Leon Borre Atari First Generation Repair

From PinWiki
Jump to navigation Jump to search

Repair Atari CPU board . Type A006020.

Machines; Space riders , Atarians, 4x4, Road runner

Time 2000, Neutron Star, Airborne Avanger

Swash4.gif

My special thanks to Harald , a German pinball-fan , who lended me his CPU board to test on .

Swash1.gif

Atari adapted their CPU board for each model, this means that not all of the lamp drivers , or coil drivers are always installed, depending on the number they need on the game there where more or less installed. This also means board changes between two Atari machines is not always possible , mostly some IC's and transistors must be added. Always compare the board you take out with the one you are going to use if you do a switch . Operators where asked to perform some changes on the board to improve the power -on sequence , not all board have this change , so check it out , you find what and how in the " pinball troubleshooting guide " Make it a general rule not to exchange CPU and Auxiliary boards between two Atari machines..

This means also that the coil drivers and the lamp drivers are on the CPU board. Atari has no switch matrix , every switch is directly checked with a game instruction using the specific switch address.

Some smaller changes excists between the schematics and the reality on the board.. You will find the "real" selection wiring of the game roms in this article...There may be other changes wich i did not see ....so be aware.

Here the Atari troubleshooting guide ................. Pinaqua.gif

Here the schematics ....................... Pinaqua.gif

The boards treated in this article have serial number A006020 followed by another number. We must have a number that is -06 or higher . Lower numbers are boards that are shorter and miss the game rom sockets E0 and E00. The first boards had no 1K game roms but rather small roms at emplacements E1 to E8. As the test eprom needs socket E0, we can only test boards with a number -6 or higher. The best board is -12 this type of board has all the coil drivers , all the lamp drivers and all the switch drivers present. Some good news , we only need 5 volts to test the board on the bench.

To do the try-outs of the test eprom I had two boards , one fom a Space riders and one from a Time 2000. One had the extra modifications , the other not, but the test run on both , so if you have a non modificated board you can leave that for the test it does not matter.

The original game roms are not pin compatible with a 2716 or a 2732. In case you want or need to replace the original game roms an adaption must be done . What to do find it here .........

In order to run 2716's instead of mask ROMs, cut the trace between pin 21 and pin 18 of E0 and/or E00, depending on which is to have the EPROM installed. After cutting the trace, install a jumper from pin 18 to pin 12 ( ground ).

For the test eprom i use a 2732 cheaper and more easy to find than 2716. To avoid that runs have to be cutted to adapt the eprom at the rom socket , i use an external adaption . The adaption of the test eprom is done by bending upwards two pins and grounding them.( see further)

Swash1.gif

Ata2volplaat.jpg

Here the cpu board , six rows of IC's numbered from A to F ( Row A is the lower row) On the lower row we see 6 empty places , these are lampdrivers that are not used in this Space Riders board and not present on the board.

Ata2misprom.jpg

Left on row E ( blue arrow) two 24 pins connectors for the game roms At the right of these 8 empty places where the older type roms where soldered .( oragne arrow) On the row below we see two missing IC's these are emplacements for extra memory rams. On this board there are only 2 of the 4 possible rams present.

Ata2mistri.jpg

At the blue arrow we see a missing coil driver transistor, the same at the row above some solenoid transistor drivers are missing as they are not used on Space riders . On top at the right corner the 5 volts stabilistor, the board receives about 10 volts on his input, brought back and stabilised at 5 volt. De orange arrow indicates the capacitor that is connected at the 5 volts , we will apply 5 directly there on that capacitor instead of 10 at the normal input, as mostly we have a stabilised power source of 5 on the bench , ATTENTION the 5 volt source must be able to deliver 3 amps.

Changes and adaptions

Examples:

Ataa1modif.jpg

Adaptions at IC A1. On the upper photo we see the adaption made by the operator, done wit blue wires. ( blue arrow) On the lower photo we see that there are two extra runs, here the board was already adapted in the factory. ( red arrow) .

Ataboardverschil2.jpg

Atamodif.jpg

An extra transistor added, ( red arrow ) This modification is in the cirquitry of the" total games played counter " Probaly done by the operator for reasons only he knows. Again we see at the blue arrow a whole row of IC's that are missing , these are lamp drivers not used in the machine .

Atadikvoet.jpg

Atahogevoeten.jpg

This board has very special sockets for the game roms and tyhe CPU IC.( blue arrow) Because the test eprom has two pins bended upwards it is impossible to insert the eprom in this socket... I needed to put in the socket first another socket with long legs , and then in that one the test eprom , the same for the CPU IC ( 6800) there to a pin is bended upwards during the test. To overcome this i first inserted another socket with "long legs"( see type at the red arrow) and in that one i could insert the test eprom. I used socket of 24 pins so two of these where a bit to long for the 40 pins CPU IC that is why there are some pins sticking out ...( white arrow) i did not cut them , perhaps i can or will use these sockets later for something else , you never know !!!

Swash1.gif

All these first comments make it clear before starting the repair we have to prepare things and we cannot start with the board " as-is" or just plug in the test eprom.On the Atari board there is no indication at all, nothing , no LED like Bally or Williams. All IC's are soldered ,memory IC included , only the CPU and game roms are socketed. Add to this two more difficulties. The watch dog cirquitry wich will reset the board when the game or test program does not run completely accurate .Secondly the DMA cirquitry ( direct memory acces) wich will interrupt any program test or game program because it is harware wired and cannot be avoided by programming. So the test will be interrupted every 16 milliseconds wich will have some effects on our measurements ...During this interrupt the memory contens ( some positions) is guided to the displays. A thirth harware interrupt comes disturbing our test program, to avoid this one, NMI ( non masqueble interrupt) we need to bend upward pin 6 of the CPU IC and put it at 5 volts. Perhaps all this sounds complicated, but do not worrie the test is easy and the test eprom takes care of these problems.Let's go at work !!!

Introduction

As always we have a test eprom that wil be plugged in at socket E0. The test will test all selections one afther another in an endless loop.During the test procedure we will check all coil driver outputs , lamp drivers , switch driver , check the watchdog and DMA cirquitry .A test-led will permanently indicate if the test program is running OK.The coil, lamp and switch outputs will be checked without using the test eprom, and are checked directly on the output connectors . ( The whole cirquit is tested, from source to output connector ) IF an output seems bad we use the test eprom to follow the signal flow from the source of the signal to the connector .

Preparation of the CPU board

We connect the 5 volts to the board , directly connected at the 5 volt capacitor, and ground .

Ata2aanslui5.jpg

The 5 volt connection at the capacitor.

Download the test eprom image here .... Pinaqua.gif This has to be burned into a 2732, BUT this IC cannot be plugged in into socket E0 , FIRST we have to bend upward pins 18 and 21 , both will be connected at pin 12 ( ground) Solder the connection at the very top of pin 12 as this pin still has to fit in the socket.

Atapin18en21.jpg

Here we see the test eprom inserted in E0 with his adaption pin 18 and 21 bended upwards and together connected at pin 12.

Last things to do is to connect a test-led between pin 40 of the CPU IC and ground.

The test-led is the same as in all other repair methods , and made like this :

Ata2ledopg.jpg

An ordinairy LED with two wires with at their end mini hooks , In series with the LED a resistor of 1000 Ohm. Before you uses the test led for the first time try it between 5 and ground to see wich side is the and wich side is the - side of the test LED. Put the red mini hook at the side the green ( or blue) at the - side.

Ataboardvolklaar.jpg

Here the complete board ready for test . The test eprom in socket E0,( white arrow) The red wire connects pin 6 ( bended upwards) of the CPU IC with 5 volts ,and at pin 40 the test led is connected (see the test led at the blue arrow)

To check the different outputs we do not use a scope, it is possible to use another test-led to check all signals.. I advise to use a second test-led that way the other one can stay connected at pin 40 to have a permanent control that indicates if the test is running ok..In this extra test led the series resistor better be a 470 ohm resistor wich will give us some more brightness .

Simple test led

Atatek1eng.jpg

Atatestled2.jpg

Swash1.gif

START .

ATTENTION !!

We start WITHOUT the test eprom in socket E0 , with no program the watchdog will reset the CPU constantly and the test led connected at pin 40 will blink. If that is not the case go to " Test led does not blink" .

It is ok the test led blinks if during the measuremants the test led stops just push on the reset button of the board to start it again, it sometimes happens.

We can measure now the outputs of;

The coil drivers , The test led is connected at 5 and we measure with the - side of the test led. For the coil drivers measure at connector J8 The test led will blink at every output that is ok. If there are outputs where the led does not blink or remains " on" all the time note the pin number for later use when we proceed with the tests eprom present. Also be aware that some outputs will have nothing if the output transistor is not present, you can clearly see wich ones are not on the board so ignore these outputs.

The lamp drivers , The test led remains connected at 5 volts . All outputs will blink the test led at the same rythme of the test led. Again be aware of missing IC's , easy to see wich connector pins have a IC present and wich not . Again note those who do not blink the test led or those where the test led remains "on" all the time. The outputs are at connectors J1,J2 and J3 Output IC's are A15 to A20 and B15 to B20.

The audio ouput.

The audio ouput. The schematic is not consistent with the reality. The schematic indicates that the audo outputs coming from D12 pins 9,10,11 and 12 should go to the J10 connector . in fact only pin 9 and pin 11 arrives.I made a design of the connector , numbered the pins , as they have no numer on the board, and made a list of what signals you should find WITHOUT and WITH the test eprom in place .

Ataj10.jpg

It are pins 8 and 7 solder side wich are the outputs of the audio IC D12 pins 9 and 11.

WITHOUT the test eprom :

Component side : pin 1 blinks . Pin 2 stays off, pin 3 blinks , pin 4and 5 faintly on. pin 6 blinks pin 7 pin 8 blinking in a irregular way.

Solderside: pin 1,2 and 3 blinks., pin 4and 5 always on. pin 6 ,7 and 8 blinks , pin 9 off , pin 10 blinks.

When using the test rom come back here and see the results with the test rom in place:

Component side: 1,3,4,5,6,7,8, all on pin 2 off

Solderside , 1,2,3,4,5,6,7,8 and 10 all on, pin 9 off

Swash1.gif

Use the test eprom

Now it is time to use the test eprom . Plug it in at E0, of course we have adapted the test eprom before using it ( pin18 and 21 grounded ) . Power up the board again, now the test led at pin 40 must remain "on" all the time no blinking, IF the test led blinks go to "Test program does not start " .

If the test led glows constantly this means the selection cirquitry is working ok. Nevertheless we check all ouputs and selections , the selections for rom, ram, switches, audio , watchdog etc..

The selection signals

The test led remains connected at 5 and we test at pins 4,5,9,10,11,12,and 13 of C6.:

Pin 9 is the watchdog ( wake up) the led will glow continously

Pin 5 is the rom selection, the test led will glow continously.

Pin 4 is de audio reset selection, the test led glows continously but faint.

Pin 10 is audio set , the test led glows continously but faint.

Pin 11 is the switch selection, the test led glows continously but faint

Pin 12 is the selection that activates the bit-latches , the test led glows continously.

Pin 13 is the ram selection the test led glows continously

If there is a signal missing there are two possibilities, IC C6 is bad , or there is a short on the missing output. Desolder the missing output pin and check again, if the signal is now present at the output pin you have a short on this output line, if the signal is still missing C6 is bad.

The schematic has not the changes applied to use the game roms at E0 and E00 . Here you find how the selection of the game roms is done , selection on the game roms is at pin 20.The start of the selection signals at C8 and D1 is where you start at the schematic. F0 and F1 are added an both 7400 IC's

Ataselschem .jpg

The bit latch signals.

The bit laches are C13 to C20

All bit latch outputs light up the test led very bright.The outputs are at pins 4,5,6,7,9,10,11 and 12 of all the IC's

If one of the bit laches has no outputs it is almost certain the IC is bad. All inputs for this 8 IC's are the same except at pin 13 ( the data bit) To be completely shure check if the data bit arrives at pin 13. If it arrives the IC is bad , if not check where the dataline is broken, it is certain the data is present at the output of the CPU IC the only thing between there and the arrival at pin 13 is a smal buffer ( 7404) wich can be found at emplacement BC8 or CD8 ( two letters here because the IC is between rows BC or CD ( the only IC's that are placed horizontal) use the schematic to see wich in and output pins are used for the bit you are missing.

The lamp outputs.

These where tested already. You have noted wich ones where missing or "on" all the time . For those wich where "on" all the time replace the ULN2003.

Go back to these wich had no output , now using the test eprom they have to be present all the time as there is a permanent signal for each output .

I give one example how to follow and find the missing signal:Use the schematic ( page 45)

J3 pin K has no output ( example) the output comes from IC A13 pin 15, the input at this IC is at pin 2.. Is the signal there at pin 2 and not at pin 15 then A13 is bad. If there is an input go to BIT latch bit 6, output latch 1000 , this is at pin 12 of IC C14, if nothing there look at the input at pin 13 ( still IC C14) If the signal is there replace C14. Stil no signal there go to pin 9/10 of IC CD8, there you must find the signal otherwise the test can't work ( missing data bit) As you had nothing at the output od CD8 you have to replace that one .

The switch outputs.

You find these at connector J6 and J7. Output IC's are F3 to F13; Again here several IC's are not present, at the schematic we can see wich IC's go to wich connector pin, so if ouputs are missing look first if the IC is present.Output pins at the F3 to F13 IC are: pins 1,2,3,4,5,6,7 and 9.

Output from F2 and F4 are tied to the dipswitches ,, their ouputs do not come at the connector but can be monitored at the left side of the switches if the switch is open at both sides if the switch is closed.

If all ouputs of the same IC ( F3 to F13 ) are missing. something can be wrong with the selection of the IC ( or the IC is bad) The selections comes from E11 look if the selection signal for the faulty Fx IC is present there , outputs at E11 are at pins 3 to 13. Missing an output, look at the input signals of E11 at pins 1,2,14 and 15. Missing one there look where it has to come from , if all inputs are there the IC E11 is bad.

The audio outputs.

Go back at the testing without the test eprom in place.

Swash1.gif

Test program does not start.

This means someting fundamental for the CPU or test rom is missing. The special sockets lose their pression and bad contacts can occur. Check with a ohm meter if the contact between the copper of the socket and the pin of the ic is ok. ( When one pin has a bad contact do not hesitate to replace the socket ( This was one of the faults i had on the Space riders board !)

REMOVE the test eprom , and power up without any IC in E0 or E00.

We check the minimum signals we need at the CPU IC..

Pin2, HALT, a pulsing signal ,measured with a volt meter 3 volts.The signal comes from trhe DMA cirquitry , if you mis the signal here go to the test wich describes this DMA cirquit.

Pin 3 clock signal about 2,5 volts .

Also check the external clock cirquitry ( page 38 ) The clock starts at the crystal and gate D10, follow the signal via C11, B10, B9, B11 and C9 and as last gates C10. The IC where the signal gets in and does not come out is bad.

Pin 5 VMA pulsing about 2 volts , an output signal of the CPU , if not there the CPU is bad BUT this signal can only occur IF there is a clock signal present OR if the halt signal ( pin2) is at 0 volt. So the clock and halt must be ok before we can judge the VMA signal.

Pin 8 .. 5 power

Pin 9 to 25 included . ( except pin 21 this is ground) These are the address lines, pulsing ,value between 1 to 2,5 volts. If one is missing bend upward the pin and check again ,is the signal present , then there is a short on that addres line, still missing the CPU IC is bad. Also check the buffers as the address signals A0 to A9 are buffered by D2, D4 and D3 these outputs to must give the signal.( use the schematic ).

Pin 26 to 33 included, these are the data signals. Value from 1 to 2,5 depending wich data line . The same way as the addres lines if one is missing bend upward, if the dat signal appears , there is a short on that line , if not the CPU ic is bad.

Pin 34 R/W signal .. 3 volts , an ouput signal from the CPU. Check afther previous signals are ok. If not present bend up pin and check again, like the address line ...

Pin 36 and 37 clock signals the same as pin 3.

Pin 40 reset .. 5 volt , pulsing as there is a regular reset by the watchdog cirquitry , but must come up to 5 in between resets at 0 by the watxch dog.

If all that is ok , leaves us with the selection of the test eprom on his pin 20. We need at signal there from about 2 volts volts... If not check selection cirquit use the schematic below , you need to find signals at al these gates.

Ataselschem .jpg

The test led does not blink .

If the test led is not blinking at the start of our testing where we are testing without the test eprom in E0. then you have a problem in the watchdog cirquitry. Follow the describtion here below " DMA and timing " as the watch dog signal is created there. You will find out that the watchdog signal comes from A10 pin 9, and passes then A2 where it leaves as RESET from pin 10.

Het DMA en timing circuit:

This whole cirquit consists of a few dividers one afther another. This results in the fact that the original Q2 clock and the DMA clock ( created at A6) are slowed down . At several point in this chain ,signals are used for the DMA int and the NMI and at last for the watchdog.

The Q2 pulses coming from D10 and the DMA clock generated at A6 are the base starters. The chain is composed of A6, A7, A8, A9 and A10, at several point you see signals leaving this chain to create the DMA int , the NMI and the watchdog signals , easy to check . The watchdog is normally coming out at A10 pin 9 IF there is no reset from the wake up reset at pin 2/3 . When we work without game or test rom there can't be a walke up reset and watch dog at A10 pin 9 must be there .

Extra informations.

The test-led is used to check the signals on the output connectors, when you have to follow a "missing" signal you have to use a logic test lamp or a scope . This because the test led will light up as wel on a steady o volt at the input/output of a gate , as with the pulstrain we are looking for ..

When using a scope to follow the signals , be aware that the outputs towards the connectors are done by transistor arrays wich have the "open collector" structure, so you must place a resistor ( 1Kohjm) on the output pin to watch the signal .

Another handy tool is a connector with 11 test led's and their resistor , with that tool you can watch half of the connector outputs at once , and when turning it 180� the other 11possible outputs.. The connector to be used is de easy to find JAMMA connector from and older video game.( Jamma edge connector 2x 22pins) See the model below.

Ataveelleds.jpg

The Jamma connector and the print with 11 test -leds

Atatestergreen.jpg

Here connected at the output connector of the coil drivers. The connector fits also the output connector of the lamp drivers the audio connector and the switch outputs.

Atatesterred.jpg

Written on the connector is RED and GREEN this indicates clearly wich color leds are at the side you are monitoring as the board connectors are at the component as well as on the solder side of the print.

There is no memory test , this was impossible because of the continous disturbing by the DMA cirquitry . In the worst case you need to replace both memory IC's . Perhaps in the future i will take some time to resolve this last problem.....

Swash1.gif

P.S The whole test progam consists of only 37 instructions and 6 wake -up reset instructions and is the smallest of all pinball test programs i wrote up to now. It took anout 10 days and 17 reburns to have the final image.