend more time playing pinball. That way it can be reverted back to 6808 operation in future. Good instructions are currently at www.robotron-2084.co.uk. A useful site for arcade video and pinball and Dave has a nice Defender pinball to look at while you're there.

6.19.4 Modifying the System 6/7 Speech Board to Use 2732 EPROMs

The original System 6/7 speech board is setup to only use 2532 EPROMs. With the following simple modifications, a speech board can use 2732 EPROMs instead, if the need arises. Please be aware that although these modifications are reversible, it is not nearly as easy to change a modified board using 2732 EPROMs back to a board which uses 2532 EPROMs.

Procedure:

  1. On the component side of the board, there is a trace which runs beneath the ribbon cable connection and between pins 11 and 12 of U5 (the lowest two pins on the left). Cut this trace.
  2. Turn the board over to the solder side. The orientation of the board in the pic below is with the ribbon cable connection on the right.
  3. Locate pin 18 of U5. There is a trace between pin 18 of U5 and a via down and to the right of it. To verify the correct via, check continuity between pin 13 of the ribbon cable (fourth pin up on far right of the ribbon connections) and the via. Cut the trace between this via and pin 18 of U5
  4. There is another via slightly down and to the right of the previously discussed via. This second via will have continuity between it and pin 12 on the ribbon cable (third pin up on far left of the ribbon connections). The via will also have continuity between pin 21 of U5. Using wire wrap or a thin gauge wire, tie these two vias together.
  5. Locate pin 18 of U5 and pin 12 of U4 (down and to the left of U5 pin 18). Again using wire wrap or a thin gauge wire, tie pin 18 of U5 to pin 12 of U4.
  • System 6/7 Speech Board 2732 EPROM Modification
  • Trace cut on component side
  • Trace cut and two jumpers added on solder side


6.19.5 Replacing the on-board volume control with a remote volume pot (System 4)

System 4 original volume pot
Connector 10J4, with pins 1 & 2 bridged with solder.


If the volume pot on your system 4 or later sound board fails or otherwise needs replacement, it's often difficult (or impossible) to find a direct matching part. In such cases adding a remote volume pot may be the only solution.

The first thing to do is flip the board over and make sure that somebody hasn't bridged pins 1 and 2 of connector 10J4, as shown here. If it has been bridged, remove the solder bridge and solder the pins independently to the board.

Taking a reading across volume pot legs.
Taking a reading across volume pot legs.


Take a resistance reading across the existing volume potentiometer's legs. If you get a reading above 4K ohms, the potentiometer is not mostly shorted and will not require its leg to be cut as detailed below. (i.e. you may skip that step below.)

This potentiometer is not mostly shorted and could be left un-cut.

Cutting the volume pot's left leg.
After cutting volume pot's left leg.


If the previous resistance reading reads less than 4K ohms, then one leg of the potentiometer must be cut. (This one was cut for the sake of demonstration.)

After the cut, the potentiometer should look similar to this.

Jumper added across volume pot legs.
Replacement potentiometer.



In all cases, a jumper must be added to the back side of the board to bypass what is left of the old potentiometer.

A new 5k-ohm, audio (logarithmic) taper potentiometer must be sourced.

Wiring of new potentiometer.
Wiring of 10P4.



Here a radio shack 271-1720 potentiometer was used.

Three wires must be run between the potentiometer and 10P4, the plug that connects to 10J4. The wires must then be connected to 10P4, as shown.

Plug the new 10P4 into 10J4, and the on-board volume pot will be bypassed with the new external pot.

6.19.6 If All Else Fails

You can source a replacement sound/speech board. It may be more cost effective than sending out your sound boards to be fixed, especially as you may be able to sell the original boards to get back some of the initial outlay. These are good quality products, the price isn't bad and they should last for many years to come. James Kohout's pinballpcb.com

6.20 Solenoid Problems

6.20.1 Controlled Solenoids

In addition to the ground braid being screwed to this post, a white wire with red stripe is also connected.


Some Williams games provide ground for coils via an extra wire (white/red) that must be connected to the backbox ground stud as shown in the picture at left. If this wire is not connected, some coils will not work. An example is Pharaoh, which uses this path to ground for both "Magna Save" magnets and the hidden tomb coils.

Note: in the example picture, a nut has been used to replace the typical "wing nut".

6.20.1.1 Special Note Regarding Coin Lockout Coil and Free Play

When some games (perhaps all games but at least Flash using Green ROMs) are set on "free play", by setting adjustment 18 to zero, the coin lockout coil will not be turned on. This makes sense since when free play is enabled, coins should drop straight through the coin mech and into the coin return.

6.20.2 Special Solenoids

A closeup of the special solenoid circuitry showing transistors Q2 through Q12. Note that these transistors are not numbered row and column sequentially. Argh!
Diagram of Williams System 3-7 Driver Board Special Solenoid Section


The Special Solenoid drives switch power to ground for pop bumpers and sling shots. This "special" circuitry was implemented with the thought that the switch matrix could not be polled rapidly enough for these devices to be highly responsive. The drive transistors can be switched on either by a switch closure on the device, or under software control by the CPU.

The Special Solenoid drive transistors, Q2 through Q12, are located just above the flipper power ground enable relay Z1.

The flaw in this circuitry is that a "one shot" was not implemented (ala Gottlieb System 80 Pop Bumper Driver Boards). One shot circuitry only allows the associated solenoid to engage once per switch closure. In the case of the Williams special solenoids, as long as the switch is tied to ground, the solenoid will be engaged. As long as a device switch (pop bumper or sling) is closed, then coil power will be switched to ground continuously, eventually overheating and killing the associated TIP-122 (or TIP-102) transistor and / or the solenoid.

There is also a lack of "intelligence" in the software as was implemented eventually in WPC games. WPC games disable a device if the device's associated switch is sensed closed 10 times in succession without sensing any other switch closure.

6.20.2.1 Theory of operation

The trigger switch is connected to ground on one side. It is connected to 5V on the other side via a 4.7Kohm pullup resistor and also connected to one of the 4 NOR gate inputs of a 74LS02 IC. This holds the input of the 74LS02 at logic high until the switch is closed and thereby shorted to ground, changing the logic state to low.

A 74LS02 is a quad NOR gate. NOR gates require both inputs of the NOR gate to be logic low (0) for the output to yield logic high (1). If either of the 2 inputs to the NOR gate are high, the NOR gate output is low. For example, if input pins 11 and pin 12 are both low (logic "0"), then output pin 13 will be high (logic "1"). If either input pin 11 or 12 is at logic high, then output pin 13 will be low (logic "0").

This implementation allows the CPU to "disable" the special solenoids by simply holding one NOR input high, rendering the output low (logic "0") regardless of the switch being closed or not. The output from the 74LS02 drives a 2N4401 pre-driver transistor which is used to switch the TIP-122 (a TIP-102 or BDX-53C can be substituted) drive transistor on, and hence proved a path to ground for coil power and energizing the coil.

6.20.2.2 Arc Suppression Resistor/Capacitor for Special Solenoids

See also: Wetting Current Resistor/Capacitor for Special Solenoids

While a certain amount of current is necessary to help keep contacts clean, too much current can result in arcing when a switch is opened, which over time, can cause pitting and erode the contacts. To prevent this, an RC network is placed across the contacts.

When the contacts in an arc suppression circuit open, the applied voltage is placed across the capacitor and not the contacts.

The capacitor charges at a rate faster than the contacts open which prevents an arc from forming across the contacts.

When the contacts close, the inrush current from the charged capacitor and source can be substantially higher than the contacts can safely conduct, causing the contacts to deteriorate. This is why it is important to have a resistor in series with the capacitor.

The resistor acts as a current limiter which reduces the inrush current by a significant amount. The arc produced at the contact closure is greatly reduced, thus extending the service life of the contacts.

Reference: https://testguy.net/content/362-Arc-suppression-circuits


6.20.2.3 Wetting Current Resistor/Capacitor for Special Solenoids

A resistor/capacitor connected across a pop bumper spoon switch.
The schematic for the circuit as implemented for slingshot switches.


Williams used a 22uf capacitor and a 100ohm resistor, themselves connected in series, wired across special solenoid switches to provide a minimum "wetting current" across the switch contacts. This implementation also provides some measure of "electrical debouncing" of the switch. In the image at left, the black wire is coil power from the power supply. The orange wire is the path back to ground via a driver board special solenoid drive transistor. The green and white wires are switch matrix wires for the "scoring switch". A 1N4004 switch isolation diode is used just like any other switch matrix switch.
See: https://en.wikipedia.org/wiki/Wetting_current

The special solenoid switch contacts are pulled high to 5 VDC through a 4.7K ohm resistor on the driver board. When the special solenoid switch closes, this voltage is shunted to ground resulting in a current through the contacts of approximately 1 mA. This very low "wetting current" is insufficient to keep the contact surfaces clean and free from dirt and oxides. An RC network is placed across the contacts to increase the wetting current and help break through any dirt or oxidation that has formed on the contact surfaces. When the contacts are open, the capacitor charges to 5 VDC through the 4.7K and 100 ohm resistors in series. When the contacts close, the capacitor discharges through the contacts and the 100 ohm resistor, resulting in a momentary peak current of approximately 50 mA. This discharge wetting current is sufficient to burn off any surface film that has built up on the contacts. Once the contacts open again, the capacitor begins charging again and effectively extending the contact closure time. This also provides some amount of electrical debouncing.


6.21 Lamp Problems

6.21.1 General Illumination Lamps

An extreme example of high current damage to the general illumination traces of a System 7 power supply.


One of the largest weaknesses of the System 7 board set (starting with the later run of Black Knight) is the general illumination PCB connection. The configuration of that connection evolved from a two pin Molex connector into four (two for AC in, two for AC return) 18 gauge wires soldered directly to the board and connected to the transformer secondary with a 4-pin Molex connector.

6.21.2 Controlled Lamps (Lamp Matrix)

One source of problems with the lamp matrix can be the 27 ohm 3 watt resistors used for the lamp columns. These resistors put out a ton of heat! In some extreme instances, they can get so hot that they will literally melt the solder which holds them to the driver board, and fail off the board. It is best to replace these resistors with 27 ohm 5 watt resistors. The 5 watt variants can dissipate the heat a little better. When replacing these resistors, make certain to allow enough space under them to allow for air cooling. Even though it is best to use higher wattage resistors, they still put out a fair amount of heat.

  • Williams System 3-7 Driver Board Lamp Matrix Resistors
  • Old resistors in questionable condition due to overheating. Note these are 3 watt resistors as shipped by WMS. Later boards used 5 watt versions.
  • Using a pencil to allow proper spacing of the resistors from the driver board. One by one, place the legs of the resistors through the circuit board, and bend the legs over.
  • Flip the board over, and allow the resistors to "hang" off the board.
  • New resistors soldered on the board with ample air flow beneath them.
  • Another method of raising the resistors. Be sure to remove the cardboard when complete.
  • Side view of resistors raised with ample air flow beneath them.


6.21.2.1 Testing the Lamp Matrix

This is a generic Lamp Matrix Diagram which you can refer to for testing the lamp matrix.

6.21.2.2 Reducing heat by replacing the TIP-42 lamp drive transistors and eliminating the 27 ohm current limiting resistors

The original Williams (and Data East) lamp matrix circuitry dissipates a lot of heat through the 27 ohm current limiting resistors. Often, this part of the circuit board is badly heat damaged. The TIP42 transistors are bipolar PNP transistors, and require a lot of drive current on the base. The current through the 27 ohm resistors does not even go to lamps, just to the base. The issue starts when the 40 pin interface connector becomes faulty. If the lamp scan update is not received by the driver board, it will sit on one selected column of lamps.. and promptly burn one 27 ohm resistor off the board, possibly burning a hole in the board.

The first method below is simpler to implement, but the second method is included, as it has been implemented for many years.

In the first upgrade method, replace the 8 TIP42 transistors with the darlington version, which require much less current. Get 8 TIP125 transistors, and replace the 8 original transistors, being careful to keep them oriented the same way. Masking tape can keep all 8 lined up while soldering. The second step is to remove all 8 of the 27 ohm resistors, and replace them with 10K 1/4W resistors.

The second method uses modern MOSFETs (IRF9530, IRF9Z34N or FQP17P06).

Once MOSFETs are substituted for the TIPs in the lamp driver circuit, it is possible to replace the 27 ohm resistors with simple jumpers or even leave the 27 ohm resistors in place if they are in good shape as the low resistance will have a negligible effect on the operation of the MOSFET. However, the problem with either of these is that the MOSFETs are driven by 18 VDC Source to Gate voltage which is dangerously close to the absolute maximum rating of the IRF9530 and IRF9Z34 of 20 VDC (this also holds true for Data East MPUs). The FQP17P06 has a higher Vgs rating (25VDC) than either the IRF9530 or IRF9Z34N - but you are still driving the part at over 70% of its absolute maximum. Typical uses for these MOSFETs drive them by about 10 VDC Source to Gate voltage (or -10V for P-Channel MOSFET's used here).

By using a "Voltage Divider" circuit with two 1K, 1/4 watt resistors to divide the source 18 VDC in half, the MOSFET will be driven with 9 VDC Vgs.

1K resistors installed on the component side of the board.
1K resistors installed on the solder side of the board.

Parts Required

  • 8 P-Channel MOSFETs (IRF9530, IRF9Z34N or FQP17P06)
  • 16 1K Ohm 1/4 Watt resistors

Procedure

  • Remove the TIP-42 transistors Q63, Q65, Q67, Q69, Q71, Q73, Q75, Q77
  • Remove the 27 Ohm resistors R149-R156
  • Remove the 2.2K ohm resistors R141-R148
  • In place of the 27 Ohm resistors, install 1K Ohm resistors
  • In place of the TIP 42 transistors, install the P-Channel MOSFETs, oriented the same as the TIP-42s were oriented
  • On the solder side of the board, install eight 1K Ohm resistors between the MOSFET gate and the 18VDC source.


Another implementation of 1K resistors installed on the component side of the board.
Closeup of 1 kohm resistors added to solder side of board


Pay close attention when adding the 1 Kohm resistors to the solder side of the driver board on heavily burnt / delaminated boards. The resistor leg and / or the end of resistor leg comes within close proximity of the same column or adjacent column traces.

Another implementation of 1K resistors installed on the solder side of the board. This implementation requires less resistor leg bending and hence, a much easier install.
A "component side only" implementation. Image courtesy of Victor, Pinsider "DumbAss".


Yet more ways to efficiently add the voltage dividing resistors.

6.22 Flash Lamp Problems

System 3-7 flash lamps (and System 9/11 also) are driven by the same kind of circuitry that drives solenoids. Flashers are generally wired two in series (like old Christmas lights). If one lamp fails, the neither lamp will light. They are wired in series so that they split the DC voltage across two lamps, preventing them from burning out quickly.

To make the flashlamps "agile" (i.e. turn on rapidly) and to extend the life of the flashlamps, a "warming" circuit was employed which connects two series flashlamps to coil power via a 330 ohm 1 watt resistor. This causes the lamp filament to turn on very slightly, and to "warm up", ready for the power to find ground via a 1 ohm (5 ohm in System 11) resistor.

Over the years, these games have undergone "maintenance" to replace parts on the resistor board and have sometimes also been rewired, attaching the wires to different solder lugs on the flashlamp warming board.

Note: If LED flash lamps are installed instead of incandescent lamps, the 330 ohm warming resistor must be removed from the circuit. If it's not removed the lamp will remain on, albeit not at full intensity.

  • Williams System 3-7 (and 11) Warming Board Implementation
  • Firepower warming board. The 330ohm resistor has clearly failed. This board is connected to the game wiring differently than Gorgar and is not factory, yet it works. The red wires are power. It's difficult to see in this pic but a yellow wire is also connected to the red wires. That yellow wire connects to two flashlamps in series. The other yellow wire returns from those flashlamps and is connected to the 1ohm and 330ohm resistors. Power finds ground through the 330ohm resistor and the black wire. This forms the "warming circuit". Power also finds ground through the 1ohm resistor, the brown/violet wire and the driver board drive transistor.
  • Firepower warming board. The 330ohm resistor has been replaced. This board is also connected to the game wiring differently than Gorgar and is not factory, yet it works. The two yellow wires connect to two flashlamps in series. Red is flashlamp power. Black is ground. The remaining wire is a path to ground via the drive transistor.
  • Gorgar warming board. The two yellow wires connect to two flashlamps in series. Red is flashlamp power. Black is ground. The remaining brown/violet wire is a path to ground via the drive transistor. This warming board wiring matches the schematics and corresponds to factory wiring.
  • System 11 (High Speed) warming board. Note that flashlamp power is no longer routed through this board. The orange wire connects to two flash lamps in series which are then connected directly to flashlamp power. The two black wires are ground. The gray-blue wire is the path to ground via a drive transistor.


Warming Resistor circuit shown in Stellar Wars schematics. Note the two #89 lamps wired in series. They are dimly lit with power from the red wire, through the 1 ohm resistor, through the lamps, through the 330 ohm resistor, and then to ground. They will fully light when the drive transistor that services the GRY-BLK wire is switched on, providing a low resistance path to ground and effectively removing the 330 ohm resistor from the circuit.


The diagram at left shows the physical connections to be made by the warming board. The schematic snippet at right is from Stellar Wars (System 4) and accurately depicts connections. The schematic contained in Gorgar is not correct for this circuit. Warming boards are not shown in any of the System 11 documentation.

6.23 Switch Problems

WARNINGS:

  • Never use a file, sandpaper, or anything more abrasive than an old business card to clean gold flashed switches. Doing so will remove the gold flashing and permanently ruin the switch.
  • Never adjust switches with the game powered on. It is far too easy to short the switch, or your adjustment tool, to adjacent power (lamps or solenoids) and damage the switch matrix circuitry on the driver board.

The only switches that can be filed are tungsten switches which are both materially and physically quite different from gold flashed switches. In a System 3-7 game, the only tungsten switches are the cabinet flipper buttons and the flipper mechanism end-of-stroke switches.

Tungsten cabinet flipper switches and end-of-stroke switches "arc" and will become blackened (or whitened) and pitted over time. Once this happens, the resistance between the switch pair increases, which reduces electric current, which reduces coil power, which reduces flipper strength.

Switches vary from game to game, most are standard for WMS System 3-7.

A few of the common switches are listed here:

System 3-7 Switches
Part Number Common use
SW-1A-118 Spinner Blade switch
SW-1A-124 Rollover Lanes (inlanes/outlanes as well)
SW-1A-130-1 Switch Rightmost on Ball Runway (Ball Locks, with round white nylon piece on leaf end)
SW-1A-136 2nd Ball Runway Switch (typical, depending on game)
SW-1A-137 3rd Ball Runway Switch (typical, depending on game)
SW-1A-138 Shooter Lane Switch
SW-1A-139 Lane Change Switch (can use SW-1A-150, usually stacked to right flipper or right EOS)

These were used throughout the Williams games: Examples are for the above include System 3, System 4 (Flash), System 6(a) (Firepower and Alien Poker), System 7 (Black Knight). Some appear on F-14 Tomcat (’87) and further, even after most Ball Ramp switches had changed over to micro switches. Black knight also used micro switch kits, which Williams provided as a field replacement kit for operators. Drop Targets also had factory micro-switch adapters fitted later to make them more reliable.

Switch Assembly Note: The Williams factory (or more accurately their suppliers) assembled many switches so one of the blades was facing the wrong way. It should have two gold contacts facing each other, where in fact one blade faces so that the rough-sided rivet is making contact. If you have an intermittent switch, it can be worth the time to de-solder it and remove from the game and then carefully pry the switch stack apart and reassemble the switch so the gold contacts face toward each other. Keep the exact order of spacers and just reverse the one blade. While there, carefully clean the switch contacts with naptha and a clean cloth. Again, please don't use emery (or files) on these switch points! Once the gold has worn, they form "dead spots" and won't work reliably. Another symptom of the misassembled switch stacks is multiple switch closures.

6.23.1 How the Switch Matrix Works

First of all, a switch is usually an electro-mechanical device with moving parts. This includes leaf blade switches, tilt switches (where the ring is fixed and the plumb bob is the moving part of the switch) and micro-switches. When a mechanical switch closes the contacts tend to oscillate before coming to rest. It literally 'bounces' several times before remaining closed.

The CPU could see the rapid open/close events as multiple switch closures, so there is usually some 'debounce logic' built into switch matrix reading programs. The CPU sees a switch closure and then 'checks back' to see if it remained closed as few milliseconds later. If it is closed, the CPU processes the event once, scoring the correct value or in some cases triggering a solenoid to fire. In the case of Williams 3-7 games, the switch debouncing for normal switches is done at the software level; unlike other contemporary manufacturers' method, Williams' software is more sophisticated in that the definition of switches include the ability to treat a switch as fast-response (example: spinner switch), normal response (lane rollovers - longer 'ignore time' to prevent duplicate reads on a slow traveling ball), and the longest time (trough switches - where the ball would travel over the first 2 switches on a multiball machine, you would not want those to react unless a ball were truly in the trough).

As an aside, an optical switch (or opto) has no moving parts and so doesn't suffer from mechanical switch 'bounce'. (An optical switch can suffer from electronic switch bounce - many rapid activations per second - which is why they usually utilize circuitry known as a Schmitt Trigger to lower the probability of debounce errors). It also shouldn't need adjustment or in theory wear out as quickly as say a micro-switch. However, an opto needs additional circuits and is not as simple to work on or replace as other switch types. Although you can not see an infrared transmitter with you naked eye, you can look at it with a digital camera. Optos were never used in System 3-7 games, although they were engineered and prototyped just at the end of System 7.

Wiring up every switch separately with multiple wires running to the backbox would be expensive in the amount of wiring and in the number of inputs required to the game's logic. Rather than do that, the engineers designed a switch matrix with 8 column wires and 8 row wires, creating a 'matrix' of 8x8 giving a total of 64 possible switches. Not all of these switch positions are used.

The first column of the switch matrix (COL 1) is dedicated to the same switches on the System 3-7 games. You can check in a manual, but these will be the cabinet tilts, the coin switches which sense coin drops, the credit button (to start a game) and the high score reset which is also in the cabinet (on the coin door). In order from row 1 through row 8, they are the plumb bob tilt, ball roll tilt, credit button, right coin chute, center coin chute, left coin chute, slam tilt, and high score reset button.

For the most part other switches will be on the playfield (exceptions could be a lane change or magnasave button/switch).

So how is the switch matrix read by the MPU? It uses the most useful Peripheral I/O device in this pinball era, the Motorola 6821 PIA (earlier this was the MC6820 PIA).

The 6821 is made up of two 8 bit ports, one port is known as 'Port A' and the other 'Port B'. Any of the 16 pins can be configured as inputs or outputs. That's exactly what's needed to drive our Switch Matrix! An 8-way output port B to 'send' (or strobe) down the Columns to the switches, and an 8-way input port A to 'read' the Rows from the switches. The Columns are known commonly as 'Drives' or 'Strobes' and the Rows are known as 'Returns' for this reason. In a similar way (an 8x8 matrix) another PIA is used for the Lamp Matrix.

The switch matrix PIA is at IC11 on the Driver Board (also called PIA II). This PIA doesn't drive the switch matrix directly from it's TTL output pins. There are +5v powered 'pull-up' resistors and 4 IC's which act as drivers/buffers helping to protect the PIA from damage if the switch matrix is shorted.

Two 7406 (Hex Inverter / Buffer with open collector HV outputs) is used to 'drive' the Columns.
A 7406 or 74LS06 will work and are common parts.
Two 4049 (or 14049 "CMOS Hex Inverter / Buffer) is used to 'read' the Rows.
The MC14049 or CD4049UBCN are common examples, although any 14049 or 4049 14-leg DIP will work.
(Hex in this context refers to six inverters contained in the one IC package. It's nothing to do with the computer (or math) Hex meaning of 'base 16'.

The CPU writes to the output port of the PIA driving column 1, then the return input port is saved, holding the status of all 8 rows in one byte. The program then 'looks' at each of the rows 1-8 in quick succession. It checks to see if the signal is getting through on the row that's currently being read. Think of it as an 8 bit answer for that column showing which switches in that column are closed. Say 00010001 returned would mean switches at C1,R4 and C1,R8 are closed.

So for every 'drive' down a column, the machine does 8 'reads' of the rows. Then the CPU moves on and drives column 2, looking again at each of the rows 1-8. It continues through all of the remaining columns in this way. This is done very rapidly by the CPU, strobing the whole matrix looping over and over many times a second. Because it's so fast, even multiple switch closures are rarely missed, in fact as mentioned above it has to debounce the results to obtain an accurate result. Williams' operating system software on machines from at least system 6 up through system 11 have differing debounce values depending on the type of switch, allowing greater sensitivity on certain types of switches, and desensitizing others (such as lock or trough switches).

The diodes on every switch help to steer the drives only to the row that's being read and not back into other parts of the matrix. That's why good diodes on switches are so important, as they must only allow current to flow in one direction: from column to row. Essentially that's how the switch matrix reads switch closures.

This article by Aeneas describing a later game's switch matrix may be useful as a different explanation with a diagram. Transistors can be used on the drives (as with modern games like DE or Stern ), different buffer chips can be used, but the basic switch matrix design hasn't changed over the years.

Finally a warning about protecting the switch matrix. Please note that nothing will protect the switch matrix IC's and PIA if (for example) the +50v solenoid voltage is shorted into the matrix. This happens if the 'lane change' switch is attached in a stack on the flipper mech and shorted to the EOS switch. In later designs the 'lane change' was moved to be stacked with the flipper buttons (nearly as bad) or to a separate button which was a better design. Be very careful if working under the playfield not to short the solenoid or lamp power into a switch. It can be a lot of work to repair the damage done to the driver board, so be careful with you screwdriver and loose solenoid power wiring. Most people who have been repairing games long enough learned the hard way, and won't work under the playfield with the power on.

6.23.2 Special Switches

There are a few switches that are dedicated, and not part of the switch matrix. The coin door Advance and Auto/Up Manual/Down switches, which let you start diagnostic tests when the switch matrix is faulty.

Also, there are a maximum of 6 special solenoid switches, which directly connect to the solenoid logic on the driver board. When Williams first designed these games, they were worried that the CPU couldn't always 'keep up' with the ball hits to the pop bumpers or slingshots. So during game play the switches on the playfield fire the special solenoids directly from the playfield. These usually include the spoon switches on the pop bumpers and slingshot stand-up switches.

There is a 22uF capacitor and a 100 ohm resistor mounted across the special solenoid switches. The use of the capacitor and resistor creates what is called an RC circuit. The RC circuit is used to filter noise from the switch signal as well as to ensure a minimum pulse length for the solenoid activation. If using a polarized capacitor the positive terminal goes to the tie point with one end of the resistor attached. Should a special solenoid lock on, and the switch leaves are properly gapped, the issue may be a shorted switch capacitor or resistor.

All pop bumpers and slingshots have a secondary switch which is part of the switch matrix. It closes when the coil is fired and tells the CPU to increment scoring and in some cases to trigger sounds for these devices. The scoring switch is mechanically closed, by the 'elbow' of the slingshot arm or by the bakelite yoke connected to the pop bumper ring.

Confusingly, the CPU can also fire the 6 special switches from PIA lines for the diagnostic tests, but these PIA signals are never used during game play. If all these solenoids work during diagnostic tests, but not in game play it points to the switch inputs or the 7408s at IC6/7. It won't be the 7402s at IC8/9.

I have seen a game that worked in play perfectly, but the diagnostics could not fire two of the pop bumper solenoids during tests. This was a Switch Matrix PIA with faulty output at pins 19 (CB2) and 39 (CA2), other than that the PIA was working correctly and so it could still be used to drive the Switch Matrix. If a switch row or column went out, the PIA would then have to be replaced.

Here is a list of the PIAs and where the pins are that fire the 'special solenoids' during diagnostics:
Special Solenoid Diag. PIAs
ST# PIA CHIP Pin # Location / Board
1 III IC10 Lamps 19 Middle PIA on Driver Board
2 III IC10 Lamps 39 Middle PIA on Driver Board
3 II IC11 SW Matrix 19 Right PIA on Driver Board
4 II IC11 SW Matrix 39 Right PIA on Driver Board
5 IV IC 5 Solenoids 39 Left PIA on Driver Board
6 I IC18 Displays 19 Via 1J1-26 on CPU Board

6.23.3 Testing the Switch Matrix

To test the switch matrix in the game, first remove both Switch Matrix connectors on the top right of the driver board. J2 is Column (green wires) and J3 is Row (white wires). Then run a switch test from diagnostics, you should get no switches being sensed.

Use an alligator test lead as follows:

  • Connect one end of the test lead's alligator clip to the column pin, starting at column 1. That's the bottom pin of J2.
  • Then use the other end of the test lead to touch the appropriate row pins.
  • Start at the bottom pin of J3, which is row 1. You should see switch #1 indicated (R1C1).
  • Move the probe to the next pin up on J3, which is row #2. You should see switch #2 indicated (R2C1).
  • When you got to the top of the row pins, move the clip end to COL #2 (up one pin) and start again with row #1.
  • Activate each switch in turn by connecting the appropriate 2 male pins on the CPU board with your test lead.

Using your switch matrix chart from the manual as a guide, you may find the faults as your game "sees" on the same switches.

If you get an error in sequence, more than one switch registers at a time or you are missing a row or column - then you know the problem must be on the MPU board. You can either try to fix it or send the board out for repair.

If the above test works correctly, meaning all switches register correctly then your problem must be the wiring or on the playfield.

6.23.4 Testing the Switch Matrix PIA

No section on the switch matrix is complete without mentioning the switch PIA. That IC and 4 buffer IC's are the only logic in the switch matrix. The complete instructions for testing all the PIAs are beyond the scope of this section.

The only sure-fire way to test the PIA is with the 'Leon Borre test ROM' in the MPU board. You do this by taking the MPU and Driver boards out of the game and putting them on the bench. This will also let you trace back through the circuit to find a fault more easily, and it's not as hard as you may think. It will also exercise the CPU memory and test all the PIAs on the Driver Board, not just the switch matrix PIA. Here are his excellent instructions:

Leon's Repair Pages for WMS System 3-6 games including a free test ROM.

Leon's Repair Pages for WMS System 7 games including a free test ROM.


These Belgium guys are knowledgeable about pinball.

Should any of the chips in the switch matrix prove to be faulty, be particularly cautious in replacing them. They have a greater than usual tendency for lifted traces and pads when desoldering. Heat rising from the lamp matrix resistors has a tendency to not only cause these chips to fail, but also to loosen the bond between the circuit board and the solder pads and traces on this area of the driver board.

6.23.5 Switch Matrix Components

Final thoughts about components which make up the switch matrix.

The passive components: On the driver board there is an RC network, made up of 16 x 4.7K pull-up resistors that you can measure with a DMM. Tolerance is not critical, but should measure between 4.5K to 5K which is about a +/- 5% range. There are also 16 x capacitors (470pF at 50v, ceramic) these should not measure as a short, if in doubt just replace them. And 8 x 1K ohm resistors at the switch inputs (rows) at R196-R203 these can not be open or shorted and must be around 1K each.

You can see the above chips inverting signals with the game running and a logic probe. Check if signals look weak or suspect. If you see a signal on the input side and then nothing inverted on the output side, then that's your problem.

Beyond that IC15-18, which are the Switch Matrix inverting buffer / drivers:
IC Location Inverting Pairs. Leg a to b listed as a,b Switch Row/Col Numbers IC Part Number /Eqivalent
IC15 3,2 7,6 14,15 9,10 in that order Rows 1-4 (MC)14049U / 4049U
IC16 3,2 7,6 14,15 9,10 in that order Rows 5-8 (MC)14049U / 4049U
IC17 2,1 6,5 12,13 8,9 in that order Columns 1-4 7406(S) / 74LS06
IC18 2,1 6,5 12,13 8,9 in that order Columns 5-8 7406(S) / 74LS06


J2 (Column) and J3 (Row) connector pins:
Connector Male connector Pins Switch Row/Col Numbers IC Location IC Part Number /Eqivalent
2J2 pins 1-3 and pin 5 Columns 5-8 IC18 7406(S) / 74LS06
2J2 pins 9-6 Columns 1-4 IC17 7406(S) / 74LS06
2J3 pin 1 and pins 3-5 Rows 5-8 IC16 (MC)14049U / 4049U
2J3 pins 9-6 Rows 1-4 IC15 (MC)14049U / 4049U

Except for the 6821 PIA which is getting harder to find, all the switch matrix components are readily available. However there are acceptable substitutes on the market. Any of the following part numbers are drop-in replacements for the 6821. 68A21, 68B21, 6521, W65C21N.

6.23.6 Wire Jumpers on System 7 Driver Boards

A Williams System 3-7 Driver Board, retrofitted with wire jumpers at R04-R211

Williams removed some resistors from the switch matrix inputs and used wire jumpers instead. This happened from the start of System 7. The first thing to do with a driver board (which may not be from that game) is to measure these resistors and make sure you have the the driver board set up for the correct game. While a board with wire jumpers can be used in System 4-6 games, using a board with resistors in a System 7 game will cause problems. Replacing the 330 ohm resistors with wire jumpers helps with switch sensitivity and with sensing more than one switch closure at the same time.

On driver boards from Black Knight and later System 7 games, there should be 8 wire jumpers (or zero ohm resistors) used on the switch matrix at positions W9-W16. These are located on the upper right hand corner of the Driver Board just to the left of J2, the top right connector which is the switch matrix column drive. To the left are two columns of 7 resistors each, the second column should be the wire jumpers, along with the top position of the next column to the left (column with 2 resistors only).

System 4-6 Driver boards, (games like Alien Poker, Firepower and earlier SS games) were released with 330 ohm resistors (orange, orange, brown) in these 8 locations. Starting with Black Knight and later (System 7 games) they are called R204-R211, and are zero ohm resistors (usually a tan body with one black stripe). You can replace or jumper over them with wire leaving the resistors in place. In all other aspects the driver boards are identical, so it's easy to convert between the two. This is why they are usually known as System 3-7 driver boards.

Before you start to replace parts on the driver board, be sure your playfield switches are working and the diodes are good. You need to unsolder one end of the diode from the switch to be able to test it correctly with a DMM on the diode setting. While you are there, clean the switch with a business card soaked in naptha or contact cleaner sprayed on the business card (with the game off, contact cleaner is flammable!) See if any blackness comes off on the card. You can also gently wipe the contacts until shiny with a Q-tip or the corner of a clean rag dipped in isopropyl alcohol. A pencil eraser also works well to clean switch contacts.

7 Game Specific Problems and Fixes

7.1 Black Knight

7.1.1 Magna Save Performance

There are 2 under playfield relays used to shunt high voltage to the magna saves. These are high voltage contacts and should be filed for maximum conductivity and power for the magnets.

7.1.2 Magna Save Magnets Not Working

If the magnets are not working in test or game, make certain that the white-red wire is tied to the ground screw in the backbox. This white-red wire is fed from the relay coils, which enables each appropriate magnets.

7.1.3 Fast moving ball flies through upper lock

Mount a shooter lane bracket underneath the plastic of the upper ball lock trough. With creative bending, all ball motion will stop on entering the trough. The ball will drop down ensuring credit is awarded for the lock.

7.1.4 Ball Lock Scoring Bug

There is a software bug in version L-3 and L-4 where instead of getting 5000 points with other balls in the lock you get 10,000, scored in increments of 5000. There is no current fix for this; it may have been intentional for the additional score (usually picked up by another player) but probably an oversight by the original programmer. There is a footnote in the operator's manual referencing this. It is unknown what the original intention was.

7.1.5 Right Ramp Gate Adjustment

It is erroneous to have the mystery rollunder gate score in one direction only; the switch should activate both going UP to the upper playfield as well as DOWN to the lower playfield. It can be a strategy in multiball to let it roll down the switch, multiplying the mystery value.

7.2 Firepower

7.2.1 Firepower Won't Start a Game

Firepower and other multi-ball games are infamous for being stubborn if there aren't three balls in the trough, or if one of the trough switches isn't closed. Additionally, multiball games usually have a shooter lane switch. If MORE than 3 switches are closed, the game will also not start.

My Firepower, in particular, is frequently upset if the right-most switch isn't closed, and that switch has an unusual actuator on it that doesn't work as well as a common rollover actuator. That switch is a little mangled, so it's the one that is most problematic. ts4z (talk) 18:12, 21 May 2014 (CDT)

7.3 Time Warp

7.3.1 Time Warp loses background sound occasionally

There is a software bug in Time Warp where the background sound (if enabled) will go mute at ball launch. This happens whenever the ball kicker launches twice, or when a ball drains without scoring any points. There is a fix for this identified as Time Warp L-3 however it is not officially sanctioned by Williams and so is not readily available.

8 Parts Substitutions & Replacements

8.1 Boards

NOTE: Kohout Enterprises has begun manufacturing circuit boards again.

9 Repair Logs

Did you do a repair? Log it here as a possible solution for others.

9.1 Unreliable +5V

A Firepower won't boot consistently. It was reliable before it was moved in a box truck. +5V supply measures low. For some reason, I decided the problem had to be the diodes in the power supply and replaced them, which didn't help. I replaced the +5V cap in the power supply, C16, and all was well. (Thirty year old filter caps are all absolutely suspect.)

Once I had the cap loose, I discovered I could shake it and hear it rattling inside. Presumably, the jelly roll of electrolytic was dried out. Shaking it might have the internal components in the right place, but probably not. A few miles in the back of the truck was enough to wreck it, more permanently before. The game has been very well-behaved since replacing the cap with a new one. ts4z (talk) 18:12, 21 May 2014 (CDT)

9.2 Unreliable +5V, part 2

All classic pinball boards rely on 5V logic. Most data sheets indicate 5V, +/- 0.25V. Always check the 5V on the power supply, AND on the MPU board. Note that there may be 5.2V on the power supply, but 4.8V on the MPU. This is too low.. small pulses can pull the 4.8 below 4.75 fairly easily and cause resets. If there is less than 5V on the MPU board, but the voltage is right on the supply, try checking voltage drops at each individual connector pin. One will usually find that pins will have built up tarnish, or heat has turned them grey or brown.. this is the tip off that the connector needs to be replaced. I know, groan.. it can be a pain, but replace the header in the circuit board. Make sure to remove or clip the key pin. Then replace the connector on the wires (female) with crimp and poke type. Use a ratchet crimper, and tug on them after each crimp to prove that they are good. Replace the wires one by one.. clip, crimp, poke.. copy each wire position from the original one one at a time, and keep in mind that the connectors have a ramp and go on one way. Replace and power up, and then go back to the meter, and verify that the voltage on the MPU board has risen..

9.3 CPU will not start

This happens to involve a Type 2 sound board, but it could apply to an MPU board or a Type 1 sound board. The reset pin voltage (CPU Pin 40) was low and would slowly climb after power on, but not enough to allow the CPU to start. All components in the reset section were good, and replacement parts did not make any difference. The problem was the sound board's PIA. Its reset line was pulling down the voltage so much that the CPU would not start. After replacing the PIA, the sound board booted and worked fine.

Retrieved from "https://www.pinwiki.com/wiki/index.php?title=Williams_System_3_-_7&oldid=24127"