Capcom Repair

Note: This page is a work in progress. Please help get it to a completed state by adding any useful information to it.

For history of Capcom Pinball see Capcom Pinball.

1 Introduction

Capcom turned its hand to Pinball in 1995. The company brought its electronics experience from designing video games. The Capcom platform included a number of technological advances ignored by mainstream manufacturers. In particular, Capcom made extensive use of surface-mount parts and moved to the high performance Motorola 68xxx 32-bit processor.

2 Games

Title	Date of Release	Model #	Notes
Pinball Magic	10-1995
Airborne	3-1996
Breakshot	5-1996
Big Bang Bar	3-1996*		Saw pilot production of 14 games, never was released
Flipper Football	10-1996		Has adult language setting, it will cuss at you if this is enabled
[Kingpin]]	20-1996*		Saw pilot production of 9 games, never was released

3 Board Set

3.1 MPU Board

The "ahead of it's time" CapCom MPU, this one from BreakShot

4 Technical Info

Capcom used an electronic system that differs in several ways from most WMS and DE/Sega/Stern systems. Some major differences and noticeable design points:

• Separate switch matrix board with discrete wiring to each switch. The switch matrix diodes are not mounted on the switch but on the board instead.
• No batteries on the CPU board but an SRAM chip with a built-in backup battery (this battery will eventually wear out just like standard batteries).
• Adjustments, audits and diagnostics are done using the flipper and start buttons and the menu is entered upon opening the coin door (memory protect switch activated). There is no test switch.
• It features a diagnostic system automatically detecting lamp and switch failures.
• DMD high voltages are 110V DC (125V on WMS), 98V DC (113V on WMS) and 68V DC (62V on WMS)
• It supports up to 128 CPU controlled lamps (2x 8x8 lamp matrix). All lamps are CPU controlled; therefore there's no GI. Capcom pinballs do not have the burned GI connectors issue as is common on some other pinball manufacturer's designs.

The electronic system consists of the following parts:

• CPU board A0015403 (mounted behind the DMD).
• Power supply board A0015204 (upper board in backbox)
• Driver board A0015105 (big board in the middle of the backbox)
• Sound board A0015003 (most right in the backbox)
• Switch board A0015301 (lower board in the backbox)
• Display power supply board A0015502 (mounted behind the DMD)

Only exception is the Breakshot which uses a combined Power/driver/switch-board and a different sound board:

• CPU board A0015403 (mounted behind the DMD).
• Power/driver board A0017701 (big board in the middle of the backbox)
• Sound board A0021701 (most right in the backbox)
• Display power supply board A0015505 (mounted behind the DMD)

Breakshot uses several five lead solid state relays to drive the lamp matrix. So LED bulbs do not work properly. Incandescent bulbs work best on this machine.

Flipper Football uses a very large DMD display: 256 x 64.

5 Problems and Solutions

5.1 Power Problems

The power supply provides the following voltages:

+20A = 20V DC unregulated (up to 28V at no load); 20V AC comes in through fuse F2, goes through BR2 for rectifying, C5 and C6 are the buffer capacitors (15.000uF). LED2 shows the status of the +20A line.

+18 = 18V DC unregulated comes from the +20A 20V DC unregulated line above, the 20V DC passes through an 1N5402 rectifier diode which will cut approx. 1.2V from the 20V DC. +20B = 20V DC unregulated (up to 28V at no load); 20V AC comes in through fuse F1, goes through BR1 for rectifying, C8 and C9 are the buffer capacitors (15.000uF). LED1 shows the status of the +20B line.

+50A/+50B/+50C = 50V DC unregulated (up to 70V at no load); 50V AC comes in from J20, passes through fuse F4, goes through BR4 for rectifying. LED 4 shows the status of the line at this point after which it is split in 3 lines going to F11 for +50A (LED11 for status), F10 for +50B (LED10 for status), F9 for +50C (LED9 for status).

+50R/+50L = 50V DC unregulated (up to 70V at no load); 50V AC comes in from J20, passes through fuse F6, goes through BR6 for rectifying, C12 and C13 are the buffer capacitors (2200uF). LED6 shows the status at this point. It is then split in two going to fuse F13 for +50R (LED13 for status) and fuse F12 for +50L (LED12 for status).

+12V = 12V DC unregulated (up to 17V at no load); 12V AC comes in from J19, passes through fuse F5, goes through BR5 for rectifying. C11 is the buffer capacitor, LED5 is the status LED at this point. It is then split in two, going to fuse F7 for the +12V on J18 (LED7 for status) and fuse F8 for the +12V on J15 and J16 (LED8 for status).

+5V = 5V DC regulated (should always be between 4.9V and 5.1V no matter what the load is); 9.5V AC comes in at J1 and goes through fuse F3 to BR3 for rectifying. C1 is the buffer capacitor, LED3 shows the line status at this point. From there is goes to Q1 which is an LT1084CK-5 5V regulator (5A max.) after which the DC voltage should be pretty tight between 4.9V and 5.1V with a low ripple. LED14 shows the status at this point.

Troubleshooting: First to check whenever you have a problem are the LEDs on the boards. They give a perfect indication of a missing voltage. Note that when the coindoor is opened all 50V lines will be cut-off so LED 4-6-9-10-11-12-13 will be off unless you pull the switch out to diagnostic position (or close the door of course).

When one of the LEDs is off the first thing to check is the fuse, the best way to check the fuse is to take it out of the board and check it's continuity with a multimeter, testing it in the board can cause false readings and visible inspection is really unreliable. Note that some fuses can cause multiple LEDs not to light, a blown fuse F5 for instance will cause LEDs 5, 7 and 8 to stay off. Check the fuse holders as well, the clips can go bad over the years when they have become hot. They should be clean and fit tightly around the fuse.

When the problem isn't in the fuse it is probably caused by a broken bridge rectifier (BRx). These can go bad over time but often they go bad because of worn out buffer capacitors causing more strain on the bridge rectifier. It is wise, especially for the +5V line as that is the most critical voltage, to replace both the bridge rectifier and the appropriate buffer capacitor(s).

Check the 12V DC for ripple with a DMM. If there is a ripple present, it might be time to replace the capacitors.

Note there are 6 big capacitors 15.000uF/50V in the center of the power supply board. Capacitors C1 and C11 can be replaced with a 15.000uF/25V capacitor (14V DC) Capacitors C5, C6, C8, C9 however must have a voltage rating of atleast 15000uF/50V. The voltage I measured was at 34V DC when I had replaced them. Capacitors C12 and C13 are rated at 2200uF/100V (70V DC). The last capacitor is C10 and is rated at 100uF/100V (70V DC).

Another problem may occur when you remove the big capacitors. The original caps are click-in type. When you remove them, they might/will pull out the connection between the component and solder side of the board. Drilling a hole in strategic places and passing a wire through the hole will let you link the front and back side of the board again.

5.2 MPU boot issues

If the game does not boot, first check the voltages. 5Vdc and 12Vdc must be present. If these are low (less than 10Vdc) suspect the capacitors on the power supply board.

5.4 Solenoid problems

General Solenoids in Capcom machines are driven by an STP20N10L logic level N-channel MOSFET. When a solenoid problem occurs the software will report a solenoid problem in the diagnostics.

Locked on solenoid A locked on coil is often caused by a defective MOSFET driver on the driver board. With the power off, first test the resistance between pin 2 (middle) and 3 (right) of the MOSFET, disconnect the connector to the affected coil first so you only measure on the MOSFET. When the resistance is zero or close to zero the MOSFET is defective and needs to be replaced. Before powering on the machine after replacing a defective MOSFET, test the coil's resistance and check the tie back diode and it's connections because these are the main causes for a defective MOSFET. If the MOSFET seems to be good next to test is the input of the MOSFET on pin 1 (left) for a signal. When there's no pulse on pin 1 while testing the solenoid in diagnostics it could be a defective 74LS74 driving the MOSFET.

Non working solenoid A non working solenoid is often caused by a broken wire but other causes are possible. The easiest to check is the solenoid itself so check the resistance of the coil first. When the solenoid's resistance is very high or when it doesn't have any resistance at all the coil wire might be broken from the solder tab. When the solenoid seems ok check for continuity of the wires to the coil. When the wires check out fine as well the problem might be caused by a defective MOSFET driver. Check pin 1 (most left pin) of the MOSFET for a signal while the solenoid is activated in diagnostics, if there's a signal the MOSFET is probably broken, if there isn't any signal on pin 1 it's probably caused by a defective 74LS74 driving the MOSFET. After replacing a MOSFET, check the tie back diode and it's connections because in this case this is the main cause for a defective MOSFET.

Repeatedly broken MOSFET driver When a defective MOSFET is replaced but fails again after a short period of use, or even the first solenoid activation, test the coil's resistance and check the tie back diode and it's connections because these are the main causes for a defective MOSFET. This also occurs when a wrong type of MOSFET is used, a non logic level MOSFET (IRF540 for example) will not be driven correctly because the gate voltage will be insufficient to fully drive the MOSFET causing it to get too hot and fail.

STP20N10L MOSFET alternative In case of a defect STP20N10L MOSFET, this part is getting scarce but it can be replaced by the IRL540N which is readily available. Do NOT use the IRF540N instead (not logic level)! Please note that these MOSFETS are quite sensitive to static electricity.

5.5 Lamp problems

General There can be up to 128 CPU controlled lamps in two seperate 8x8 matrices called A and B. The lamp columns are driven by VN02N solid state relays and the rows by STP20N10L logic level N-channel MOSFET's. The MOSFETS are more likely to fail then the solid state relays.

Locked on lamp row Whenever you have a locked on lamp row it is likely to be a failing MOSFET driving that row but it can also be the 74LS74 driving the MOSFET. Check pin 1 (most left pin) of the MOSFET for a signal while the pin is flashing all lamps. Suspect the 74LS74 driving the MOSFET when there's no signal (permanent 0V or 5V) on pin 1 of the MOSFET. If the signal seems good you can check for ground on pin 3 (most right) of the MOSFET, if the ground is good as well you probably have a defect MOSFET.

STP20N10L MOSFET alternative In case of a defect STP20N10L MOSFET, this part is getting scarce but it can be replaced by the IRL540N which is readily available. Do NOT use the IRF540N instead (not logic level)! Please note that these MOSFETS are quite sensitive to static electricity.

5.6 Switch problems

In general switch matrix problem resolution is similar to other brands of pinballs but for a part it's a bit easier to work on, unfortunately for the bigger part it's harder to work on. Because of the discrete wiring from the switch matrix board to the switches and the diodes mounted on the board there aren't any diodes on the switches so it's impossible to put the wire on the wrong side of the diode when replacing a switch. However, the wires to the switches are quite thin and are known to break easily, the switch matrix board is mounted very low in the backbox and it's position doesn't make it easy to reach when diagnosing, the LM339's on the switch board are SMD (surface mounted) and sometimes their solder joints crack causing bad contacts.

5.7 Display problems

The display power supply board is based on a switching power supply built around an LT1271CT switching regulator from Linear Technology and a transformer. There are little parts to fail on this power supply. Problems can occur due to cracked solder joints but it rarely ever happens. Although unlikely to fail the LT1271CT is still available but it seems to be getting scarce.

A service bulletin (95-017a) regarding Pinball Magic was issued about the display power supply to solve some issues. The originally used 1N4936 rectifier diodes on D4, D5 and D6 are too slow and need to be replaced by faster rectifier diodes (Motorola MUR160). Capacitor C6 should be removed. Resistor R6 should be changed from 47k 1/2W to 15k 1W.

The service bulletin is available on IPDB: [1]

The MUR160 is available from On Semiconductor and the available MUR160G is fully compatible with the previous MUR160.

The display power supply board can develop burnt holes in it.

Holes burnt in board

5.8 Sound problems

Cracking sound When the cracking sound comes from one speaker only that speaker is probably defective as all speakers share the same amplifier output.

No sound Although it shouldn't be happening often a common failing component on other sound boards is the TDA2030 amplifier and these are used on Capcom's boards as well. The TDA2030 (U4) can be easily recognized because of the heat sink it is attached to. A defective TDA2030 will cause none of the speakers to work as a single amplifier is used for all speakers.

TDA2030 alternative The TDA2030 is an amplifier in a Pentawatt package delivering 14W of output power. You can use the readily available TDA2030A as a replacement which is equal but provides 18W of output.

5.9 Flipper problems

Flipper assemblies are Capcom's own design with some similarities to modern Stern Pinball flipper assemblies. If the flipper coil stop is worn down or damaged, drill out or punch out the worn stop from the bracket. A Gottlieb # A-4862 replacement coil stop can then be used in the hole to restore the bracket to usable condition if a Capcom # A-00378-1 coil stop bracket cannot be located.

5.11 Slingshot problems

A weak slingshot can be caused by a broken bracket. Due to metal fatigue the bracket can crack allowing arm play making it act weak. This situation won’t last long as the crack will get bigger and the bracket will eventually break causing the arm to completely separate from the bracket. A good replacement for the Capcom slingshot bracket (MT00338) is the Data East version (515-5339-01); all the parts can just be transferred to the new bracket.

While you're at it: Check the link for wear to see if it needs replacement (WMS 03-8085 or Data East 545-5293-00 will do).

Check the plunger to see if it needs replacement. When it is mushroom shaped because of hitting the coil stop it is time to be replaced (Data East 515-5337-00, includes link); when there’s just a thin collar around the end, you can remove that with a file or mini grinding tool as it will cause friction and binding in the coil sleeve.

Replace the spring (Data East 266-5020-00 prefered or WMS 10-128).

Replace the coil sleeve (WMS 03-7066 or Data East 545-5411-00/545-5031-00)

6 Strange behavior

Locked on lamps and/or switches not working Sometimes this can be easily fixed by re-seating the ribbon cable running from the CPU board (behind the DMD) to the driver board. Simply taking the connectors off and putting them back on, on both sides of the cable, will clean the contacts a little and that might be enough to have good contacts again and solve that strange behavior.

7 Game Specific Problems and Fixes

Capcom used tiny connectors to the ball trough opto boards. Carefully wrap a small cable tie around it and it stays put better.

Capcom also used too fine a wire for switches. Easy to have one break off, which in turn can make that switch and others not work. First thing to check for when a switch doesn't work.

8 Repair Logs

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

8.1 Breakshot several feature lights out

Ball rack 1 through 7 lights not lighting and Bank Shot arrow not lighting (a total of 8 lights and all on the same circuit). Found diode D2B (type 1N4004) on the power driver board was open. Replacing the diode restored all the affected lights to proper operation.