Summary: Upgrade System3 CPU boards to run System4 and System6 game ROM software. Also minimize the number of ROM sockets that need to be replaced on System3, System4 System6 and System7 CPU boards.

In the previous section, we talked about replacing circuit board sockets. But before doing that, please check this section out. It may give some good insight on some System3 and System4 CPU sockets that do *not* need to be replaced (because they won't be needed).

Remember, until the the CPU board is running properly, it is a good idea to remove the lamp and solenoid fuses F2 and F3 from power supply board. This will minimize any risks of burning coils or lamps.

CPU Board component placement for System3 and System4 (Click here for a larger version, and click here for a schematic).

System3 CPU Board ROM Upgrade.
Most System3 games' CPU board uses four ROM sockets: IC22, IC21, IC20, IC17 (from left to right). There is a potential for a fifth ROM socket at IC14 too (but this socket, from the factory, is not installed in a System3 CPU board, but the solder pads are present). Since we have already discussed replacing all the CPU sockets, there is a trick that will allow us to replace fewer sockets! That is, to install a socket at IC14 (where there is currently no socket), and use a 2716 (2048 byte) game EPROM there. This will replace the two 3624 masked game ROMs (512 bytes each) at IC22 and IC21. This is essentially upgrading the System3 CPU board to a System 4 CPU, allowing it to be used for any System3 or System4 game. Doing this means the sockets at IC22 and IC21 will not be used, and do *not* need to be replaced.

System3 CPU Board modification to use a 2716 EPROM at IC14. Note the cut to the trace coming from IC15, between pins 6 & 7. And note the round solder pad is jumped to IC15 pin 1.

In order to complete this modification, the chip selection circuit also needs to be modified. Here are the steps:

• Find chip IC15. This is directly to the lower right of the 40 pin PIA chip closest to the battery holder.
• Between pins 6 and 7 of IC15, there is a trace that goes to a round solder pad. Cut this trace to the left of the round solder pad.
• Connect the round solder pad trace to pin 1 of IC15 with some wire wrap.
Now that the chip selection circuit is set to use IC14, it is just a matter of getting a new EPROM "burned" to replace IC22 and IC21. There are lots of people that offer this service for around $10 to $15 for this chip. Williams' web site also has the binary ROM images for these EPROMs available for free.

Converting a System3 CPU to System4.

System 4/6/7 CPU Board ROM Upgrade.
Most System 4, System6 and System7 CPU boards uses five ROM sockets: IC22, IC21, IC20, IC17 (from left to right), and sometimes IC26 (mounted above IC22). There is a sixth ROM socket at IC14 too. Again, just like in System3 CPU boards, IC14 can use a 2716 EPROM, which will eliminate the 3624 masked ROMs (512 bytes) at IC22, IC21 and also IC26! But in the case of System 4/6/7 CPU boards, the modification of IC15 discussed above is already done.

Again, since we have already discussed replacing all the CPU sockets, this trick will allow us to replace fewer sockets on a System 4/6/7 CPU boards too. That is, using a 2716 (2048 byte) game EPROM at IC14 will replace the *three* 3624 masked game ROMs (512 bytes each) at IC22,IC21,IC26. Doing this means the sockets at IC22,IC21,IC26 will not be used, and do *not* need to be replaced!

Since the chip selection circuit is already set to use IC14, it is just a matter of getting a new EPROM "burned" to replace IC22,IC21,IC26. There are lots of people that offer this service for around $10 to $15 for this chip. Williams' web site also has the binary ROM images for these EPROMs available for free.

A Firepower System6 CPU Board jumpered to use all six ROM sockets, including 512 byte PROMs at IC26,IC22,IC21. This configuration was specific to Firepower only, with it's original six ROMs installed. The "green" Flipper ROMs at IC20,IC17 are standard. The "gameROM" at IC14 is a 2316 ROM (or a 2716 EPROM).

System6 CPU Board jumper for a 2716 EPROM at IC14, and no PROMs at IC21,IC22,IC26.

System 6 CPU Board ROM Upgrade.
As with System3/4, System6 CPU boards can also use a single EPROM at IC14 to eliminate the 512 byte PROMs at IC21, IC22 and IC26. In this case, a single "jumper" change is required to use an EPROM at IC14. Just above the IC22 ROM socket, there are two jumpers called "J4" and "J3". To use an EPROM at IC14, make sure jumper J3 is installed, and jumper J4 is removed. See the picture above. Note this modification applies to all System6 games EXCEPT for Firepower (see below).

Firepower System 6 CPU Board ROM Upgrade.
Firepower is unique amoung the System6 games, in that it uses all the ROM sockets IC14, IC17, IC20, IC21, IC22 and IC26. But it too can eliminate the usages of IC21, IC22 and IC26 by using the following instructions. For Firepower only, these steps modify a System6 CPU board's IC14 socket to use a single 2732 EPROM (unlike other System3 to System6 games, which can use a single 2716 EPROM at IC14), eliminating the PROM chips at IC21, IC22 and IC26. Please see the ROM section of this guide for instructions on this modification.

CPU Board component placement for System 6 (Click here for a larger version, and click here for a schematic).

Flipper ROMs at IC17 and IC20 (all System3-7 CPU Boards).
Location IC17 and IC20 on all System3, 4, 6, and System7 CPU boards can use either 2716 EPROMs, or 2316 masked ROMs. These are the flipper ROMs (essentially the BIOS for the CPU board). These sockets will need to be replaced too. But the masked ROMs or 2716 EPROMs can be used without any jumper modifications.

I highly recommend installing new EPROM versions of the flipper ROMs. The old ROMs often get "bit rot", and the silver legs tarnish and easily break. It's just a good idea to replace these old ROMs with new EPROMs.

Remember, there are different flipper ROMs for each system revision level. They are coded by color, one color for each system of games (more or less as they did overlap). Basically the colors decode like this:

• White = System3
• Yellow = System4
• Green = System6
• Blue = System7

The big exception to this rule was World Cup. This game used one unique white flipper EPROM, different than the standard white flipper ROM at IC17. Without this special flipper EPROM, World Cup will not boot.

Also remember the game ROM (IC14) *must* match the correct color flipper ROMs. And some games had a different game ROM for different flipper ROM colors. For example, Flash, which had a very long production, has a yellow flipper game ROM (system4), and a green flipper game ROM (System6). This happened because Flash's production was so long, some Flash games used System6 boards. Now it does not matter which game ROM is used. That is, a Green (system6) Flash game ROM can be used in a System4 CPU board (or vice versa). But the Green Flash game ROM *must* be used with Green flipper ROMs, and a Yellow Flash game ROM *must* be used with Yellow flipper ROMs!

System3 CPU Board Reset Modification.
During the production of Disco Fever (the last System3 game), Williams made a change to the reset circuit to increase its reliability. It is a good idea to do this modification on System3 CPU boards:

• Remove capacitor C27.
• Remove resistors R30 and R40.
• Add resistor R96, a 10K ohm 1/4 watt resistor, from the left solder pad of resistor R30, to the banded side (top) of diode ZR1.

Below is a diagram from Williams of this modification. The three red components are the ones to remove, and the blue component is the resistor to add.

Reset modification for System3 CPU boards. Red components to be removed, blue component added.

Early System7 "Mystery" Chip next to IC33.
On very early system7 CPU board, there was an unmarked 74125 chip located above IC33. This chip is not needed, and Williams removed it from the system7 design. If there is a 74125 installed above I33, it can go bad and lock up the CPU. If your CPU has this chip, it is suggested that you cut it off the board.

ROM Software.
Williams has System3 to System6 ROM software available at www.pinball.com/tech/sys6roms.html. Unfortunately, some of these ROM files are *incorrect*, and in ROM formats that are not commonly used today. Because of this, I have known good and working software available here, ready for burning into standard EPROMs. All ROM files below use 2716 or 2532 EPROMs (file names ending in ".716" are 2716 EPROM binary files, and those ending in ".532" are 2532 EPROM binary files), except for Firepower which uses a 2732 EPROM (".732"). All the download files below are in ZIP format (use PkZip or WinZip to unpack the files).

The "Flipper ROMs" Explained.
A lot of time people will ask, "what are the 'Flipper ROMs', and is that why my flippers are not working?" The short answer is, no, the Flipper ROMs have nothing to do with "flippers". What the flipper ROMs are is the Operating System for the game. Originally it was thought the term "flipper ROMs" were named to distinguish them from the "shuffle ROMs", since shuffle alleys used the same CPU/driver board, but alledgedly different operating system ROMs. But according to Larry Demar, this story is untrue, as the shuffle alleys did in fact use the same "flipper ROM" operating system as the pinball games. So the origins of the name "flipper ROMs" is unknown.

The flipper ROMs are identified by color: white, yellow, green and blue. A different color denotes a different version of the pinball's operating system. The color originally referred to the actual color of the label on the CPU ROM chips themselves, and was probably used when the games were manufactured for easy identification.

One color for each system of games (more or less as they did overlap for Flash), but basically white=system3, yellow=system4, green=system6, and blue=system7. The big exception to this rule was World Cup. This game used a unique flipper ROM2 at IC17, different than the other four color groups of flipper ROMs. Without this special flipper ROM, World Cup will not boot (also it should be noted that Flash has two versions of IC14 Game ROMs, one for yellow flipper ROMs and one for green Flipper ROMs, because Flash's production was so long it was made with system4 and later system6 CPU boards).

There were two flipper EPROMs on the CPU board at IC17 and IC20 (both 2716 EPROMs or 2316 masked ROMs, except on System7 which had one 2532 EPROM at IC17). Note in many System7 games IC20 was mislabeled (make sure the 2532 EPROM goes in IC17, and the 2716 EPROM in IC20).

THE CORRECT COLOR FLIPPERS ROMS MUST MATCH THE GAME ROM. As Williams pinballs evolved, so did the operating system Flipper ROMs. So a Firepower (a green Flipper ROM system6 game) can not use yellow (system4) Flipper ROMs! This is because as the games evolved, the bookkeeping and adjustments advanced and added new items. These new features led to a new version (color) of the Flipper ROMs (the operating system of the game).

Note some software included here may not be for the same system of CPU board installed in a game. For example, a Flash game may have a system4 CPU board, but "green" system6 software is included here. This does not matter. That is for example, all System3 CPU boards should have already been upgraded to System4/6 specs (as outlined in the CPU Upgrade section). This should be done so that a system3 CPU board can run system4 or system6 software (as discussed previously, system3 and system4 CPU boards are upward compatible to system6, except in the case of Firepower).

It is helpful to remember the following Flipper ROM color codes when dealing with game software: White = System3, Yellow = System4, Green = System6, Blue = System7. Flash spanned multiple systems when it was produced, being made with both system4 and system6 CPU boards. Therefore Williams had both yellow (system4) and green (system6) Flipper ROM game ROM versions at IC14 available for Flash. The newest version should always be used (in this case, Green Flipper ROMs). When it comes to game software, the general rule of thumb is, "newer is better". So make sure the CPU board used is upgraded as documented in the CPU Upgrade section. Also it should be noted that there were only two games made with yellow flipper ROMs: Stellar Wars and Flash (but the yellow Game ROM version of Flash is not included here, as the newer green flipper version is instead). Also Flipper ROM1 is at IC20, and Flipper ROM2 is at IC17.

Williams has System7 ROM software available at www.pinball.com/tech/sys7roms.html. These files appear to be good and correct. Star Light ROM files thanks to Clive Jones.

Psuedo 7 Digit scoring on 6 Digit displays.
If you have a game that uses Green flipper ROMs (Gorgar, Tri-zone, Time Warp, Lazer Ball, Firepower, Blackout, Scorpion), you can use a custom Green flipper ROM2 at IC17. This custom Green flipper ROM divides the score by 10. This essentially makes the game have 7 digit score, but using the stock 6 digit score displays! If the current scoring "rolls over" too easy for your play, this is an easy way to fix that (just burn one 2716 EPROM). Download the Psudeo 7-digit scoring.

Using the Above Firepower EPROMs Files.

Firepower was an unique System6 game in that it used the four ROM sockets at IC21, IC22, IC26, IC14 for holding game's software. Three of these ROM chips (all but IC14) were 512 byte 7641 chips, and IC14 was a 2716/2316 ROM. Using the additional ROMs allowed Firepower to have extra space for the game program (in addition to the 2716/2316 green Flipper ROMs at locations IC17, IC20). The original configuration of three 512 byte PROM chips at IC21, IC22, IC26 and a 2316/2716 at IC14 required that the System6 CPU board jumper J4 be installed, and J3 be removed (there were orange labels in the cabinet and on the CPU board noting this).

The three 512 byte PROMs used in Firepower at IC21, IC22, IC26 are troublesome. In addition their sockets may need to be replaced. A better solution is to use a single 2732 EPROM at U14, eliminating all the 512 byte ROMs and their sockets! The ROM data file documented here accomplishes this, but some System6 CPU board modifications must be made. Note this System6 modification was originally designed by Duncan Brown. Note a System6 CPU board modified in the described way will no longer work in any other games unless the modifications are "undone" (but the Leon test EPROM *will* work in a Firepower modified System6 CPU board!)

If Firepower is to be run in a System7 CPU board, please see the section below. Using the Firepower EPROMs in a System7 board requires no modifications to the System7 CPU board, but it does require a different set of EPROMs than the System6 version.

These instructions and version was created by Duncan Brown. Note there is also another version of this Firepower combo ROM concept out there called the "ZIG chip" (I believe developed by Tom Callahan). It is basically a slight hardware deviation of the same theme. The difference between the two methods is that the Brown method uses two diodes and a pull-up resistor, where the ZIG method just uses straight jumpers. Technically speaking, the ZIG method electronically should *not* work, but it does because of the nature of LS TTL chips. Most people use the ZIG Firepower method as it is simplier (though if the 74LS139 chip was replaced with a 74139, the ZIG method may no longer work!)

Steps to modify a System6 CPU board a single 2732 Firepower EPROM at IC14:

1. Verify that the System6 or System6a CPU board to be modified is currently working! DO NOT attempt this modification on a non-working or unknown system6 CPU board.
2. Verify that IC15 is a 74LS139. If this chip is a 74139 (no "LS"), the "Duncan Brown" method will need to be used instead of the "ZIG" method documented here.
3. Remove all Game ROMs (IC14, IC21, IC22, IC26) from their sockets.
4. On the solder side of the CPU board at IC14 pin 21, cut a notch in the heavy circuit trace on both sides of the pin. Removing some of the existing solder on the trace may help. The idea is to get rid of the +5V connection from pin 24, while still leaving an intact solder pad at pin 21. Use a DMM and make sure pin 21 is isolated and disconnected from pin 24 of IC14.
5. On the solder side of the board, run a small jumper wire from IC14 pin 21 to interboard connector pin 34 (seven interboard connector pins from the right). This brings address line A11 to chip IC14. Be sure to count the interconnector pins correctly, as the right most solder pad actually holds *two* pins, not just one!
6. On the component side of the CPU board, install two jumpers or zero ohm resistors at jumper locations J3 and J4, just above ROM socket IC22.
7. Program a 2732 with the ROM data listed above and install at IC14. If the game doesn't already have working 2716 green "Flipper" EPROMs at IC17 and IC20, of course install those too.

Modifications to the solder side of a System6 CPU board to use the 2732 Firepower EPROM at IC14. Note IC14 pin 21 is cut from the large trace to the right and left of it (blue lines in the picture), and then a jumper wire is run from IC14 pin 21 to the interconnector pin 34.

In case there is a need to restore the board to standard System6 condition, do the following. Note if using Leon's test EPROM on a Firepower modified System6 CPU board, there is no need to "undo" the above changes, as Leon's test EPROM works fine with a modified System6 CPU board (since the above change only affects EPROM IC14).
• Bridge a piece of wire with solder between IC14 pin 21 and pin 24, which is the heavy +5 volt trace.
• Perform whatever J3/J4 jumpering and chip installations are required for the game the board is being used.

The Duncan Brown Method.
If for some reason the above "ZIG" method does not work, or IC15 is a 74139 (and not a 74LS139), a couple changes need to be made to the above modifications:

• 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, they should line up at just the right spacing to solder one lead to each of those pins. Again, make sure not to short 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.
• On the component side of the board, use 1N914 or 1N4148 diodes instead of jumpers at J3/J4, with the banded (cathode) end should be towards the TOP of the board for both diodes. This allows either of the two Game ROM address ranges to drive the active-low chip select.

System6 CPU Board Jumpers.
The are six jumpers on the System6 CPU board:

• J3/J4 (to the right of IC15). These two jumpers determine the addressing of the ROM sockets at IC21,IC22,IC26.
  Default configuration with EPROMs installed and no PROMs: J3 INSTALLED, J4 REMOVED (except for Firepower, see above).
  If original PROMs are used at IC21,IC22,IC26: J3 removed, J4 installed (this is the factory default installation for Firepower using PROMs).
• J1 (to the lower left of IC1) - default configuration: J1 INSTALLED. With J1 installed (and resistor R4 removed), the CPU can use either a 6802 or 6808 CPU processor. Having J1 removed and R4 installed (4.7k ohms), the internal RAM of the 6802 CPU is used. This disables the 6810 static RAM at IC13, allowing IC13 to be removed. If using a 6808 for IC1, J1 must be installed and R4 removed, and the 6810 RAM at IC13 must be installed, as the 6808 chip does not have onboard RAM (like the 6802).
• J2 (to the right of IC30) - default configuration: J2 REMOVED. J2 jumpers the IRQ request outputs of the PIAs to the IRQ port of the CPU chip. This jumper is not installed because the Williams programming method is a "polling" scheme, and not an interrupt driven scheme. If they had redesigned the software to poll the PIAs, then they had this jumper available to implement this programming method.
• J5 (to the right of IC11) - default configuration: J5 INSTALLED. J5 enables the Memory Protection circuit. Removing J5 disables the memory protection circuit. MIGHT BE HANDY FOR SYSTEM3/4 GAMES.
• J6 (to the right of IC11) - default configuration: J6 INSTALLED. If J6 removed, this would halve the interrupt circuit timing from 1msec to .5msec, allowing for flexibility in using the board in another type of game.

A Firepower System6 CPU Board jumpered to use all six ROM socket, including 512 byte PROMs at IC26,IC22,IC21. This configuration was specific to Firepower only, with it's original six ROMs installed. The "green" Flipper ROMs at IC20,IC17 are standard. The "gameROM" at IC14 is a 2316 ROM (or a 2716 EPROM).

System6 CPU Board jumper for a 2716 EPROM at IC14, and no PROMs at IC21,IC22,IC26.

System7 CPU Board Jumpers.
After downloading the above EPROM files for any System7 games, the CPU board jumpers may need to be changed to reflect the EPROMs being used on the CPU board. Williams System7 CPU boards were used in games from Black Knight (1980) to Star Light (1984). All System7 CPU boards should use the same jumper settings for the EPROM chips and same size ROM chips, except for Hyperball, Defender and Star Light (these games require different System7 CPU board jumpers).

System7 CPU board jumper locations. Picture by Ray.

If swapping a CPU board from Hyperball/Defender/Star Light into any of the other system7 game (or vice-versa), the CPU board ROM jumpers need to be changed according to the table below (and the EPROMs will need changed too). The following table shows the setting for each jumper on the CPU board, dependant on the installation game (basically Hyperball/Defender, or any other System7 game).

Note the following rules apply:

• A setting of IN indicates jumper is installed.
• As setting of OUT indicates jumper is removed.
• Jumper settings left blank ("-") are unknown, or irrelevant.
• Jumpers is BOLD are the jumpers typically changed when converting a CPU board from Hyperball/Defender/Star Light to any other System7 game.

Changing CPU jumpers from a System7 Hyperball/Defender/StarLight to a System7 Black Knight. This includes installing jumpers 3,10,26 (blue wire wrap) and removing jumpers 5,27,28 (red).

Changing CPU jumpers from a System7 Hyperball/Defender/StarLight to a System7 Black Knight. This includes installing jumper 6 (blue wire wrap) and removing jumper 24 (red).

Using a 2732 EPROM in a System7 CPU board at U17.
This will change a System7 CPU board from a 2532 to a 2732 at U17. Most system7 CPU board using the blue flipper ROMs have a 2532 at U17 and a 2716 at IC20, and two 2716 game ROMs at IC14 and IC26. 2532 EPROMs are sometimes hard to find, so many people want to avoid using them. A 2732 EPROM can be used instead of a 2532 at U17 if the following cuts and jumps are made:

1. On the component side of the CPU board, remove jumper w22.
2. On the component side of the CPU board, install jumper w23.
3. On the solder side of the CPU board, cut the solder trace (5 volts) going to U17 pin 21 (this is the large trace directly above U17 pins 13 to 24).
4. On the solder side of the CPU board, install a jumper wire from U17 pin 21 to the w22 jumper pad closest to U17 pin 1 (the pad with "22" next to it).

Make sure to use the right w22 pad in step 4 above, otherwise it may short U17 pins 18 and 21. Also make sure to cut the trace clean in step 3, otherwise you may short 5 volts to ground.

The modification to a System7 CPU board to allow a 2732 at U17 instead of a 2532.

System6/7 Sound Board Jumpers
Applies only to the Sound board and not to the Speech board.
Table data from Williams

Jumpers Used	Sound ROM Type	Board Format	Games Used In
W2,5,7,9,10,15	2K x 8 2516/2716	Sound& Speech	Pinball: Gorgar, Blackout, Firepower, Black Knight, Jungle Lord, Pharaoh
W1,2,5,7,9,10,15	2K x 8 2516/2716	Sound only	Pinball: Defender, Solar Fire, Barracora, Stargate, Cosmic Gunfight, Varkon, Time Fantasy. Video: Defender.
W1,3,5,7,10,14,15	4K x 8 2532	Sound only	Pinball: Hyperball*.
W1,3,4,5,7,10,15	4K x 8 2532	Sound only	Pinball: Joust Video: Joust, Robotron, Bubbles, Sinistar (cockpit rear sound only).
W3,4,5,7,10,15	4K x 8 2532	Sound& Speech	Video: Sinistar (upright & cocktail)
W1,2,4,5,7,10,15	2K x 8 2516/2716	Sound only	Video: Warlok
W1,3,6,7,9,11,12,15	512 x 8 7641	Sound only	Shuffle Alley: Big Strike

* Hyperball sound board note: The above information is different than the information which came from Williams. But the above jumpers have been confirmed as correct! Note a 6802 sound board processor can be installed and the old 6808 and 6810 RAM (IC11) removed. To do this, a trace must be cut on the back of the board below R30. It looks like Williams had a bit of a conflict of interest here because the pull-up resistor is in place for the processor swap but the resistor is permanently grounded until the trace is cut.

System6 and System7 Sound Board Jumpers.

System6/7 sound board jumpered WITH pinball speech: W2,5,7,9,10,15. Jumpers shown in blue circles. To convert to no speech on all system6/7 pinballs except Joust, just add jumper W1.

System3/4 Sound Board Jumpers.
The only jumpers on the system3/4 cabinet mounted sound board relate to the type of ROM/EPROM used on the sound board. From the factory most of these sys3/4 sound boards came with a 2316 ROM or 2716 EPROM (2048 bytes), but some also came with a 7641 (512 byte) PROM. In order to use a 2716 in place of a 7641 PROM, either the sound board jumpers need to be changed, or the 2716 slightly modified (bend-up 2716 EPROM pin 21 only, and solder a jumper wire from this pin to EPROM pin 24, which is not bent-up). But here's the jumper table (note the orientation of the jumpers, along the bottom edge of the sound card from left to right are J11,J12,J14,J13,J16,J15,J17,J18):

• 7641 PROM (512 bytes) at IC2: J11,J14,J16,J18 IN.
• 2316 ROM or 2716 EPROM (2048 bytes) at IC2: J12,J13,J15,J17 IN.

CPU Board DIP Switches - Reseting Audits and Factory Settings.
On all system3 games (pinballs with "white Flipper ROMs" installed), the CPU board uses two banks of eight DIP switches. The lower bank and some of the upper bank DIP switches controls the game's adjustments (such as number of coins for play, etc.) and audits. Three upper bank DIP switches allows the operator to zero audit, restore factory settings, and to run the auto-cycle diagnostics. Starting with System4, DIP switch game adjustments were abondoned, and all adjustments were done through software using the coin door's "advance" switch (and the audits/adjustments were stored in CMOS memory). The DIP switches were basically ignored after system3.

If running a system3 game on a later system7 CPU board, DIP switches are still used (as the software is written for DIP switches). This is important to keep in mind since many system7 CPU boards do NOT have the DIP switches installed (though the user could install them on the board).

If running any system4 to system7 game on any CPU board, all 16 DIP switches on the CPU board should be set to OFF (right most position). The game isn't using these anyway, but it's just a good idea to set them all to 'off'.

System3 to System7 games did use the three upper switches of the 16 DIP switches (switches 8,7,6 on the upper DIP switch bank). These were used for zeroing the audit totals, restoring factory settings, and auto-cycling mode (frankly zeroing audits and restoring factory setting can be done easier by just removing the AA batteries in the battery holder for 10 seconds).

Here are the system3 to system7 DIP switch reset/auto-cycle options. All upper bank CPU DIP switches assumed to be OFF (or to the standard game settings on system3), except as noted below. Again, I don't find this particularly useful (or easy to use!), but here it is:

• Upper DIP bank Switch 8 ON = Zero Audit Totals.
• Upper DIP bank Switch 7 ON = Restore Factory Settings.
• Upper DIP bank Switch 6 ON = Auto-cycle Diagnostic Mode.

System3-System7 zero audits, restore factory setting, and auto-cycle diagnostics DIP switch settings.

To activate any of the above, follow these steps:

1. With the game in attract (game over) mode, set the coin door Auto-up/Manual-down switch to Manual-down.
2. Press the Advance button, keeping the coin door open.
3. On the CPU board, set all three of the upper bank "master command" DIP switches (#8,7,6) to OFF (to the right).
4. Set the ONE desired upper DIP bank (master command) DIP switch to ON (move to the left), from the above list.
5. Press the Master Command momentary switch on the CPU board. This is the upper push button switch, above the diagnostic switch. The CPU board's LEDs should turn on as the Master command switch is momentarily pressed in.
6. Turn the game off and back on TWICE to return to game over (attract) mode.

This information applies to all system3 to system7 driver boards (with the exception of Hyperball, which used a unique driver board).

Lamp Matrix Power Resistors.
The power resistors in the lower right corner of System3 to System7 driver boards are often burnt to the point of failure. The original System3 driver boards used 2 watt resistors, which were later (System 4) upgraded to a 3 watt variety. But this is still not enough. These 27 ohm resistors are used for the CPU driven lamp matrix. The game software often did not strobe these feature lamps while in attract mode. This caused the resistors to get very hot, to the point where they can crack or even desolder themselves from the driver board.

The other problem of the 27 ohm resistors relates to CPU lock ups. If the CPU board locks up (Scanbe sockets strike again!), the lamp matrix voltage will no longer strobe. This will definately heat up those resistors enough to desolder them from the driver board.

27 ohm 3-watt R149-R156 resistors that need to be replaced with 5-watt versions on the Driver board (Black Knight). To the left of the large resistors are the eight TIP42 lamp matrix transistors.

To fix this problem, a couple things can be done. First, replace these eight driver board R149-R156 resistors with 27 ohm 5 watt sand or ceramic wire wound resistors. Also make sure to install the new resistors about 1/4" off the circuit board to allow better air flow around them. Another good idea is to use #47 light bulbs for the feature lamps, as this will decrease feature lamp power consumption by about 40%!

Yet another option to fix the burnt resistors is an idea documented by C.Eddy. He replaces the eight TIP42 lamp matrix transistors (Q63, Q65, Q67, Q69, Q71, Q73, Q75, Q77) with IRF9z34N mosfets. The MOS-FETs are installed oriented just like the TIP42 transistors. And the mosfets only need a tiny amount of current to drive them (compared to the TIP42 transistor), hence the large power resistors at R149-R156 never get hot. Because of this there is no need to replace the large resistors (the old burnt ones can be left installed, unless they are open). Heck if the TIP42s are replaced with Irf9z34n mosfets the power resistors R149-156 can even be replaced with jumper wires or zero ohm resistors.

Resistors R204-R211 replaced with wire wrap jumpers. Alternatively the original resistors could stay in the board, and the jumpers go over the resistors. This could save any potential damage to the driver board when trying to remove the original resistors.

System3 to System6 Driver Board Switch Matrix Jumper Upgrade.
Since the driver board is already removed, it's a good idea to upgrade System3 to System6 driver boards to be upward and downward compatible from System7 to System3. To do this, replace the 1000 ohm (system3) or 330 ohm (system 4-6) resistors R204-R211 in the upper right hand corner with zero ohm jumpers. System7 driver boards already have this modification done.

The decrease in these switch matrix resistor ohms was done to increase the current drive through the switch matrix. For example, if a switch or connector was dirty and had slight resistance, the switch could still be sensed by the CPU/Driver board.

There is a rumor that using a jumpered system7 style driver board in a System6 or earlier game may result in random switch closures during game play. This does not seem to be the case (but keep it in mind if having random switch closure problems). One thing for sure though is using a non-jumperd System3 to system6 driver board game in a System7 game will definately result in switch closures being missed.

Part of the problem Williams was having with switches was due to an assembly mistake, which started in the mid-1970s (pre-solidstate). It turns out Williams was assembling one of the pair of leaf blades backwards. This was not a huge deal with Electro-Mechanical (EM) games, but with solidstate games, it was a BIG problem. Because solidstate games use low voltage (5 volt) switches (unlike EM games in which all switches were high voltage 28 volts), the contact rivets are gold plated to help keep them clean (gold is a non-corrosive metal). But because one of the switch blades was reversed, a gold plated switch rivet made contact with a gnarley rough non-gold plated switch rivet. Problems occurred mainly with any switch where a ball "sat", like the ball trough, lock or kickout hole. This mistake was not realized until the Firepower era, and Williams offered retrofit kits for Firepower and Black Knight ball troughs using microswitches to fix the problem.

Replace the Driver Board Female Interconnector Pins.
As described in the burnt connector and random reset section, the 40 pin female interconnector on the driver board should be completely replaced. This is VERY important!! Do NOT skip this step! These pins are rated for "25 cycles" (25 board removal and insertions) before their plating is worn and their tension diminished. And in the high vibration pinball environment, the cycle life is much lower. For reliable game operation, the driver board's female interconnector must be replaced.

Resolder the Header Pins.
While the driver board is out of the game, it would be a good idea to resolder all the male Molex .156" header pin connectors around the edge of the driver board. Because of vibration and board flex (when the driver board is removed from the CPU board), often the header pin solder joints crack. This can give intermittent connections. Also now is a good time to inspect all the male header pins. Any that are burnt or tarnished should be replaced with new .156" Molex header pins.

Driver Board Connectors and their related functions.

The best way to resolder the header pins is to first *remove* the original solder. This can be done quickly with a desoldering tool, as documented at marvin3m.com/begin. It really is a good idea to remove the old solder, and then resolder the header pins with fresh solder. I know most people won't do this (instead they'll just reheat the solder joint, and maybe add a touch of new solder). But to do this right, the old solder should be desoldered, and new solder added (the old solder can have a hard time sticking to the header pins when just reheated).

Give the driver board a good visual inspection. In this picture, do you see the physically blown 7408 driver chip? It can be counted on that the 2N4401 pre-driver and TIP120/TIP102 driver transistors are blown too! Just hope the IC5 solenoid PIA chip survived.

Since the driver board is already out of the game for the other upgrades and modifications, it is a good idea to check all the transistors with a DMM (Digital Multi-Meter). This only takes a few minutes, and will save much trouble down the road if a transistor is found faulty here.

Driver Board Transistors.
By far the most commonly used transistor on the driver board is the TIP120 (along with its 2N4401 pre-driver transistor). It is responsible for driving all the solenoids. Also used is the TIP42, for the Lamp matrix column drives. The 2N6122 (TIP41) transistor is used for the Lamp matrix row drives, and the small 2N6427 transistor is used for overcurrent protection in the lamp matrix. There are some 2N5060 Silicon Controlled Rectifiers (SCRs) used too.

Important Note: Testing transistors (or chips) using the methods below does not give 100% proof that the component is good or bad! It's probably about 95% accurate, but it is not 100% accurate.

All transistors are tested using the diode function of a DMM (Digital Multi Meter).

Testing the TIP120 (or TIP102).
This is the solenoid driver transistor. Always replace a TIP120 with the more robust TIP102.

• Black DMM lead on metal tab (or center leg).
• Red DMM lead on either leg.
• .4 to .6 volts seen.

Testing the 2N4401.
This is the pre-driver for TIP102 (or TIP102). Note if testing the 2n4401 "in circuit" (installed in the driver board), this transistor can test as "bad" (yet still be good), if it's controlling TTL 74xx chip is bad (which is actually fairly common).

• Red DMM lead on center leg.
• Black DMM lead on either leg.
• .4 to .6 volts seen.

Testing the 2N6122 or TIP41 (TIP41 is a sub for the 2N6122).
This is the lamp matrix row transistors.

• Orient the transistor's writing facing towards you.
• Read DMM lead on the *left* leg (base) of transistor.
• Red DMM lead on the center (collector) leg (or metal tab), .4 to .6 volts seen.
• Red DMM lead on the right (emitter) leg, .4 to .6 volts seen.

Testing the TIP42.
This is the lamp matrix column transistors.

• Orient the transistor's writing facing towards you.
• Black DMM lead on *left* leg (base) of transistor.
• Red DMM on center leg (or metal tab), .4 to .6 volts seen.
• Red DMM on left leg, .4 to .6 volts seen.

Testing the 2N6427 (or MPSA14 or NTE46).
This is the pre-driver for 2N6122/TIP41 and TIP42 transistors.

• Red DMM lead on middle leg.
• Black DMM lead on right leg, 1.0 to 1.3 volts seen.
• Black DMM lead on left leg, .6 to .8 volts seen.

Testing the 2N5060.
This is the pre-drivers for the lamp matrix, S1-S8.

• Red DMM lead on middle leg.
• Black DMM lead on right leg, .5 to .7 volts seen.
• Black DMM lead on left leg, .9 to 1.1 volts seen.

The Coil Diodes and Why they are Important.
After testing all the driver board transistors, it is important to examine all the coils in the game. If any of these coils has a bad diode, this can almost instantly kill its associated driver transistor! The coil diode prevents a coil's collapsing voltage from "backwashing" to the driver board, damaging the driver transistor.

Since you spent the time to test/replace the bad driver board transistors, it only makes sense to also check for bad coil diodes. Since these 1N4004 diodes are mounted right to the coils under the playfield, vibration can crack or damage them.

The best way to test a coil diode is to just grab the diode by its body with the forefinger and thumb, and gently give it a pull. If the diode has a cracked body or broken lead, it should be pretty easy to see.

The 1N4004 coil diode mounted on a Firepower slingshot, showing the proper orientation of the diode and power wires.

Testing 1N4001/1N4004 Diodes on Coils, Lamp Sockets, Switches, Etc.
This test applies to all 1N4001 to 1N4007 series of diodes. When mounted on coils or lamp sockets or switches, one end of the diode should be removed, so the diode is tested "out of circuit". On circuit boards, this is usually not required. DMM set to the diode function:
• Turn the game off.
• Unsolder or cut one end of the diode from the coil, lamp socket or switch.
• Use a DMM set to the diode function.
• Put the blank DMM lead on the banded side of the diode.
• Put the red DMM lead on the NON-banded side of the diode.
• .4 to .6 volts should be seen.
• Reverse the DMM leads, and a null reading should be seen.
• If these values are not seen, replace the diode with a new 1N4004 diode.

In all situations, remember to mounted the new diode correctly. For example, on coils, install the diode with the diode's band on the power lug of the coil. It usually pretty easy to tell which is the power lug of a coil. The power wire, which daisy chains from coil to coil, is usually the thicker wire on a coil lug. The banded lead of the 1N4004 diode should be connected to the coil lug with this thicker daisy chained power wire attached. The non-banded end of the diode attaches to the coil lug with the thinner wire, which leads to the driver board transistor, and ultimately ground.

Burnt General Illumination (GI) Connectors on System3 and System7.
As a continuation of the power supply section, one problem that plagues Williams machines Hot Tip/Lucky Seven (System3) and System7 and later is burnt GI (General Illumination) connectors. The GI connectors that burn are the ones going from the transformer to the power supply, and from the power supply to the backbox/playfield, as seen on System7 games, but *not* System3 (except Hot Tip/Lucky Seven) to System6 games. Hot Tip and Lucky Seven System3 power supplies used .156" header pins on the power supply board for the GI connectors. System7 machines routed the GI power through a round two pin connector on the power supply board, and back out through a .156" header pin connector. Because of this, System7 and early System3 GI connectors tended to burn very early in their life. This is unlike (later) System3 (World Cup) to System6 machines which had the GI hard wired right to a fuse card, which really prevented the burnt GI connector problem fairly well.

To fix burnt GI connectors, they must be replaced completely. That means both the circuit board mounted connectors AND the wire mounted connectors. The output GI connector, a straight .156" header, is no problem (part number listed above in the non inter-board connector section). Just make sure to use Trifurcon pins for the wire mounted connector (they have greater surface area, and last longer than standard pins), and a new square 9 pin .156" header for the circuit board.

Note on system7 games there is often a connector the breaks the GI playfield wiring harness for removal of the backbox. This connector will usually have 6.3 volt AC yellow and purple wires, and white wires with yellow and purple traces. Also sometimes this connector will have 12 volts DC too - this powers any under-playfield GI relays (if the game uses them).

The Single Pin Cabinet/Backbox G.I. Connector.
On system3 to system6 games, there is a single pin connector in the wiring harness that connects the backbox GI fuse to the playfield (this connectors allows the backbox to be removed from the game). Often this single pin connector can also burn, since all the playfield GI lamp power goes through this. ADD PICTURE

The GI relay on BlackOut, Scorpion, & early Black Knights with System6 power supplies is located next to the lamp matrix capacitor in the bottom of the backbox. The yellow wires are the GI power wires, which attach to the contacts on the relay, and open/close, turning the GI off and on, as required by the game. The relay's red/black wires go to the power supply board connector 3J7, and provides power/ground to energize the relay, via a driver transistor on the driver board. Picture by Mark.

Other Hot Tip/Lucky Seven and System7 General Illumination things to check:
• Using a DMM set to AC volts, check for 6.3 volts AC right at the transformer. The yellow and yellow/white wires are the General Illumination wires. Put a DMM lead on the yellow transformer wire, and the other DMM lead on the yellow/white wire.
• Measure the GI's 6.3 volts AC at the power supply input plug (two pin plug at the lower right).
• Measure the GI's 6.3 volts AC at the power supply fuse (the fuse is good, right?).
• Measure the GI's 6.3 volts AC at the power supply output plug (the nine pin .156" molex plug at the lower right, next to the 2 pin plug).
• If there is GI power at the input power supply GI plug, but nothing at the output GI power supply plug, the GI relay (mounted on the power supply with System7, or on the backbox's bottom floor on BlackOut & Scorpion and early Black Knight games next to the lamp matrix capacitor) could be the problem. To check this, disconnect the relay connector; if the relay is shorted/seized (or the driver board transistor/circuit that turns the relay on has failed), the GI lights should come on. On System7 games, this is the power supply connector 3J7. On BlackOut, Scorpion and Black Knights with System6 power supplies and the GI relay on the backbox floor, just remove the black/red connector from the wiring harness.
• Many system7 games also have 12 volt DC G.I. relays under the playfield (or even on the insert panel of the backglass), so the CPU board can toggle a particular G.I. playfield string on and off. These relays have five lugs (center lug unused). One side of the GI string is interrupted by the relay, so two solid yellow or purple GI wires go to the relay. The other two lugs are 12 volts power and the 12 volt ground path back to the driver board TIP122 transistor. When the relay is not energized, the G.I. relay's switch is normally closed, allowing G.I. power to pass through the relay. Often the problem with no G.I. is the lug connectors on these G.I. relay(s). Recrimp new lug connectors or squeeze the old ones with pliers for a better connection.

System7 (Black Knight) General Illumination connectors. The two pin plug on the right is the GI coming into the power supply from the transformer. The .156" header pin connector on the left is the GI coming out of the power supply board. The fuse is the GI fuse. Note the owner repair made with the added wires, because these connectors burned. Note the GI relay can be barely seen above the fuse.

On System7 power supplies, the two pin input GI connector is more troublesome, because it is an older Molex part (as seen on the right in the picture above). The power supply mounted male two pin Molex part number is 15-31-1026 (specs here), and has solid pins to withstand heat better. The female plastic wire mounted housing is Molex part number 19-09-1026 (specs here, see page two). The plastic housing uses standard female Molex .093 diameter series terminal pins, series 1189, part number 02-09-1104 (tin contacts) (specs here). Mouser.com claims to handle these parts.

To summarize, get these parts to repair the GI connectors on Hot Tip/Lucky Seven and System7 power supply board:

• .156" header pins with lock (9 pins), part# 26-48-1095 (Mouser).
• .156" Trifurcon pins (three wipers): Molex part# 08-52-0113 (tin plated phosphor bronze) or 08-50-0189 (tin plated brass), for 18 to 20 guage wire. Digikey part# WM2313-ND. Mouser and Competitive Products (#06-2186) also sells these.
• .156" white housings (9 pins), part# 09-50-3091 (Mouser)
• PCB mounted two male pins, Molex part number 15-31-1026 (not needed for Hot Tip/Lucky Seven).
• Female plastic wire mounted housing, Molex part number 19-09-1026 (not needed for Hot Tip/Lucky Seven).
• Female Molex .093 diameter terminal pins, part number 02-09-1104 (not needed for Hot Tip/Lucky Seven).

G.I. Relay Replacement.
If the G.I. relay has failed or burned, a replacement PC poard mount 24 volt DC 10 amp DPDT relay can be used. The original is a Guardian Electric 1395P series, and a Mouser equivelant is the NTE R14-11D10-24. Just make sure to get a DC coil relay and (not the AC). Playfield or backglass insert G.I. relays are 12 volt DC 10 DPST relays.

Dead batteries (or worse, dead corroded batteries) will cause problems getting a system4 to system7 game to power-up properly. On these games, if the batteries are dead, the game can still boot, but will come up in "audit" mode (system3 games do not have this boot-up audit mode). To get to attract mode, turn the game off and on again quickly. If done fast enough, the game should come up in attract mode. Note on System6 and System7 games, the coin door *must* be open for this to work (memory protection switch open).

If the game still doesn't come up in attract mode, even with the "turn it off and on quickly" trick, the CMOS 5101 RAM chip at IC19 is probably bad.

Please refer back to the Batteries and Battery Holders section for more information on this subject, including how to test the batteries, blocking diode D17, and the CMOS 5101 RAM at IC19.

Booting into Audit Mode Explained.
On system4 to system7 pinballs all the game's options and audits are stored in CMOS memory. If the batteries are dead, or the battery holder is damaged, or the blocking diode D17 has failed, or there's a bad IC19 RAM 5101 chip, or battery corrosion has damaged the CPU board, the game will power up into "audit mode". Audit mode is shown in the picture above, and is saying that the game has lost its CMOS memory, and there's a problem. It's a big red flag when the game is turned on, since the game goes into audit mode instead of attract mode (game over mode). Operator assistance required!

AUDIT MODE: On system4 to system7 games, the dreaded audit mode. A Firepower powered-on with dead batteries and/or a dead IC19 5101 RAM chip, booting into audit mode. Audit "00" shows the game number (#497) in the player one score display, the operating system revision (the preceding "1", meaning "green" flipper ROMs, where "blue" flipper ROMs have a "2", and yellow flipper ROMs have a "0"), and the software revision (version "2") next to the game number. The "00" in the ball-in-play displays shows audits number zero, and "04" in the credit window indicates audits (remember "01" is lamp test, "02" is solenoid test, and "03" is switch test).

In audit mode ("04 00" in the credit/ball-in-play display, where "04" is audits, and "00" is the first audit number), the numbers shown in the player1 score display are the value for the audit number shown in the ball-in-play display. For audit "00", which is the software identification audit, the last number is the game's current software revision number (version 2 in the above picture). The middle three numbers are the game number (i.e. 497 is Firepower; see the Game List section for all the game titles and game numbers). And the first number determines the "flipper ROMs" version installed. Remember Williams used a color coded system:
• 0 = Yellow Flipper ROMs (system4)
• 1 = Green Flipper ROMs (system6)
• 2 = Blue Flipper ROMs (system7)

These software identification numbers made it easy to see if the wrong Game and/or Flipper ROM software was installed in the machine.

Note the lack of a code for White flipper ROMs (system3). This is because the boot-up "software versions" was not implemented until System4 and the Yellow flipper ROMs, when adjustment were also stored in memory (system3 used DIP switches for the adjustments). Williams did the audit mode routine to show instantly upon power-on that the game's adjustments/audits were lost, and that the batteries needed to be replaced. The main reason this was done was to protect the game from having garbage in an adjustment that may put the game into free play (or some other equally accidental bad mode), since now all the game's adjustments were store in memory instead of being "hardcoded" with DIP switches. With system6 and its memory protection circuit/coin door switch, it also keep miscreants from drilling through the bottom of the game and activating the switches to change the settings (like one quarter equals 25 credits!), since the coin door now had to be open to change an adjustment/audits.

A system3 game (Hot Tip) with dead batteries booting into audit mode. Here the audit number is in the credit window ("01"), the audit mode ("04") is in the ball-in-play window, and the value for the first audit ("090000") is in the player1 score window.

On system3 games, a dead battery or failed CMOS memory still comes up in audit mode, but there is no indication of software revisions. The audits in system3's white flipper ROMs looks a bit different too, with the audit number in the credit window, and the "04" (to signify audits) in the ball-in-play window (this was reverse of system4 to system7), and the audit value in the player1 score display. If the manual-down/auto-up switch is in the auto-up position, the game rotates through all the audit numbers automatically also.

For info on the Blanking signal, chick here to jump down to that section.
For info on the Leon's Test chip, chick here to jump down to that section.

Introduction to CPU/Driver Board Repair.
The System3 to System7 CPU/Driver board combo are a real circuit board design nightmare. After spending a great deal of time fixing 1977-1985 Bally/Stern pinball and Gottlieb System80 machines, I would personally have to say the Williams System3 to System7 games are way worse than 1977-1985 Bally games (Williams sys3-7 is tied with Gottlieb System80, or maybe is even worse, thanks to all the bad Scanbe sockets and 40 pin interconnector!)

System3 to System7 games were designed so that a field tech would only have to swap out a failed board, and then send the board to the distributor for repair. The boards were not designed to be diagnosed and fixed on location. Looking at the Williams "repair" manuals, the instructions tell pretty much only how to pinpoint which board is bad. The final "fix" of most repair steps in a Williams manual is to "replace the board". Unlike more recent games which report many problems, early Williams board provided little diagnostic help. This was unlike the approach Bally took with their famous "seven flashes" diagnostic system (which makes component level repair pretty easy).

Where Bally took the approach that the boot-up diagnostics were very important and game diagnostics were not, Williams took the opposite approach that boot-up diagnostics were *not* important but game diagnostics were! Hence 1977-1985 Bally games have the famous power-on seven flash test, but don't really have a lamp, display, coil or switch test. Williams system3 to system7 games have basically no significant power-on test indicator, but do have good versions of lamp, display, coil and switch tests (assuming that the Williams game in question will boot up properly!)

Initially, when these games were new, Williams pinballs were dependable, about as good as Bally games (and better than Gottlieb' system80). But after a few years, as the Scanbe sockets got old and the 40 pin interconnector wore, dependability of these games made them way worse than Bally (and probably on par with Gottlieb). And with no power-on indicator system like Bally had, operators didn't even know where to start with a Williams repair.

It's not that the Williams system is "hard" to fix, it's just time consuming. Where the Achilles' heel of the Bally system is battery corrosion and Gottlieb had board connector and ground issues, Williams System3 to System7's Achilles' heel is the design of the boards themselves. Passing address and data lines over a 40 pin .156" Molex interconnector is a bad idea. If *any* address or data line drops out for even a millisecond, the whole game locks up. Same thing with the Scanbe sockets; if a single pin become intermittent on any Scanbe socket, the game locks up. Add vibration to the mix, and things only get worse. Having some PIAs on the CPU board and some PIAs on the driver board forced the address and data lines to cross the two boards at this Achilles' heel (the 40 pin interboard connector). And if the game locks up when a coil is energized, that can take out more components on the driver board, making the repair even worse.

Also some components on the Williams System3 to System7 boards are obsolete and unavailable. For example, the System3/4 clock chip (MC6875) is obsolete and impossible to find. This is the companion clock chip to the 6800 CPU processor. Luckily, the MC6875 doesn't fail too often. Also the 8T28 and 8T97 chips are discontinued (there is a work arounds for most of those). But this is why many repair facilities will not repair Williams System3 or System4 CPU boards.

To make matters even worse, the software contained in the "flipper ROMs" (essentially the operating system) adds to the problems. If any of the PIAs can not be found in the exact state the software expects, this locks up the software, and hence the entire game. So say the solenoid PIA is partially dead (common, because of a locked-on driver transistor which cascaded back to the PIA and damaged it). This may make a coil or two not function. That's fine; the game maybe could still be operated and played without a bumper or two. But instead, the whole game locks up and becomes useless. And because the CPU is locked up, diagnosing the problem becomes very difficult too.

To add insult to injury, the driver board is an integral part of the CPU board. That is, the CPU board will *not* run without a functioning driver board attached! (Unless a special test EPROM is used, which has only recently became available.) So the whole theory of spliting the system into smaller components (a separate CPU and separate driver board) for easier diagnostics is missed*. The two boards need each other, and are linked together through the incredibly stupid and troublesome .156" 40 pin interconnector. This is unlike Bally and Gottlieb, where their CPU boards can be run independently of the driver board, making diagnosing problems much easier, since the system can be broken down into smaller, more managable pieces.

* Note there is a trick that allows the Williams CPU board to (semi) boot without the driver board. The driver board can be completely removed from the game, and the CPU board booted. If the CPU board is OK (trying to run), the two CPU LEDs will blink on, then go off, and then come on steady (since the CPU board is looking for the Driver Board). If installing the Driver Board locks the CPU (LEDs steady on, no blinking at bootup), then the CPU board is probably Ok (and there's problems on the Driver Board.) On a system7 CPU board, if the CPU were OK (or trying to run) the Numeric Led would blink “0”, go off and come on steady (looking for the Driver Board). If putting the Driver Board back in locks up the game ("0" steady on), then there are some problems on the Driver Board

The Internal System3-System7 Diagnostic Firmware.
AKA, this is NOT "Bally World".
Lastly (really I promise!), the System3 to System7 internal diagnostic software is very limited. Remember the Bally power-on LED flash test, and how nifty that is at helping identify bad CPU board components? Well unfortunately, there is nothing like that in the Williams' firmware. Either a Williams system3 to system7 CPU/driver board boots, or it does not. There's not much middle ground here. The Williams' LED(s) at boot up provide some very basic information, but nothing like the information a Bally MPU LED provides. Yes there is a diagnostic test switch SW1 on the Williams CPU board. But in order for this test to work properly, the CPU board has to be successfully booted and running! But if the CPU board has successfully booted, the need for this test is really quite limited.

To make matters even worse, the diagnostic SW1 test, even on a working CPU board, can confuse even a veteran user. The diagnostics are a memory test only, and tests the CMOS RAM IC19 (5101) and the two static RAM chips (IC13/IC16). But the static RAMs IC13/IC16 rarely die. And the user will already know if the CMOS 5101 RAM is dead well before the diagnostics are run. If the game boots into "audits" mode, and the CPU batteries are good, it's 99% for sure the 5101 RAM is dead.

Also, even if the CPU board has seemingly "booted correctly", the flipper ROMs IC17 and IC20 can still have problems. These two ROMs hold the diagnostic code, and if one of these ROMs has a problem, false indications can result from the SW1 diagnostic switch (but usually the CPU board didn't boot anyway and the LEDs are indicating a locked-up board, and these two ROMs don't even get a chance to start working).

To make matters worse, the diagnostic LEDs just tend to confuse the newbie repair person. For example, the CPU board does not boot, but the user presses the diagnostic switch SW1 anyway. The LED reports back the suspected failed component. But that's the problem... since the CPU board never booted properly, the output from the diagnostic test CAN NOT be trusted! What ever the test indicates is surely incorrect, and the newbie is replacing good CPU board components, based on the failed/incorrect test results (I believe this is called, "chasing one's tail"). This is especially a problem if the newbie came from "Bally world", where Bally's LED actually has good boot-up component diagnostic results.

The bottom line is this: if the Williams System3-System6 CPU board's LEDs lock-on at power up, the CPU board is not working! Likewise for System7, if "0" comes on immediately at power-up, the CPU board is not working. Why that is happening, well, you're on your own to figure it out! Because the Williams diagnostic firmware is *not* going to help.

OK, So That's the Bad News. What's the Good News?
The good news is fixing a dead Williams CPU/Driver board is pretty systematic. The same things seem to cause the same problems, and it doesn't stray too far off this path too often. The bottom line is this: all the CPU board and driver board Scanbe sockets need to be replaced. Change the flipper ROMs and gamerom to EPROMs, to use less chips/sockets and to replace the old and undependable original ROM chips. The 40 pin female interconnector needs to be replaced. The CPU board's 5101 RAM at IC19 dies often. The driver board's solenoid and switch matrix PIAs die. Sure lots of other things can and do happen, but that's the majority of the CPU/driver board problems!

Again, use the simple test to trick the Williams CPU board to (semi) boot without the driver board. The driver board can be completely removed from the game, and the CPU board booted. If the CPU board alone comes on with both LEDs on (no blinking), then the CPU board is faulty. If the CPU board is OK (trying to run), the two CPU LEDs will blink on, then go off, and then come on steady (since the CPU board is looking for the Driver Board). If installing the Driver Board locks the CPU (LEDs steady on, no blinking at bootup), then the CPU board is probably Ok (and there's problems on the Driver Board.) On a system7 CPU board, if the CPU were OK (or trying to run) the Numeric Led would blink “0”, go off and come on steady (looking for the Driver Board). If putting the Driver Board back in locks up the game ("0" steady on), then there are some problems on the Driver Board

Before Starting...
Before starting to fix a dead CPU/driver board, go back to the "Before Turning the Game On" sections and follow all those steps. It's not even worth going on until that work is done. For example, the power supply needs to be working and outputting the correct voltages. Also the assumption is that the CPU board's LED(s) are locked on, or at least flash. If they don't do that, make sure there is +5 volts getting to the CPU board, and then check the section below for more help.

Before removing any boards from the game, and assuming the power supply is working, it's time to do some preliminary diagnostics. Remove the game's backglass, so the CPU board's LED(s) can be seen (if the CPU board in question is not in the game, it can be powered up "on the bench" with an external +5 volt power supply, those details are below). Right now, before doing anything, remove fuse F2 (solenoid power) and fuse F3 (lamp matrix power) from the power supply board. Now go ahead and power the game on, keeping an eye on the LED(s) on the CPU board. Here's a list of what could happen:

• CPU LEDs come on and stay on (or system7 LED displays zero), with no flash:
  The CPU either isn't booting, isn't resetting, or has locked up attempting to boot. Common problems are bad CPU sockets (Scanbe) or a bad inter-board interconnect board. Go to the Leon Test ROM section below.

• CPU LEDs flash once (or system7 CPU LED flashes "zero" once), and then turns on and stays on:
  The CPU board booted and then ran into a problem. Most like a ROM related problem, like bad sockets (Scanbe) or bad ROMs, or a bad PIA on the CPU or driver board. Go to the Leon Test ROM section below.

• CPU LED(s) flash once quickly then go out, but the score displays do not turn on:
  The CPU has started and the game program is (probably) running, but there is either a ROM problem or a problem with the score displays (missing high voltage?) or a blanking problem. Turn the game off, and replace fuse F3 (the lamp matrix power). Turn the game on, then turn the game off and on quickly; if the score displays remain off, but the game goes into "attract" mode (feature lights flashing), there is a problem with either the master display driver board or the display circuitry on the CPU board. Remember system3 games (white Flipper ROMs) do not have an "attract" mode other than the backglass high score lamp flashing on and off. So for these games, try starting a game to determine if there is a score display problem (add some credits and press the start button). If the feature lights do not come on (or a system3 game can not be started), Go to the Leon Test ROM section below.

• CPU LEDs flash once then go out and player 1 display shows what looks like an error code:
  The game is booting up into "audits" mode. This is good news, as the CPU/driver board problems are usually minimal. The CMOS RAM memory has been erased for some reason (dead batteries or bad CMOS RAM), and this is a warning that all audit and adjustment information has been forgotten. With the coin door open, turn the game off and on quickly, seeing if the game goes into attract mode (might have to do this several times). If the game comes up in attract mode, then chances are good the CMOS IC19 RAM chip is Ok, and all that is needed is new batteries or battery holder (please see the section on batteries). If the game continues to come up in audits mode, then the IC19 CMOS RAM 5101 chip is probably bad. Replace this chip (use a socket!), and the game should be good to go.

• To determine if the problem is the CPU board or driver board, again, use the simple test to trick the Williams CPU board to (semi) boot without the driver board.
  On a system3-6 game, the driver board can be completely removed from the game, and the CPU board booted. If both CPU board LEDs should come on, go off, and then come back on and stay on. If they do, that usually means the CPU board is Ok, and the driver board needs some work. Go to the Leon Test ROM section below.

Get Leon's CPU Test EPROM.
All the other "Before Turning the Game On" work is done, right? But the CPU/driver board is still dead. Well now is the time to get Leon's test EPROM. This chip is really needed to do any serious diagnosing of the CPU board, as it allows the CPU board to be run without the driver board. Chances are good you do not have an EPROM programmer. So you'll need to find someone to burn this chip into a 2716 EPROM (or 2532 for System7 CPU boards). The 2716 EPROM code for System3 to System6 games can be downloaded by clicking here (2716) (new system3-6 version 11/01/03). The 2532 EPROM code for System7 can be downloaded by clicking here (2532) {new system7 version 11/01/03}.

EPROM Tip: If getting an EPROM burned is really a big deal, the System3 to System6 version of Leon's test EPROM can be "doubled up", burned into a 2532, and used in *any* System3 to System7 CPU board. Yes the 2532 is twice the size of the expected 2716 at IC17 on a System3 to System6 CPU board. But because of the way a 2532 EPROM is addressed, the added address space of the 2532 will just be ignored by the System3 to System6 CPU board. But there is a downside to this; Leon's memory test will *not* work in a system7 CPU board. This is because memory for System3 to System6 is at locations $0080 to $0100, and in System7 memory is located from $0000 to $0100 (but the basic part of Leon's test ROM will work, but the memory test will not). This works in a pinch if the correct EPROM is not available.

After the above EPROM code is downloaded and burned into the appropriate EPROM, remove all the existing ROM/EPROM chips from locations IC17, IC20, IC14 (and IC21, IC22, IC26, but there should not be any chips in those locations anyway if the CPU board was converted to EPROMs and new sockets installed). Now installed Leon's test EPROM into the flipper ROM socket at IC17.

Note: Leon's chip will work with the other ROM chips installed. This happens because Leon's test program starts at the "boot up" memory location, and it does not access the other ROM chips. So basically the other ROM chips get ignored, as Leon's chip is accessed first at power-on, after the CPU board resets. BUT I highly recommend removing ALL the other ROM chips! If one of the other ROMs is bad (shorted), it could "lock up" the CPU board. By removing them, it's just one less thing to go wrong.

Separate the CPU board from the Driver Board.
The test EPROM is burned and installed at IC17. Now it's time to separate the CPU and driver boards. Remember how I mentioned the CPU board would *not* work without the driver board? Well that's correct, except when using Leon's test EPROM! This is the *only* firmware that will allow the CPU board to run without the driver board attached (thank you Leon!) This alone makes diagnosing a bad CPU/driver board much easier.

IMPORTANT: Leon's test EPROM can be run with the CPU board installed in the game, but this is *not* recommended. The test EPROM will toggle all the PIA outputs. If installed in a game, this will turn all the lamps and solenoids on and off! If the test EPROM really must be run in the game, make sure to remove fuses F1 (H.V.), F2 (solenoids), and F3 (lamp matrix) from the power supply board, and/or all the connectors are removed from the driver board and the CPU board (except CPU connector 1J2). This will ensure the solenoids and lamps and score displays are not energizing. For example, if the lamp matrix F3 fuse is not removed, a full 18 volts (not strobed!) will go to the CPU controlled lamps, and burn all of them out instantly! It is also a good idea to remove the power going to the sound board as the PIA outputs could toggle the sounds.

Removing power supply fuses F1 (H.V.), F2 (coils), and F3 (lamps) before powering-on with the Leon test ROM and the boards mounted in the game.

Booting the CPU board on the "Bench" with an External Power Supply.

The next trick to fixing a CPU/driver board is to move the boards from the game and to the work bench. In order to do this, an old AT style computer power supply that outputs +5 volts DC (and +12 volts) is needed. These are pretty easy to come by, any computer store should have one for free to $20 dollars (heck new AT computer power supplies can be purchased too for less than $30). An old video game switching power supply can also be used.

Newer ATX power supplies can also be used, but these do not have a physical power switch. Instead they get a signal from the computer's motherboard connector to turn the power supply off. But these power supplies can be fooled to turn on when their power cord is plugged in. Just tie the green /PS-ON wire (power supply on, active low, normally pin 14 on the motherboard connector) to the black COM ground wire. (a diagram of the 20-pin ATX connector can be found at wired.hard.ru/data/atxpower.shtml).

Hooking up a system3 to system7 CPU board to an external PC power supply. The red wire is +5 volts, the yellow wire is +12 volts, and the black wire is ground.

After the power supply is obtained, power it up and figure out which wires are the ground COM (usually black), +5 volts DC (usually red), and +12 volts DC (optional, but usually yellow). Use a DMM to test the voltages. Turn the power supply off. Now get three alligator clip test leads, and hook up the power supply to the CPU board's connector 1J2 as follows:

• Ground (black) = CPU board 1J2 pins 1,2,3 (connect to pin 1)
• +5 volts (red) = CPU board 1J2 pins 4,5,6 (connect to pin 5)
• Unregulated +5 (or +12) volts = CPU board 1J2 pin 9*

* Is the Unregulated +5 Volts Needed?
Williams calls this "unregulated +5 volts", but in reality it's actually closer to unregulated 12 volts. System6 and System7 CPU boards may be easily jumpered to work on only +5 volts, with no need to use a power supply's +12 volts too. Interestingly, System3 and System4 CPU boards also do not require the +12 volts IF they have the "reset modification" performed (in the CPU board modification section). I personally don't recommend using the +12 volts unless it is needed, as it's just another thing to mis-connect, and most CPU boards don't need it.

To get around not connecting power to the unregulated +5 volt power pin on System6 to System7 CPU boards, try the board first without the unregulated +5 volts. If it doesn work, use an alligator jumper clip and do this: Connect the power supply's +5 volts (the electroylic cap C23's positive lead) to resistor R27 (see which side of R27 in the bullet points below). Because the layout of System6 and System7 is different, pay particular attention to the instructions below. Because one side of R27 is ground, and connecting +5 directly to ground would cause the power supply to short. So be careful which side of R27 is used! In all cases, cap C23 and resistor R27 are just to the right of connector 1J2 on the CPU board.

• System6 CPU board: alligator clip on the TOP side (positive lead) of C23. Connect the other jumper to the bottom lead of R27.
• System7 CPU board: alligator clip on the BOTTOM side (positive lead) of C23. Connect the other jumper to the BOTTOM lead of R27.

At Power-Up, the LEDs or 7-Segment Display Does not Come On.
On system3 to system6, if the LEDs don't come on at all, or on system7 the segment display doesn't come on, there is a problem! This is usually a sign of a short on the board (battery corrosion?) or no power. Make sure the power is hooked up properly to the board. On System 6/7 CPU boards, check test point TP9 for +5 volts. On all system3-7 CPU boards, also test for +5 volts at pin 8 of the CPU chip IC1 (ground is IC1 pin 1) and at interconnector pin 1 on the far right.

If there's no +5 volts at the CPU chip, then trace back through the power circuit to see where it's losing power. Check to make sure there isn't a short to ground. There are capacitors (.01 mfd non-polarized) at the Vcc (+5 volt) connection of each chip on the board to ground. Check each one of these for a short. Also check the board very carefully for solder splashes. Remember a prior repair may have gone bad and the board was junked for this reason.

On System3 to system6 boards, check for +5 volts at IC2 pins 2,5,16 (the 8T28 that drives the LEDs). Also remember both LEDs could be bad! The LEDs will always come on at power-up unless turned off by an executing program on the CPU board. If there is proper voltages on the board, but no voltage at pins 2 and 5 of IC2, check for a running CPU chip (see below). If there is CPU activity on the address and data lines, then the problem is most likely IC2 (8T28) on system3 to system6 boards.

On System7 CPU boards, if there is no activity on the 7-segment display, it is common for IC34 (7447) segment driver chip to fail. This will give a false indication that the board is completely dead or that there's no power. Using a logic probe, test for pulsing at IC1 pin 15 of the CPU chip. If there is activity there, then the problem is most likely IC34 or the segment LED itself. Also if the LED comes on and stays on (even though the game is seemingly working fine), often IC2 (74125) has shorted internally.

Note two test LEDs (like on system3-6 CPU boards) can be added to the System 7 CPU board. This will allow testing to the board before replacing IC34 and/or segment LED. All that is needed is two LEDs and two 100 ohm 1/4 watt resistors. Solder the two resistors into the board, just to the left of test point TP9. The LEDs then should be soldered into the two top pads pairs next to the segment LED, with the flat side of the LEDs towards the 7-segment LED. With the two LEDs installed, they will "flash" just as the LEDs flash on system3 to system6 CPU boards.

Adding LEDs to a System7 CPU board. Note the flat side of the LEDs goes towards the 7-segment display, and the new LEDs mount on the upper two solder pads (of the four pairs). Two 100 ohm resistors are also added next to TP9. If the 7-segment LED or IC34 that drives the 7-segment display is bad, the installed LEDs can be used to diagnose other CPU board problems.

Step 1: How the CPU board boots (or "Getting Started").
Ok, so the CPU board is all ready, with the test EPROM installed, power supply connected, and the driver board removed. Now we can power on the CPU board. But first, it would be helpful to know what is happening, and how to check it.

Here are the steps involved in a correctly booting CPU board. In the case of a dead CPU board, each step can be diagnosed with a DMM and/or logic probe.

• Make sure there is +5 volts at CPU IC1 pin 8 (and IC1 pin 35 on system6/7). There should also be a slightly lower voltage at IC1 pin 2, and ground at IC1 pins 1,21.
• At power-on, the /RESET (U1 pin 40) 6800/6802 CPU chip is held low for about 50 milliseconds by the reset circuit. This gives the power supply a chance to stablize the +5 volts before the CPU chip starts to "boot".
• Next the /RESET pin 40 goes high, and the CPU chip begins to execute a startup program inside IC17 (the Flipper ROM). Using a DMM connected to ground and U1 pin 40, check that the voltage comes up to at least 4 volts (from zero volts) when the CPU is first turned on. If the reset pin is not going high (to at least 4 volts), suspect CPU transistors Q1-Q4, Q6-Q9 and diodes D19,ZR1,ZR2 on System6-7, and CPU transistors Q1-4, Q6 and diodes D17-D18,ZR1 on System3-4. Using the schematics, work back through the circuit, testing each transistor, diode and resistor until the faulty component is found. Interestingly, the reset circuit can be overridden by connecting +5 volts to IC1 pin 40 of the CPU. If the reset circuit is the only problem, the CPU should start when +5 volts is applied to pin 40. This is a good method if testing for a bad reset circuit.
• The "clock" circuit must also be running. This gives the CPU chip a timing pattern.

  On System3 and System4 CPU boards, the clock can be seen with a logic probe at chip U1 pins 3,36,37. Also the the same signal can also be seen at the MC6875 (left of the crystal) IC5 pins 7,13,15 (IC5 pin15 feeds CPU IC1 pin 3, and IC5 pin13 feeds CPU IC1 pin36,37). On system3 to system6, hopefully all the signals are at IC5, as this chip is obsolete.

  If there is no clock signal on the IC5 MC6875 chip, first replace the crystal as they are much cheaper than trying to find the obsolete MC6875 chip. The manual calls for a 3.58 Mhz crystal, but 3.579545 Mhz is really what is available today, and this works fine.

The clock signal on System3 and System4 CPU boards at IC1 pin 3. This same signal can also be seen at IC1 pins 36,37.

On system6 and system7 CPU boards, the clock is slightly different because Williams replaced the 6800 with a 6802/6808 CPU processor. With this upgrade, the obsolete MC6875 chip was no longer needed, and hence was eliminated. The clock circuit on the 6802/6808 is built into the CPU chip itself, so the only external clock parts are the crystal and two capacitors. On system6/7 the clock can be seen on the CPU chip U1 at pins 38,39 (which feeds to the 3.58 mHz or 3.579545 Mhz crystal CR1, and the two capacitors C25,C26 which are 27 pF). If the clock is not pulsing on system6/7, replace the crystal and check caps C25,C26. If still no clock, then the IC1 CPU 6802/6808 chip is bad.

The clock signal on System6 and System7 CPU boards at IC1 pins 38,39.

• The CPU chip now begins to run a program in the IC17 ROM chip (this is one of the Flipper ROMs). In order to do this, the address and data lines fetch the information from the ROM chip, and the CPU chip executes it. The RAM chips are also used in this processes too. If everything checks out prior to this step, but the CPU board's LEDs lock on or lock off, suspect the EPROM IC17 or its socket (also suspect all the other EPROMs and their sockets, but with Leon's test EPROM, there shouldn't be any other ROMs installed!)

  All the IC17 EPROM address lines A0-A7 (pins 8-1), A8-A10 (pin 23,22,19), Chip Select2 (pin 20), and data lines D0-D7 (pins 9-11,13-17) on the EPROM IC17 chip can be checked with a logic probe. They should be pulsing for the most part, except for maybe address lines A8-A10 (pins 23,22,19). If any are missing (not pulsing), suspect a broken or shorted circuit board trace (use the DMM's continuity function, as all the address/data lines go from the CPU's buffer chips to each of the EPROMs and RAM chips, though the pin numbers for the RAM chips will be different from the above EPROM pin numbers).

  Also check the CPU chip IC1 for pulsing at address lines A0-A7 (pins 9-16), data lines D0-D7 (pins 33-26), and perhaps address lines A8-A10 (pins 17-19).

• There is a chance the CPU chip at IC1 could be bad. It doesn't happen often, but it does happen. Remember system3 and system4 CPU boards use a 6800 or 68B00 CPU chip, and system6 and system7 CPU boards use a 6802 or 6808 CPU chip. A 6800 can not be put in a system6 board, and likewise a 6808 can not be put in a system4 CPU board! Keep this in mind.

  Note in System 6 and system7 CPU boards a 6802 or 6808 CPU chip can be used. Most system6/7 boards all shipped with 6808 chip, as these are slightly cheaper than the 6802 (which has onboard RAM). If a CPU board originally had a 6808, either a 6802 or 6808 CPU chip can be used without any jumper changes. If a system6 CPU board has a 6802 chip and a 6808 is desired instead, check to ensure that CPU jumper J1 is installed and that a 6810 RAM chip is installed on the CPU at IC13. The 6802 CPU chip is identical to the 6808, except it has internal RAM. This internal RAM can be used instead of the 6810 at IC13, on the System6 CPU board ONLY (this is why the system6 CPU board's IC13 is socketed and IC16 is soldered). I have never seen a System 6 board configured like this, but it could be done.

• At power-on, the blanking signal (pin 37 of the interconnector) should go from low to high (around 4 to 5 volts), just like the reset line. The blanking signal is a flag that goes to the driver board and says, "Hey, the CPU is up and running". When the driver board sees the high blanking signal, the driver board's PIA outputs are enabled, allowing the solenoids, lamp matrix and score displays to work. If there is a problem on the CPU board (or on the Driver board), the blanking signal will never go high. Note with Leon's test chip installed, the blanking signal will alternate from low to high once every second.
• Check the IRQ circuit, as generated by IC25 (4020) on system6/7 CPU boards, or IC23 (556) on system3/4 CPU boards. Using a logic probe, check IC1 pin 4 for the a pulsing IRQ signal (or check TP5 on System6/7 CPU boards).
• Don't forget to check the CPU socket IC1, especially if it is a Scanbe socket. That was replaced right?

How to Tell if a CPU Board is Locked-up or Good.
The following applies to CPU boards installed with game sofware (and *not* Leon's test EPROM).

For System3 to System6 CPU boards with game software installed, a good working and fully booted CPU board will, at power-on, flash both LEDs for a moment, and then turn both LEDs off. A locked-up System3 to System6 CPU board will have either or both of the LEDs turned on and stay on when the CPU board is powered on.

On System7 with game software installed, a good fully working and booted CPU board should flash the 7-segment LED or the two small LEDs next to the 7-segment, seeing a "0" briefly on the 7-segment display. Then the LED(s) go out and should *not* come back on. If there is a "0" from the instant the power is turned on, with no flicker or activity, then the System7 CPU board is locked up.

In addition, if the CPU board is up and running, the blanking signal will be high (around 4 volts). This is pin 37 of the 40 pin interconnector, or CPU connector IJ3 pin 4. The blanking signal can be fooled into being high on a non-working CPU board, or it can never go high on an otherwise working CPU. But for the most part, a working CPU board will have a high blanking signal.

How to Tell if Leon's Test EPROM is Running.
The most obvious thing that will happen if Leon's test EPROM is running correctly is this: On a System3 to System6 CPU board, the two LEDs on the CPU board will start flashing together, at about one second intervals. If the reset and clock circuit are working, and the address and data lines are active, both LEDs should start flashing about one second after the power is applied to the CPU board. On a System7 CPU board, a "zero" will flash every second on the 7-digit LED display.

The LEDs Come On & Stay On (or 7-segment display shows "0").
This is certainly the most common problem, and perhaps the most difficult to fix. By default, the two system3-6 LEDs or all the system7 LED segments come on and stay on when powered on. This happens because that's what the hardware dictates; the LEDs/segments do not turn off until the program in IC17 turns them off.

The Leon test ROM and the game program do this by setting the output ports PA6 through PA10 on CPU IC18 (the 6821 display PIA) low. If the test or game program isn't running, these ports remain high, and the LEDs/segments stay on.

When the MPU board is powered on, the CPU chip attempts to read addresses $FFFE and $FFFF in EPROM chip IC17 (the test ROM or flipper ROM 2) to obtain the jump address for the program to execute. This program then takes over and controls the LEDs. If the program is not able to start reading IC17, then the LEDs stay locked on.

The standard game software in IC17 then tries to read and write to the RAM chips IC19, IC16, IC13, and to all the PIAs on both the CPU and driver board (this of course assumes that the boot up process got far enough to run the program in EPROM IC17!) If there is any problem encountered, the program will lock up and not turn off the LEDs. The Leon test ROM however will run even if the PIAs or RAM chips are bad (or missing for the most part), so this narrows down the areas to check if the test ROM doesn't start its alternating LED on and off rythmic flash pattern.

Good Reset and Clock but No 'Leon' Alternating LED Flashes.
So the reset and clock test good, but the CPU board, with Leon's test EPROM installed, is still dead! What could it be? According to Leon's documentation, if the U1 CPU chip 6802 is good, his program should run (so try replacing the U1 6802/6808 chip with a known good chip, just to make sure). BUT I find this to not be the case. For example, the following things could make the CPU board not run Leon's program:

• Bad IC1 chip or socket (6802/6808 microprocessor).
• Bad IC17 EPROM chip (Leon's test EPROM).
• Bad IC17 socket (or any socket on the CPU board!)
• Bad IC18 PIA (6821) chip.
• Bad IC19 chip 5101 RAM. A really fried 5101 can stop Leon's program from working. If there is any question, remove the 5101 chip completely (Leon's program will work with no 5101).
• Bad IC11 (74LS10) and/or IC12 (7408) which are in the chip selection circuits.
• Address or Data line is broken (board trace broken). Using a DMM, make sure the address and data lines all show continuity to the EPROM (IC17), RAM IC13/IC16/IC19 chips, and PIA chip(s).
• Bad buffer chips at IC9/IC10 (CPU boards prior to System6A only).

Some of the above chips can be just removed from the CPU board with Leon's test EPROM. For example, the CMOS RAM at IC19 (5101) can be removed. Also the static RAMs at IC13/IC16 can also be removed. Removing these chips on a dead CPU board with Leon's test EPROM is a good idea. If these chips have a short, removing them may allow the CPU board to boot. This especially applies to the 5101 chip at IC19.

If the CPU board PIA at IC18 is bad, it may prevent Leon's program from flashing the LEDs. Using a test LED (as decribed below) and *not* a logic probe, connect the non-resistor end to +5 volts, and touch the other resistor lead of the LED to the CPU IC1 pin 15. If the test LED flashes on and off about once a second, then the test program is running, and the CPU board PIA at IC18 needs to be replaced.

The next thing to check now is if the CPU chip is running the program in the test EPROM. Use a logic probe and test CPU IC1 for pulsing activity on the address lines A0-A7 (pins 9-16), data lines D0-D7 (pins 33-26) (and perhaps address lines A8-A10 pins 17-19). If the CPU is running a program, there should be some pulsing on these pins. If even one of the A0-A7 or D0-D7 pins is low or high (not pulsing), then the CPU can not do its job and properly run the program in IC17. If there are pulsing signals on just some of these pins, then the CPU is trying to run, but the test program can't run correctly.

For both the address and data bus lines, use a DMM and check the continuity between lines. There should be no shorts between any of the data and address lines! Use a DMM and start with address line A0, checking continuity between IC1 address lines A1 through A15 (IC1 pins 9-20, 22-25), and data lines D0 through D7 (IC1 pins 33-26).

Solder splashes and solder bridges are quite common, especially if chip sockets have been replaced during previous repairs. If a socket was replaced, or testing a board that never worked, a previous repair attempt may have resulted in a short between lines.

While testing the address/data lines, its a good idea to test address a