PINBALL: Repairing Bally, Stern Pinball Games 1977-1985, part two

Support this Pinball Repair Website & PHoF. Please purchase the Marvin3m.com/top

Repairing Bally Electronic Pinball Games
from 1977 to 1985, Part Two.
by cfh@provide.net, 12/03/08.
Copyright 1999-2008 all rights reserved.

Scope.
This document is a repair guide for Bally electronic pinball games made from 1977 to 1985. Since Stern electronic pinball games use nearly identical electronics, these games are covered too.

Internet Availability of this Document.
Updates of this document are available for no cost at http://marvin3m.com/fix.htm if you have Internet access. This document is part two of three (part one is here, and part three is here).

IMPORTANT: Before Starting!
IF YOU HAVE NO EXPERIENCE IN CIRCUIT BOARD REPAIR, YOU SHOULD NOT TRY TO FIX YOUR OWN PINBALL GAME! Before you start any pinball circuit board repair, review the document at http://marvin3m.com/begin, which goes over the basics of circuit board repair. Since these pinball repair documents have been available, repair facilities are reporting a dramatic increase in the number of ruined ("hacked") circuit boards sent in for repair. Most repair facilities will NOT repair your circuit board after it has been unsuccessfully repaired ("hacked").

If you aren't up to repairing pinball circuit boards yourself or need pinball parts or just want to buy a restored game, I recommend seeing the suggested parts & repair sources web page.

Table of Contents

1. Getting Started:

2. Before Turning the Game On:

3. When Things Don't Work:

3a. When things don't work: Making a MPU Test Fixture.

The J4 MPU Board Connector.
If you don't want to make a work bench test fixture for repairing your Bally MPU board, there is an easy way around this. Assuming your Bally power supply is working correctly in your game, when testing a Bally MPU board, you only need to use connector J4 on the MPU board. The J4 connector is the power connector. With it only attached and a working power supply, you can quickly test your Bally MPU right in the game.

Making a Test Fixture.
To make a test fixture, you will need some sort of power supply that will deliver +5 and +12 volts. You can buy a video game switching power supply (about $25), or you can use an old computer power supply. Either will work fine. Old computer AT or ATX power supplies are plentiful, and often $15 or less used at computer stores.

On the left is a video game switching power supply.
On the right is a used computer AT power supply. Either
will work fine for a test fixture. Note the computer
power supply connector on top of the box is the one
that will supply our +5, GND and +12 volts. This plug
was used to power drives (hard drive, CD rom, etc).
Red is +5, black is GND, and yellow is +12.

Also remember switching power supplies like the AT and ATX are switch-mode power dependant on load. This means that the power is switched on and off rapidly to meet the load. If there is no load (nothing attached), the output power is switched off. So using a DMM (digital multi-meter) alone to test the power supply output voltages may not be enough draw to make the power supply turn on.

In either case, all you need to do is connect the power supply to a wall plug (110 vac), and put aligator clips on the GND, +5 vdc and +12 vdc. That's all you need to power up a Bally MPU board on your workbench! On a computer power supply, the best way to do this is to use one of the plugs that connected to the computer's hard drive, floppy drive, or CD ROM drive. There are usually at least three of these 4 pin plugs on each computer power supply. These plugs have two black wires, and one red and one yellow wire. The two black wires are ground, the red wire is +5 vdc, and the yellow wire is +12 vdc. I used some crimp-on male connector pins and jammed them into the plug. Then I connected these to aligator clips, which ultimately go to your MPU board.

Using a computer power supply on the workbench to
power a Bally MPU board.

TP5 = +5 volts (right next to the battery). Red power supply lead.
TP4 = GND (opposite end of the board from the battery). Black power supply lead. Don't get this confused with TP1, which is right next to it!
TP2 = +12 volts (on the same side of the board as the battery; don't get this confused with TP3, which is right next to it!) Yellow power supply lead.
TP3 = +21.5 volts (right next to TP2). No power supply lead since neither a computer power supply or a switcher supply this voltage. This assume this is *not* a -133 MPU board. If a -133 (with a 1N4148 diode at R113, as used on Granny and the Gators, Baby Pacman, & Grand Slam), temporarily convert R113 from a 1N4148 diode to a 2k ohm resistor.

Once you have your test fixture constructed, you can diagnose Bally MPU problems with the flashing LED on the board easily. Instead of putting the board back in the game to test it, just turn your power supply on and count the flashes!

Warning!!

Problems Using the Computer Power Supply.
The original linear power supply for the game powers on with 5 volt very quickly. That is, within 50 miliseconds or so. Remember the purpose of the MPU board's reset section is to hold the U9 6800 pin 40 low for a short period of time, allowing the +5 volts to stablize. Then the U9 6800 pin 40 goes high, which starts the 6800 running code out of U2/U6 ROMs.

The problem with *some* computer switching power supplies is they take longer than 50 miliseconds to get the 5 volts up and stablized. If you are using one of these slower switchers, a perfectly good MPU board may not boot with that switching powe supply (that is, the MPU LED will be lock on). Until you find a faster computer power supply, you can get the MPU board started by shorting U9 6800 pins 39 and 40 for just a moment (I use a small screwdriver to do this). This simulates the reset section of the MPU board manually, holding pin 40 low for moment. This should start the boot-up process and the MPU LED flashing.

To get the last flash from the MPU board, 20 to 25
volts DC is needed at test point TP3 on the MPU (or put a
temporary alligator jumper wire from chip U12 pin 3 to U14 pin 14).
This can be done by purchasing an inexpensieve 25 volt
transformer and bridge rectifier. I screwed this make-
shift power supply to a board for easier use.

The Last LED Flash (missing voltage at TP3).

There are *two* ways to solve this problem. The first is to trick the MPU board into thinking the 43 volts is present. To do this, put a temporary jumper wire on the MPU board, connecting chip U12 pin 3 to chip U14 pin 14. This can be done easily with an alligator jumper wire, connecting the top leg of resistor R23 (leg closest to chip U12) to the top leg of resistor R17 (leg closest to TP3). This is certainly the easiest approach.

The other solution is more realistic, but involves some parts (a transformer and a bridge rectifier). Radio Shack sells a 120 volt AC to 25 volt AC transformer (#273-1366) for $6 that works well for this. They also sell a 25 amp 50 volt bridge (#276-1185) for about $2.70, which is also needed.

To wire this up, connect the two black transformer wires that are together on one side of the transformer to a 120 volt power cord. The remaining three wires (two yellow, one black) on the other side of the transformer are the output. Connect the two yellow wires to the AC terminals of the bridge. The black (center tap) transformer wire you won't use. The bridge should have a label signifying at least one AC lead. The second AC lead is diagonal or across from the labeled lead.

The J4 connector of the MPU board. Notice
the TP3 and TP2 test points at the upper right
corner of this picture. The R113 resistor takes
the +43 volts from J4 pin 15 and lowers it to
+21.5 volts at TP3.

The output from the bridge will be about 25 volts DC. 43 volts DC is not needed because this connects 25 volts to the MPU board's TP3 test point, which is looking for 21.5 volts DC. What happens is the 43 volts DC comes from the rectifier board and goes to MPU connector J4 pin 15. Then the 43 volts goes through resistor R133, which lowers the voltage to 21.5 volts. Then a "zero crossing" voltage detection circuit looks for that 21.5 volts. If present, a working MPU board will flash the last time. If the voltage is missing, there is no last LED flash.

So instead of supplying 43 volts to connector J4 pin 15, we just skip resistor R133 and connect our 25 volt DC power supply directly to TP3. It's pretty easy to get a 25 volt transformer, but getting a 43 to 50 volt transformer is more difficult. The zero crossing circuit will work fine if you supply 25 volts DC (instead of 21.5 volts) to test point TP3.

The voltage supplied above with the transformer and bridge gives 25 volts DC to test point TP3. This DC voltage is NOT smoothed with a filter capacitor! This is done for a reason. The original Bally power supply doesn't filter this voltage, and for good reason. The zero crossing circuit requires the DC voltage "ripple" to work!

(Re-)Booting a MPU Board with the Test Fixture.
After connecting the test fixture to the MPU board, simply turn the power supply on and count the LED flashes. To re-boot the MPU board, simply short together (with a screwdriver blade) pins 39 and 40 of the U9 CPU. This will quickly reset the CPU, and the flash sequence will start over, without having to turn the power supply off.

3b. When things don't work: Fixing the MPU (LED flashes and the such).

Skip right to the LED Flash codes by clicking here.

LED Flash Code Introduction and Assumptions.
I am assuming if your MPU board had any battery corrosion, that you have fixed it (see the Removing the MPU Battery and Fixing Corrosion section). Fixing a MPU board with corrosion is like a dog chasing its tail; a never ending path of frustration. Corrosion can cause intermittent problems. Take care of the corrosion first before trying to fix a MPU board.

Fixing a Bally MPU board, generally speaking, is one of the easiest game boards you will ever repair. The reason for this, is Bally uses a "LED flash test". The green LED (light emitting diode) lamp is right next to the battery on the MPU board. As the MPU board boots, it does self diagnostics. For each diagnostic test the MPU board runs, it flashes the LED when that diagnostic is complete. If a diagnostic fails, the boot up process stops, and the LED stops flashing. There are a total of 6 or 7 LED flashes during the boot-up process. This makes determining MPU board problems fairly easy.

Preliminary Step: Power.
The Bally/Stern MPU boards need only 12 volts and 5 volts to start booting (43 volts is needed later, but for now let's keep it simple). Before starting the LED flash code diagnositics, check the power train. Always check the end of the powerline first, the MPU board. It must have good 5 volts DC at TP5 and an unregulated 12 volts at TP2.

If the LED is lit solid and the MPU is not booting, it can be assumed that the 12 volts is present (because 12 volts is needed to light the LED). So check for 5 volts DC at TP5, it should be in the 4.8 to 5.1 volt DC range. Many times the MPU is locked only because the 12 or the 5 volts are missing or low.

If the MPU LED is not lit, check TP2 for 12 volts DC. If 12 volts is there but LED does not (never) lights, cross the two legs (on right side) of Q2 with a small screwdriver. This should turn-on the LED. If it doesn't, assume that the MPU board's LED is bad and replace it.

If the 12 volts is missing from MPU TP2, forget about the LED and go to the rectifier board and check for voltage there. The following info applies mostly to games prior to Xenon (where the power supply is in the backbox). But it can be applied to the later games, especially when testing the 12 volt and 5 volt DC circuits of the solenoid driver and MPU boards. Different wire colors may be used in newer games, but the test points (TP) and voltage readings are the same.

Check rectifier board TP3 for 12 volts. If missing, check the fuse F3. Remove the 4 amp slo-blow and buzz it out with a DMM. Check the fuse clips for damage or burns, as this is very common.

If still no 12 volts DC at TP3, check E11/E12 test points on the rectifier board. These are the two AC inputs and should read about 12 volts AC with your DMM connectors to both E11 and E12. If no voltage check the wiring to the transformer or the transformer itself. If the 12 volts AC is present at E11 and E12, then the bridge rectifier BR2 is open. If fuse F3 blows immediately at power-on, then the bridge rectifier BR2 is shorted.

If there is 12 volts DC at TP3 on the rectifier board but no 12 volts on the MPU board, likely it is just one wire causing the problem. Check J3 pin 8 (orange wire) of the rectifier board, as it may be burnt. It is a good idea to replace this connector pin as it handles the 12 volts distribution and is often burnt.

Next the orange wire goes to the solenoid driver board connector J3 pin 12. It makes a U-turn and heads back out J3 pin 11 and goes to the MPU board (this is why there is often a jumper between the solenoid driver board TP1 and TP3, to make sure this U-turn does not get cut). At the Solenoid Driver, the 12 volts DC heads also heads to the 5 volt regulator going through the big capacitor C23 (this smooths out the raw 12 volts DC). Often there is a problem with the begative side of cap C23 that causes problems. The negative side of this cap goes back to the Power Supply in a separate wire, independent of other ground circuits. Even though all grounds end up at the same spot, this independent wire helps limit line noise. It also can lead to some weird problems like shutting down the MPU or causing massive resets.

This single white/brown return wire goes from the solenoid driver from J3 pin 10 to the rectifier board J3 pin 17. Note that these pin numbers at the rectifier board may not the same the game you are working on. The grouping of the pins at rectifier board connector J3 are often swapped around between pins 1-4 and pins 14-20. These are all ground returns and over the years perhaps the wires were moved. Keep this in mind.

The point is if this ground return wire from the solenoid driver board capacitor C23 to the rectifier board gets cut, it will shut down the 5 volt regulator and the MPU. The fastest way to do a check is to jump the left (negative) side of the cap C23 to any ground in the backbox. If the game comes up, you have found a problem with the white/brown wire.

You should now have +5 and +12 volts DC at the MPU board, and can start diagnosing the MPU LED board flash codes.

First Step: Count the LED Flashes.
The first step in fixing a non-working game is to count the LED flashes. On a working MPU board, when you turn the game on, the LED should "flicker" (a very faint and quick flash), and then flash seven times (only six times on Baby Pacman or Granny and the Gators). If the board does not do this, the internal diagnostics have determined a problem. At this point you should remove the board from the game, and hook it up to your test fixture (which you made above in the Making a MPU Test Fixture section). Working on your MPU board on a workbench is much easier than trying to fix it, and then test it in the game itself.

The Last Flash.
If you game does the initial flicker, and then follows with all the flashes but the last flash (flash number seven, or flash six on Baby Pac/Granny), check your +43 volt solenoid power F4 fuse on the rectifier board before proceeding. This last flash verifies that the game has +43 volts for the solenoids. If it doesn't (the F4 fuse is blown), the game won't boot and won't finish the flash sequence.

No Flashes: the LED is Permanently On.
If you turn your Bally game on, and the LED stays on continually, this is one of the hardest problem to fix on these boards. A stuck on LED can be caused by any (or all!) of the following:

The reset circuit is damaged (after all, it's in the battery corrosion area!) This includes transistors Q1 and Q2 (both 2N3904), Q5 (2N4403), and diodes CR5 (1N4148 or 1N914) and VR1 (1N959B or 1N4738A).
The program ROM (read only memory) at U6 is bad.
The ROM jumpers are incorrectly set for the U6/U2 ROMs installed.
The U9 CPU microprocessor is bad.
The U11 PIA (peripheral interface adapter) is bad.
ANY of the sockets for chips U6, U9, U11 are bad.
The ground trace to U11, U6 or U9 are bad. U11 is especially important to check the ground trace, as it's right in the corrosion zone. And corrosion just loves to travel up the ground traces. Use a DMM set to continuity to check ground at U11 pin 1 (and U11 pin 20 for +5). Also check U6 pin 12 for ground (and U6 pin 24 for +5). U9 pin 1,21 (gnd) is less of a problem as it's well outside the corrosion area.

The first and last two points are probably the most common problems.

Go "Bare Bones" (What to Remove for the first "Flicker").
To get the initial flicker out of the MPU board's LED, you only need three chips on the MPU board. These three chips are U9 (the CPU), U11 (PIA) and U6 ROM (except on some Stern games, which also require the U2 and/or U5/U6 ROMs). Remove all other socketed chips! Once you get the initial flicker out of the MPU board, then you can re-install the other removed chips.

The Reset Circuit.
If the reset circuit is damaged, the MPU board will never start, and the LED will stay on at power up. Since the reset circuit is in the "corrosion zone", it is often damaged from battery corrosion. The reset circuit holds pin 40 (reset) of the U9 CPU low until +5 volts is "stable", and then sets pin 40 high (telling the CPU it can start the boot up process). While the reset section is holding pin 40 of the CPU low, the LED will be on. This is one part of the initial "flicker" seen when a working Bally board is first powered on.

Here's what to try first:

With the MPU board OFF, Put a logic probe or DMM set to DC volts on pin 40 of the U9 CPU. The logic probe will be "low" (or the DMM at zero volts).
Power the MPU board on.
In just an instant, the U9 CPU Pin 40 should go "high" on the logic probe, or about 4.75 to 5.5 volts on the DMM. Repeat this for U9 pin 2, which also should also go high.

If the above happens, the reset circuit is probably working fine. If pin 40 of the CPU is low and never goes high, there is a problem with the reset circuit. Since pin 40 never goes high, the CPU will never start the boot process, and the LED will stay on.

Here's some other stuff to try:

With the MPU board powered on and pin 40 of the U9 CPU high (4.75 to 5.5 volts), short the junction of resistors R1 and R3 to ground. To do this, find the junction where resistors R1 and R3 connect together (the right side of R1 is the junction of R1 & R3). With the MPU board on, use a wire and short this junction to ground (TP4).
U9 CPU pin 40 should now go low to zero to .5 volts while the R1/R3 junction is grounded.

The above procedure is simulating what the reset circuit is meant to do. If shorting the junction of R1 and R3 does not pull U9 Pin 40 from high to low, then there is a problem with the reset circuit. Most likely there is a problem with Q1 (2N3904), Q2 (2N3904) and/or Q5 (2N4403). Replace them all (they are cheap!) and repeat the above procedure. Also check/replace diode CR5 (1N4148 or 1N914). Finally check/replace diode VR1, which is a 1N959B (1/2 watt) or 1N4738A (1 watt), 8.2 volts (note VR1 is mis-labeled in the manual as "1N9598"). If you still can't get U9 CPU pin 40 to go low, there is some other problem in the reset circuit (corroded traces?).

Here are some other points to check. Turn the MPU board on and test:

U9 pin 3 = 2.4 volts
U9 pin 36 = 2.6 volts
U9 pin 40 = 4.75 to 5.5 volts
U9 pin 5 = 2.8 volts

If the above voltages check out, then the reset section is probably working correctly. This means your problem probably lies with chips U6, U9 or U11 (or their sockets), or incorrect MPU board jumpers.

A Reset Trick (LED stuck On) - Adding a Capacitor.
If having problems with the reset (LED stuck on), try shorting together (with a screwdriver blade) pins 39 and 40 of the U9 CPU. This will manually reset the CPU, since reset pin 40 gets set low by ground, pin 39. Essentially, the screwdriver blade is doing what the reset does. So if the flashes start and proceed, there is a problem with the reset section of the MPU board.

It could be the +5 volts is not stablizing within the "window" the CPU expects (about 50ms). To make the reset window longer, try putting a 470 mfd electrolytic capactor in the reset section (on the solder side of the MPU board). Solder the positive leg of this cap to the collector (top leg) of Q1. Then solder the negative leg of the cap to ground (the emitter of Q1). This will increase the reset timing length. Now test the MPU board. If the MPU board works and continues with the flashes, the +5 volts was just not getting "stable" in the 50ms window (there could be a power supply problem, probably with the large 12 volt C23 rectifying capacitor on the driver board).

One note about this added reset capacitor. Since the 470 mfd cap increases the reset timing (the time pin 40 of the CPU is held low), the initial "flicker" may not longer be a "flicker". The LED will stay on longer (because the reset is longer), making the "flicker" more of a "flash".

Buy a Bally Reset Section Repair Kit.
Having problems with the reset section on your Bally board? Instead of ordering all the parts separately, I suggest just buying a "Bally Battery Corrosion Repair Kit" BALLY35-BA-KIT from Ed Krzycki (gpe@cox.net). This kit includes all the resistors, capacitors, diodes, transistors and chips typically gone bad in the reset section. For a mere $10, or $4.50 (without a 5101 RAM chip), this kit is well worth it. See Ed's web page at www.greatplainselectronics.com for more information. When buying the kit, install *all* the parts, even if the originals "look good".

If for some reason the $4.50/$10 is too much money, here are the typical parts needed for reset section repair:

Q1,Q2 - Transistor, 2N3904
Q5 - Transistor, 2N4403
VR1 - Diode type 1N4738A, Zener, 8.2V (alternate for 1N9598)
CR8 - Light Emitting Diode, Green.
CR5,CR7 - Diode, Switching. 1N4148
CR44 - Diode, Rectifier. 1N4004 (or better)
C1,C2 - Capacitor, 820pF, Axial Ceramic.
C5 - Capacitor, 4.7uF, Radial Tantalum.
C3,C13,C80 - Capacitor, 0.01uF, Axial Ceramic.
R1,R3,R24,R28 - Resistor, 8.2K, 1/4W, 5%
R2 - Resistor, 120K, 1/4W, 5%
R11 - Resistor, 82, 2W, 5%
R12 - Resistor, 270, 1/4W, 5%
R16 - Resistor, 2K, 1/4W, 5%
R17 - Resistor, 150K, 1/4W, 5%
R29 - Resistor, 470, 1/2W, 5%
R107 - Resistor, 3.3K, 1/4W, 5%
R112 - Resistor, 1K, 1/4W, 5%
R134 - Resistor, 4.7K, 1/4W, 5%
R140 - Resistor, 20K, 1/4W, 5%

Using a Dallas/Maxim DS1811 in the Reset Section.
There is also another way to fix the reset section. This involves using the new Dallas/Maxim Semiconductor D1811 reset chip (TO-92 package). This single reset chip should replace a whole bunch of the stock reset components on a Bally MPU (thanks to Neil for doing the leg work on what can be removed). This inexpensive Dallas device looks like a transistor, but is really a three leg chip in a TO-92 transistor package (I also show how to use this same reset chip on Gottlieb System80 MPU boards in the Gottlieb System80 Repair Guide). Click here or here (PDF, more info) for the specs on this chip. Cost is less than $1, and can be ordered directly from Dallas/Maxim Semiconductor at www.dalsemi.com via their phone number 888-629-4642 (but orders must be faxed in at 408-222-7174). Be sure to order the TO-92 package (part number DS1811-10), as this chip also comes in a surface mount SOT23 package.

Interestingly, the new Alltek Systems replacement MPU board (allteksystems.com) also uses a Dallas-Maxim reset chip. The only difference is their board uses a DS1233 (which is the same as the DS1811, except the reset timing is 250ms, where the DS1811 is 150ms).

The Dallas DS1811 comes in three TO-92 flavors of "normal reset threshold":

DS1811-15 = 4.13v
DS1811-10 = 4.35v
DS1811-5 = 4.62v

The DS1811 is installed with pin 1 going to /RESET (Q5's leg closest to resistor R11), pin 2 to +5 volts, and pin 3 to ground. If the MPU board boots fine by shorting pins 39 and 40 of the U9 CPU, installing this part may be the "quick fix" to your reset problems. Also the benefit of removing many of the stock reset components on boards with battery corrosion certainly helps.

There is a side affect of the changed reset circuit: Since the Dallas DS1811 increases the reset timing (the time pin 40 of the CPU is held low) from about 50ms to 150ms, the initial "flicker" may not longer be a "flicker". The LED will stay on longer (because the reset is longer), making the "flicker" more of a "flash".

Installing the Dallas DS1811.
Here are the steps to replace the Bally reset section with a single Dallas DS1811 reset chip.

Replace R11 with a *new* 2 watt, 82 ohm resistor.
Remove Q1, Q5, R140, R138, R139, VR1, R112, R1, R2, R3.
Put a jumper from the lower hole of R138 to the upper hole of R139 (note "lower" and "upper" being relative to the MPU board in its normally installed position in the game).
Connect the reset pin of the DS1811 (bottom pin, as the flat side of the DS1811 is towards connector J4) to the collector hole of Q5 (the single left most hole of Q5 closest to connector J4). Wait to solder this connection until the next step.
Add a 1k ohm pull-up resistor from +5 volts to the DS1811 reset pin (which is connected to the left most hole of Q5). In the picture below, the banded side of CR5 was used to access +5 volts.
connect the VCC pin of the DS1811 (the middle pin) to +5 volts, which is the bottom right pin of Q5.
Connect the GND pin of the DS1811 (the top pin) to ground, which is the top right pin of Q5.
Leave CR5, CR7 installed.

Replacing the reset section with a Dallas DS1811 reset chip. Picture by Neil.

As Neil explains, replacing resistor R11 is a good idea. These carbon resistors eventually heat up and break down, and once you replace the rectifier circuit with better diodes, the leakage is smaller and that results in around 14 volts on the 11.9 volt line. This is enough to push the stock resistor out of spec. It will heat up and change resistance, and in some cases will go open, which causes a reset. As long as R11 is a 2 watt resistor, it should be fine. If the current resistor is discolored or of a carbon type, it should be replaced.

A "Good" U6: EPROMs, Masked ROMs, and Jumpers.
If the LED is stuck on, next give yourself the advantage of knowing that at least the program jumpers match the type of Game ROMs being used (EPROMs, Masked ROMs, etc.). Jumper are discussed in the Game ROMs, EPROMs, and Jumpers section below. Basically these jumpers tell the MPU board what type, size, and how many game ROMs are being used on the MPU board. Bally did this as a convenience factor; if you wanted to use an older MPU board in a newer game (that had a larger program and hence larger or more ROMs), you could. All you needed was to change the jumpers. It also allowed you to use the board with EPROMs (erasable, programable read only memory), or factory created "masked" ROMs. The factory masked ROMs are black with numbers screened on them. These are not re-programable, and must be used in the game they are intended for. EPROMs on the other hand, can be re-programmed and re-used (if you have an EPROM programmer, which costs about $150 and can connect to your computer). An accurate jumper chart is mandatory here and a basic knowledge of how to recognize the different types of ROMs is also a plus.

U6 is the program ROM chip which contains the code for the boot-up diagnostics (and on some Stern games, U2 and/or U5/U6 ROMs are also required for the initial "flicker"). If you don't have a good U6, and don't have the MPU jumpered properly for that U6/U2, the diagnostics can't even start to run. If you don't have an EPROM programmer to make your own U6 EPROM, contact Tom Callahan and order the ROMs needed for your particular game. Make sure you plug them into the sockets correctly (notch oriented correctly) when you install them. Also make sure you have the jumpers set correctly for the type of ROMs you are installing.

Don't Change the MPU board's ROM jumpers unless you have to!
If the U6/U2 ROM jumpers are incorrectly set on your MPU board, this can cause a locked-on MPU board LED. For this reason, I suggest not changing the MPU board jumpers until getting the initial flicker from the LED. If the jumpers are changed, TWO problems could exist instead of just one! If not sure if the U6 ROM is good, try and install a known good ROM of the same type (or test the U6 ROM in another game). This way the board jumpers don't need to be changed.

Using 2532 EPROMs instead of 9332 Masked ROMs.
If the game in question is a later Bally game with 9332 masked ROMs, these can be changed to 2532 EPROMs with NO jumper modifications! This can be handy and convenient if the original black 9332 masked ROMs need changing, but the repair person doesn't want to mess with the jumpers.

A "Good" U9 and U11.
These two chips MUST also be good to get your MPU board to even start to work. The diagnostics won't even start to run if you don't have a good CPU chip at U9. Also a good PIA chip at U11 is required.

If your LED is locked on, and the reset circuit is working (see above), I would next invest in a new 6800 CPU for U9, and a new 6821 PIA for U11. You can buy these for just a few dollars. Get a couple extras of each chip. While you're at it, get a few 5101 RAM chips too (as these go bad a lot on the MPU board).

What Chips are Required to Get the MPU "Unlocked"?
The only socketed chips that are required to get the MPU board to do its first flicker, are a working U6 (program ROM), U9 (the CPU), and U11 (a PIA). Of course this assumes the reset circuit is working properly. All other socketed chips can be removed until the board becomes "unlocked" and does its first flicker.

How Can I Ensure I have good U6, U9 and U11 Chips?
I keep a working MPU board around, and plug the unknown chips (one at a time!) into the working MPU. Then I power the MPU on and see if it still works. This works best for me, but you may not have a good MPU spare laying around.

If you have no spare working MPU, make sure you buy some new U9 (6800 CPU) chips and U11 (6820 PIA) chips. They are inexpensive and worth keeping "in stock". And get a set of known good EPROMS (U6/U2) too. You can buy these from many sources if you don't have an EPROM programmer

Good Chip Sockets (replace the "closed frame" sockets!)
If the sockets for U6, U9 and U11 are "bad", nothing will make the LED start flashing until this is fixed. If the sockets are the closed frame brown or black sockets (closed frame means the circuit board can not be seen under the socket), or have the brand name "SCANBE" or "RS" impressed into the sockets, replace ALL of them! These types of sockets are known to be troublesome. Another approach to replacing these problem sockets is to plug a new machine pin socket into the problem socket, then plug the chip into the new socket. This is a good temporary solution until you get the MPU board working. Then you can later replace the old bad sockets with new machine pin sockets. Using your DMM, buzz-out the chip legs to the under side of the MPU board to make sure the sockets are good. Note: do NOT get gold plated sockets! These react with the different metal in the chip legs, and can cause intermittent problems. Also remember that U11 is in the "corrosion zone". This means this socket could have easily been affected by battery corrosion.

A brown and a black closed frame socket.
There's no way around it, ALL these sockets will
need to be replaced on any Bally MPU board.

Re-seating the Chips.

Component layout on the Bally MPU.

Do I have the Right Jumpers/ROMs Installed?

On the CPU chip U9 pin 5 is the VMA line (valid memory address), and it should never be stuck HI or LOW (it "polls" the U6 ROM). A bad U6 ROM can affect the CPU U9 pin 5's behavior. The CPU's U9 Pin 5 behavior is controlled by U9, but tying it to a bad or mis-jumpered U6 ROM can cause many problems.

So if you are confused about the ROM jumpers, try this: Boot the MPU board with only U9 and U11 installed (all other socketed chips removed). The LED will stay on, but using a logic probe check U9 pin 5. It should be pulsing. Power down, add ROM U6 and repeat. If U9 pin 5 is now stuck high or low, U6 is either bad, or you have the wrong jumpers.

The LED is Still Locked On.
If the U6, U9, U11 chips and sockets are known to be good, the reset section tests good, and the board is jumpered correctly, you should be getting at least a initial flicker from the LED. The inital flicker indicates these three chips are good, and that the MPU has started its boot-up process (a good sign!) Remember, on some Stern games, a good U2 and/or U5/U6 ROMs chip are also needed to get the initial flicker.

If the reset section is working, get a head start and put in your own "known good" U9, U6 and U11 (and re-jumper the board if necessary). This will usually get at least a flicker out of the LED. If the LED is still locked on, here's some other things to test:

Reset Section Components to Replace:

If the zener diode VR1 (1N959B or 1N4738A, 8.2 volt) on the MPU board is bad, the LED will remain on.
If the reset transistors Q1 (2N3904) and Q5 (2N4403) are damaged, the LED will remain on. As a general rule, it's always a good idea to replace these inexpensive transistors.
If diode CR5 (1N4148 or 1N914) is bad, the LED will remain on.
If transistor Q2 (2N3904) is bad, it could leave the LED on (regardless of what the MPU is doing!)

After the Reset Section is Rebuilt, try this:

If your CPU board has brown sockets, or sockets labeled "SCANBE" or "RS", you may want to replace ALL the sockets. These sockets are known to be problematic. An alternative to replacement is to plug a machine pin socket into the problem socket, then plug the chip into the new socket. This is a good temporary solution until you have the MPU board working.
Measure the voltage at the CPU board. Even if you are using your test fixture, measure the voltage right at the chips to verify. The +5 line should be 4.8 to 5.2 volts DC. Check the power at the required socketed chips (U6, U9, U11). This will tell you if power is getting through the board to the chips
- U6, U2 (ROM): +5 = pin 24, GND = pin 12
- U9 (CPU): +5 = pin 8, GND = pin 1 or 21
- U10,U11 (PIA): +5 = pin 20, GND = pin 1
- U8 (5101): +5 = pin 22, GND = pin 8
- U7 (6810): +5 = pin 24, GND = pin 1
You don't really need to check the +12 at the board. If the LED is lit, you know the +12 volts is already present.
Using a logic probe, check U9 (CPU) pin 2 and pin 40. They should be high. Pin 40 is the Reset line, and this pin will start LOW at initial power-on (for about 50 ms), and then go HIGH. As you turn the power on, the logic probe's low LED should light briefly, then the Hi LED should turn on and stay on. If there is no initial LOW followed by about 4.5 volts at pin 40, the "valid power" reset circuit in the lower left corner of the board is at fault. The reset section's job is to hold the CPU pin 40 low until the +5 volts stablizes, and then allow pin 40 to go high to 4.5 volts. If pin 40 stays low, the board will never reset, and the LED will stay locked on. Most likely there is corrosion damage to transistors Q1 (2N3904), Q2 (2N3904), and Q5 (2N4403). Also check diode CR5 (1N4148 or 1N914). Replace all of these if you suspect them as bad. Or better yet, order a reset rebuild kit from Ed (about $10) at www.greatplainselectronics.com, and rebuild the entire reset section.
If U9 CPU pin 40 is high (about 4.5 volts), then jump the junction of R1 and R3 (the right side of R1) to ground. Now check U9 CPU pin 40 again. This pin should be low (about .5 volts). This is simulating what the reset section does at power-on. If jumping R1/R3's junction to ground does not give .5 volts at U9 pin 40, there is still a problem with the reset section. Try replacing the reset transistors Q1 (2N3904), Q2 (2N3904), and Q5 (2N4403). Also check the diode CR5 (1N4148 or 1N914). To do this, put your DMM on diode setting. You should get a reading of .3 to .6, and reading of null (zero) when the leads are reversed. Or again, install one of Ed's reset section rebuild kits (which includes all the reset parts needed).
Using a logic probe, make sure chip U9 (the CPU), pins 3 and 36, 37 are "pulsing". These are the "clock" signals. If you don't have a logic probe, put your DMM on these pins and you should get about 2.5 volts (the average of the pulsing waveform). If you don't have a pulse or about 2.5 volts, suspect chips U15 (MC3459L or 74S37) and U16 (9602 or NTE9602). Also sometimes the C14 and/or C15 (470 pf) capacitors have failed. However the most common cause is a bad U15 (MC3459L or 74S37).
Check that U9 (CPU) pin 2 is high (+5 volts). This is the Halt signal on the CPU. If it's not HIGH, check for a broken or missing resistor at R135 or a short on the J5 connector.
Using a logic probe, check that U9 (CPU) pin 5 is pulsing (or use your DMM and you should get about 2.8 volts). This is the VMA (valid memory address), and is an output from the CPU. If it's not pulsing, the CPU is not running! It could be a bad U9 chip, or some fault on the address and/or data lines cause U9 to "crash". Watch U9 pin 5 on the logic probe after turning the power off and then back on. You should see it pulse for a second or two before stopping. If any CPU output pin is shorted, this can cause the CPU to crash.

Left: Chip U9 (CPU),
pin 3. Note pins 36, 37
will have the same
signal pattern too.
This is the clock signal.
Middle: Chip U9 (CPU),
pin 40 and pin 2. Should
be a straight line "high",
except initially at
power-on.
Right: Chip U9 (CPU),
pin 5. This is the Valid
Memory Address line.

Trace the VMA signal through its path. Using a logic probe, test the following path. You should get a pulsing signal at all points while the game is in attract mode (or when the MPU is on the bench in your test fixture):
1. U9 pin 5 (VMA line on the CPU): Pulse
2. U19 pin 8 (input to U19): Pulse
3. U19 pin 9 (input to U19): HIGH
4. U19 pin 10 (output from U19): Pulse
5. U14 pin 11 (input to U14, about 1.3 volts measured with a DMM): Pulse
6. U14 pin 12 (output from U14): High Pulse
7. U15 pin 4 (input to U15): Pulse
8. U15 pin 5 (input from U16 pin 10): Pulse
9. U15 pin 6 (output from U15): Pulse
10. U15 pins 1, 2 (input to U15 again; both pins tied together): Pulse
11. U15 pin 3 (output from U15, about 1.8 volt measured with a DMM): Pulse
Tracing the above, you can see where the signal runs astray. Replace the suspected component. U15 (MC3459L or 74S37) seems to be the component that fails the most here, followed by U14 then U19.

Left: U9 (CPU) pin 5, and U19 pin 8.
Right: U19 pin 10, and U14 pin 11.

Left: U14 pin 12, and U15 pin 4
(U14's wave form is tighter)
Right: U15 pin 5
(only seen on a fully booted MPU).

Left: U15 pin 6, and U15 pins 1 and 2.
Right: U15 pin 3.

Check U9 (CPU) address and data lines with your logic probe for pulsing signals. This includes pins 9 to 20, 22-24 (address lines, pin 24 is only used if 2732 EPROMs installed), and pins 26 to 33 (data lines). Note pin 21 is ground.
If the PIA at U11 is bad (or its socket is bad), all the above pulses can be present and acting correctly. Yet the LED will still be stuck on. This PIA is right in the "corrosion zone", so make sure the U11 socket and chip are in good shape.
If C23 on the solenoid driver board is not filtering the full wave rectified DC volts effectively, this can cause the MPU to lock and leave the LED on (see Upgrading the Voltage Regulator/Solenoid Driver Board for details on this).
If Q20 (the +5 voltage regulator) on the Solenoid Driver board is bad, the LED will remain on.

If you've gotten to this point, and the MPU's LED is still staying lit, you may want to consider sending the board out for repair.

Strengthen the Reset Signal by Replacing Caps C14 and C15.
The two capacitors at C14 and C15 (470 pfd) should also be replaced. If these capacitors are weak, the reset signal will not be as crisp and strong. This is very apparent if you have a scope.

Was there Work Done on Chips U15 and U16?
If chips U15 (MC3459L or 74S37) or U16 (9602) were worked on, suspect problems. Even though these chips are not directly involved in the reset circuit, there are very fragile traces which travel underneath the U15/U16 sockets. If someone replaced these chip and/or sockets, these traces could be damaged. Using a ohm meter and check for continuity on all of these traces (this may require removing the U15/U16 sockets; always replace these sockets with strip socket headers, so the traces maybe seen and worked on easily in the future).

More on the RESET Signal and a Locked-on LED.
A major part of the reset circuit is transistor Q1, and there is an easy way to test it. As we learned above, when the game is first turned on, the reset line (pin 40 of the U9 CPU) goes low for just a moment, then sticks high. You can similate a reset by shorting the ground junction of resistors R1 and R3. This will force a reset in a good MPU board. As you do this, watch pin 40 of U9 with your logic probe. It should go low for a moment, then stick high. If it doesn't, try replacing Q1 (2N3904). If Q1's collector doesn't go to about 4 volts DC when R1/R3 junction is grounded, then Q1 is bad.

Q5 gets Cooked by R11, and the Game Re-boots Intermittently.
Another strange problem relates to transistor Q5 (2N4403). This transistor is also part of the reset circuit (along with Q1), and is located right below R11, an 82 ohm 2 watt power resistor. The problem here is resistor R11 gets HOT, and is sometimes touching Q5. The heat is transfered to Q5, and this can cause the MPU board to suddenly stop running, and reboot the game. To fix this, provide an air gap between Q5 and R11. You should also probably replace Q5 too. If you replace R11, make sure you leave some "air" between R11 and the MPU board.

The Game only Starts up "sometimes".
If the MPU only starts up properly "sometimes", replace Q1 and Q2 (both 2N3904), Q5 (2N4403), VR1 (1N959B or 1N4738A) and CR5 (1N4148 or 1N914). These transistors and diodes need to be in perfect working order for the MPU to boot reliably.

LED Still Locked On: MPU Board Signals.
If having play problems with an otherwise working MPU, check all the chip's signals with a logic probe, and cross reference them to this chart. If a signal is other than shown below, check the schematic. If it's an input line to the chip, the device that feeds that pin is probably bad. If it's an output line from the chip, the chip itself is probably bad.

All the following signals are with the MPU board in "attract" mode, and a MPU -35 loaded with Kiss 2732 EPROM's at U2 and U6. Because these signals are with the game in attract mode, this chart won't help much if your MPU is in some other state.
Key:

H = Logic High (+5)
L = Logic Low (GND)
P = Pulse
HP = High Pulse (signal is mostly high but pulsing)
LP = High Pulse (signal is mostly low but pulsing)
X = No signal

Chip	Pin/Signal		Chip	Pin/Signal
U16 (9602)	L 1	16 H	U15 (MC3459)	P 1	14 H
	LP 2	15 L		P 2	13 P
	H 3	14 LP		P 3	12 P
	P 4	13 H		P 4	11 P
	H 5	12 P		P 5	10 P
	P 6	11 H		P 6	9 P
	P 7	10 P		L 7	8 P
	L 8	9 P

Chip Pin/Signal Chip Pin/Signal

U20 (4502) P 1
16 H U17 (7400) P 1
14 H

P 2
15 L P 2
13 P

P 3
14 P P 3
12 P

L 4
13 HP P 4
11 P

LP 5
12 HP P 5
10 P

P 6
11 LP P 6
9 P

LP 7
10 P L 7
8 HP

L 8
9 LP

Chip Pin/Signal Chip Pin/Signal

U18 (4049) H 1
16 X U19 (4011) HP 1
14 H

P 2
15 P HP 2
13 P

P 3
14 P P 3
12 P

P 4
13 X P 4
11 P

P 5
12 P H 5
10 P

P 6
11 P HP 6
9 H

HP 7
10 H L 7
8 P

L 8
9 L

Chip Pin/Signal Chip Pin/Signal

U14 (4049) H 1
16 X U12 (555) L 1
8 H

P 2
15 LP HP 2
7 P

P 3
14 HP HP 3
6 P

P 4
13 X H 4
5 H

HP 5
12 P

HP 6
11 P

P 7
10 HP

L 8
9 P

Chip Pin/Signal Chip Pin/Signal

U2 (ROM) P 1
24 H U6 (ROM) P 1
24 H

P 2
23 P P 2
23 P

P 3
22 P P 3
22 P

P 4
21 P P 4
21 P

P 5
20 P P 5
20 P

P 6
19 P P 6
19 P

P 7
18 L P 7
18 P

P 8
17 P P 8
17 P

P 9
16 P P 9
16 P

P 10
15 P P 10
15 P

P 11
14 P P 11
14 P

L 12
13 P L 12
13 P

Chip Pin/Signal Chip Pin/Signal

U7 (6810) L 1
24 H U8 (5101) P 1
22 H

P 2
23 P P 2
21 P

P 3
22 P P 3
20 P

P 4
21 P P 4
19 P

P 5
20 P X 5
18 P

X 6
19 P P 6
17 H

P 7
18 P P 7
16 P

P 8
17 P L 8
15 P

P 9
16 P P 9
14 P

P 10
15 P P 10
13 P

P 11
14 P P 11
12 P

P 12
13 P

Chip Pin/Signal Chip Pin/Signal

U10 (6820) L 1
40 H U11 (6820) L 1
40 HP

P 2
39 HP HP 2
39 LP

P 3
38 HP H 3
38 HP

P 4
37 HP P 4
37 HP

HP 5
36 P P 5
36 P

P 6
35 P P 6
35 P

P 7
34 H P 7
34 H

P 8
33 P P 8
33 P

P 9
32 P P 9
32 P

HP 10
31 P H 10
31 P

HP 11
30 P H 11
30 P

HP 12
29 P H 12
29 P

HP 13
28 P H 13
28 P

HP 14
27 P H 14
27 P

HP 15
26 P L 15
26 P

HP 16
25 P H 16
25 P

HP 17
24 P H 17
24 P

LP 18
23 P L 18
23 P

LP 19
22 P H 19
22 P

H 20
21 P H 20
21 P

Chip Pin/Signal

U9 (6800) L 1
40 H

H 2
39 L

P 3
38 L

HP 4
37 P

P 5
36 P

H 6
35 L

L 7
34 P

H 8
33 P

P 9
32 P

P 10
31 P

P 11
30 P

P 12
29 P

P 13
28 P

P 14
27 P

P 15
26 P

P 16
25 LP

P 17
24 P

P 18
23 LP

P 19
22 P

P 20
21 L

The LED is Always OFF, and Never Comes On.

TP2: should have +12 volts.
R29: should have +12 volts on both sides of this 1/2 watt 470 ohm resistor.
Q2: should have +11 volts on the leg closest to U11 (PIA). <-- CLAY CHECK THIS.
Q2: Cross the two legs that are next to each other with a screw driver. The LED should come on.

If +12 volts is at the above three points (or at least two of them), the LED is probably bad. If you have the above voltages, cross the two adjacent legs of Q2 together with a screw driver. This should light the LED. If it doesn't, and the above voltages exist, replace the LED.

Ok, the MPU's LED "Flickers": What's Next?.
At this point you have an "unlocked" MPU board. By "unlocked" I mean you get the initial LED flicker upon power-on. This tells us that U6, U9 and U11 are good, and the CPU is "running". Now the MPU will test the other components on the board. So at this point, you should have all the other socketed chips like U7, U8, U10, U2, etc (if you removed them) back in the board. The U6 ROM will continue to run its diagnostics on the other components on the MPU board and will FLASH when each of its tests are COMPLETED. What follows is a description of these flashes (or lack of them), and what they mean.

The MPU Diagnostic LED Flash Sequences Explained.

Introduction: The following LED "flash" sequences are indications to the user that the MPU board's components are being tested, and passes these tests. The number of flashes counted can determine very precisely what MPU board component has failed (1977-1984 Williams pinball owners wish they had it this good!) People familiar with repairing Bally MPU boards are indeed very thankful of these tests, as most often they identify the exact problem on the MPU board.

Remember, the indication of a particular flash means the part being tested passed. That is, if a MPU board boots with just a flicker and then two LED flashes, this mean the reset section and chips/sockets U6/U9/U11 are good (the flicker indicates they passed), ROM chips/sockets U1 to U6 are good (the first flash indicates they passed), and the U7 RAM 6810 is good (the second flash indicates it passed). Since no third flash was seen, this means the item being tested at the third flash (the 5101 RAM at U8 or its socket) did not pass the test, and therefore no third flash was seen. At this point, the boot-up program locks up, and will not continue testing the MPU board until chip U8 (the third flash test item) is fixed.

Of course if the MPU board's LED locks on at boot up with no flashes, this means (as discussed above) that the reset section and/or chips/sockets U6/U9/U11 are bad. At this point, I am assuming the LED is not locked on, and that the reset section and chips/sockets U6/U9/U11 are good (basically that the MPU board at least gives the first LED "flicker").

1st Brief Flicker Completed:
The Fakers Guide: If the LED briefly flickers on power-up, the CPU chip has booted and found the startup code in ROM U6, and is starting to execute startup/test program in ROM U6. So U6 (ROM), U9 (CPU), U11 (PIA), the reset components, and +5 volts DC are good (remember some Stern games also require the U2 and/or U5/U6 ROMs for the first flicker). If the LED locks on, one of these components (or some other supporting components such as the MPU's Q1, Q2, Q5, VR1, or C23, Q20 on the solenoid driver board) are bad (see above). Also the traces connecting these components together could be bad (battery corrosion!)

Techno Guide: On power-up, the U9 CPU chip requires +5 volts DC be applied before the reset line is allowed to swing from 0 to +4.8 volts. It also requires the presense of a two-phase, non-overlapping clock pulse. If these conditions are met, and if the U9 CPU chip itself is good, the LED on the MPU board briefly flickers.

The brief flicker indicates the operation is proper. The MPU has gone out to memory and obtained the starting address of the self-test from memory. The flicker indicates that it then went to that address in the U6 ROM and started to execute the self-test program in the U6 ROM.

The Valid Power Detectors circuit on the U9 CPU works with the +5 volts DC regulator Q20 on the solenoid driver board. This prevents the reset line from going high until +5 volts DC is proper at the U9 CPU chip. Q20 is supposed to go into regulation when +7.5 volts DC is applied to its input. This means that when the game is turned on, and a sufficient time (milliseconds) has passed so that C23 on the solenoid driver board has charged, Q20 switches into regulation. This supplies +5 volts DC to the MPU board.

Q1 on the MPU board (in the valid power detector circuit) does not allow the CPU chip to turn on immediately. The zener diode VR1, in series with the base of Q1 delays application of the reset voltage until C23 charges. At this point, Q1 and Q5 on the MPU board go into conduction, and the reset line at the MPU is caused to go high. Only then is the U9 CPU chip "on".

The importance of the Valid Power Detection circuit can be appreciated when the following fact is known; should the reset line be allowed to go high before the +5 volts is applied and proper, or should the +5 volt supply fail and go out of regulation, the U9 CPU chip can jump out of the program. The reason this happens is that the U9 CPU goes out to the program memory bank U1-U6 for instructions. The logic levels are wrong because the +5 volts is not proper. The MPU misinterprets the data, jumps out of the program, and executes this misinterpreted program! The U9 CPU is now like a train that has left the tracks, and it can end up anywhere. The difference is that a train will eventually stop. But the U9 CPU may continue as long as the clock circuit continues to run.

If the U9 CPU jumps out of the program, it is said to be in "run away". While it is mis-interpreting the program, it invariably overwrites the Bookkeeping function in U8 and the scratch pad RAM. An indication of a "run away" would be false data in bookkeeping. Probable cause is a faulty Q20 or C23 (or both) on the solenoid driver boared, or a leaky zener diode VR1 on the MPU board.

First Flash Completed:
The Fakers Guide: flash number one means ROM U1 to U6 are good. No first flash means one (or more) of the game program ROMs U1 to U6 are bad. Could be a mis-jumpered board, or a bad ROM chip at U1 to U6, or a bad ROM socket at U1 to U6.

Techno Guide: the U9 CPU chip next goes out to the program ROM's (read only memory) U1 to U6. It tests each chip in the bank, in accordance to how the MPU board is jumpered. When it finds the bank is correct, it flashes the LED for the first flash. A fault in the U1 to U6 ROM chips is indicated by the absense of the first flash.

The U9 CPU tests each ROM chip's function like this: in a game with ROM chips U2 and U6 (typical), the CPU first goes to U2. It fetches the first byte in U2, and adds it to the second byte in U2. It will add to this sum the third byte in U2. This continues until all bytes in the chip have been added up. If the sum of all the bytes is "0000 0000", the U9 CPU proceeds to U6 and repeats this process. If U6 has a sum of "0000 0000", the U9 CPU causes the LED to flash the first time. Fault in either U2 or U6 is indicated by the absence of the first flash.

The contents of each ROM chip have byte locations called checksums, reserved for this test routine. There is one checksum byte reserved in each 512 bytes of ROM memory. The game programmer at Bally must insert a btye with the proper value in each checksum byte location to force each 512 byte checksum to equal "0000 0000".

During the life of an electronic game, if a ROM chip U1 to U6 fails by so much as a single bit, it will be detected during this CPU test. The CPU will not continue until the defective ROM chip is replaced.

Second Flash Completed:
The Fakers Guide: a second flash means the 6810 RAM at U7 is good. No second flash means U7 (6810) or its socket are bad.

Techno Guide: The U9 CPU chip goes out to the U7 RAM and erases the contents of the first byte (U7 is a 128 byte scratch pad memory). It then tries to read back the word "0000 0000" (indicating erased). If it can read it back, it adds "1" and continues. 256 tries later, it writes the word "1111 1111". If it can read it back, it has determined that the first byte in U7 is good. It repeats this process for each of the 128 bytes of RAM in U7, one at a time. If at the end of this 256 x 128 (=32,768) tests, each time the CPU writes, it can read the same word back, the CPU cause the LED to flash a second time.

Note the pause between the first and second flashes. This is the CPU doing 32,768 tests to the RAM at U7 and repeats the process.

Third Flash Completed:
The Fakers Guide: a third flash means the U8 5101 RAM is good. No third flash means U8 (5101) or its socket is bad.

Techno Guide: The U9 CPU goes out to U8 (CMOS 5101 RAM) and makes a copy of the contents of the first half byte. It does this because U8 is battery supplied, non-volatile memory where the bookkeeping functions are stored. It then erases the contents of the first half byte, and tries to read back the word "0000 xxxx". If it can read it back, it adds "1" to the previous word (giving "0001 xxxx"). It continues to write and read until it reaches the word "1111 xxxx". When this is done successfully, the CPU restores the original contents to the first byte located in U8. It then makes a copy of the contents of the second byte, and repeats the process. It does this for the entire 256 bytes, one at a time. If at the end of the 256 x 16 (=4096) test, each time the CPU writes and reads the same word correctly, the CPU caused the LED to flash a third time.

Since the 5101 U8 RAM is in the "corrosion zone", here is a quick check of the U8's connection to other components. This will help determine if the U8 socket is bad. (Note the letter/number in parends is the address, or data, or chip enable number.)

U8 pin 1 (A3) - U7 pin 20, U6 pin 5
U8 pin 2 (A2) - U7 pin 21, U6 pin 6
U8 pin 3 (A1) - U7 pin 22, U6 pin 7, U11 pin 35
U8 pin 4 (A0) - U7 pin 23, U6 pin 8, U11 pin 36
U8 pin 5 (A5) - U7 pin 18, U6 pin 3
U8 pin 6 (A6) - U7 pin 17, U6 pin 2
U8 pin 7 (A7) - U7 pin 15, U6 pin 1, U11 pin 24
U8 pin 8 - Ground
U8 pin 9 (D10)*- U7 pin 6, U6 pin 14, U11 pin 29
U8 pin 10 (D00)*- U7 pin 6, U6 pin 14, U11 pin 29
U8 pin 11 (D11)@- U7 pin 7, U6 pin 15, U11 pin 28
(* U8 pins 9&10 shorted together)
(@ U8 pins 11&12 shorted together)

U8 pin 12 (D01)@- U7 pin 7, U6 pin 15, U11 pin 28
U8 pin 13 (D12)*- U7 pin 8, U6 pin 16, U11 pin 27
U8 pin 14 (D02)*- U7 pin 8, U6 pin 16, U11 pin 27
U8 pin 15 (D13)#- U7 pin 9, U6 pin 17, U11 pin 26
U8 pin 16 (D03)#- U7 pin 9, U6 pin 17, U11 pin 26
U8 pin 17 (CE2) - Q5 rt upper leg, U9 pin 40, U11 pin 34
U8 pin 18 (OD) - U18 pin 6
U8 pin 19 (CE1) - U17 pin 8
U8 pin 20 (R/W) - U7 pin 16, U11 pin 21, U9 pin 34, U18 pin 7
U8 pin 21 (A4) - U7 pin 19, U6 pin 4, U11 pin 22
U8 pin 22 (Vcc) - C13 left leg, R12 upper leg, CR5 lower leg
(@ U8 pins 11&12 shorted together)
(* U8 pins 13&14 shorted together)
(# U8 pins 15&16 shorted together)

If all the U8 RAM lines ring out as described above, and the 5101 U8 RAM is known "good", yet only two LED flashes are received, next replace U19 (4011). This often fixes a stubborn two-LED-flash MPU board.

Also be aware of the speed of any replacement 5101 RAM. A too-slow 5101 will may pass the initial boot-up LED code test, but can cause wack score display behavior. Turns out a bad 5101 RAM or one that is too slow will not keep up with the CPU's demand for accessing the display data. Here's a list of 5101 RAM speeds:

PCD5101P : 150nS (Philips)
5101-1 : 450nS
5101L-1 : 450nS
5101-2 : 450nS
5101L-2 : 450nS
5101 : 650nS
5101L : 650nS
5101-3 : 650nS
5101L-3 : 650nS
5101-8 : 800nS

The Phillips brand of 5101 is the best, as it's the fastest and will work in all Bally/Stern MPU board applications. Sterns MPU-200 need 450ns or faster RAM, Bally MPU -35/-17 need 650ns or faster. So you can see that some 5101 RAM just won't work properely in Bally or Stern MPU boards.

Fourth Flash Completed:
The Fakers Guide: a fourth flash means the U10 PIA (6821) is good. No fourth flash means U10 (6821 PIA) or its socket are bad. Or battery corrosion had broken a trace going to U10. Or the chip leading to U10 is bad, which is U20 (4502).

Techno Guide: The U9 CPU chip now tests the first 6821 PIA chip. There are two of these chips on the MPU board, which are identical and interchangable. The test for both is the same.

To determine if a PIA chip is good, the U9 CPU does the following:

The CPU accesses, by means of input RS0, RS1, CS0, CS1 and CS2 each of the two full byte registers used to store the port initialization information. If does this, one register at a time. After it completes the first register, it repeats for the second. It goes through 256 tests similar to that used to check each byte in U7 (second flash). If each time the CPU writes a word into the register, it can read the same word back, it continues to test until completion.
The CPU accesses, by means of input RS0, RS1, CS0, CS1 and CS2, each of two full byte registers used as data output registers when PA0 to PA7 and PB0 to PB7 are used as outputs. It does the same type of test on each register as described just above. Again if no faults are found, the test is continued until completion.
The CPU then accesses, by means of input RS0, RS1, CS0, CS1 and CS2, the two ports CA2 and CB2. The port is initialized as an output. The port is then written into to see if it can store a "1" and then a "0".

A total of 4 x 256 + 4 (=1028) test steps are required to test the PIA chip. However, there are internal buffer amplifiers used with the PB0 to PB7 output registers and CB2 port register which can not be tested by the CPU. Access is only to the register; if the buffer is open, it does not interfere with the registers ability to be written into and read from by the CPU. It is this uncertainity that reduces the accuracy of these test to 99.5%.

Also check and make sure that the MPU connector J1 and the lamp driver connector J1 are not mixed up! These two connectors are about 4 inches apart, and are keyed the same. If only three MPU LED flashes are seen, this problem can also sometimes show the score displays with an '8' or a '9' in the hundreds place.

Also I recently fixed a game (Bobby Orr Power Play) where sometimes it would not get the fourth MPU flash, but sometimes it would get all seven. If it did boot, the score displays would flicker and only show some of the six digits. The problem was a trace going to U10 on the component side of the MPU right near the battery that was intermittent from corrosion. After I cleaned up the corrosion (sanded and neutralized) and repaired the broken trace (it was obviously open after sanding), the MPU booted and the game worked correctly.

Fifth Flash Completed:
The Fakers Guide: a fifth flash means the U11 PIA (6821) is good. No fifth flash means U11 (6821 PIA) or its socket are bad. Note the flipper enable relay will usually click on/off at about the fifth flash. The lack of a fifth flash could also be a bad resistor connecting to U11 like R134. Another issue could be a bad score display, which can cause the MPU board to stop during the 5th LED diagnostics. Disconnecting all the score displays (power off), then reconnecting them one at a time (power off) and rebooting can help identify a bad display that is dragging down the fifth MPU LED flash.

Techno Guide: Same test is performed on U11 as was performed on U10. See above.

Sixth Flash Completed*:
The Fakers Guide: a sixth flash means the PIA at U11 (6821) and the U12 timer (5550 chip is good. No sixth flash means either PIA U11 (6821) or its socket is bad, or the U12 (555) timer is bad. Lack of the 6th flash (five flashes only) can also mean a problems with the zero-crossing detector. Often this is the MPU board's R113 resistor (2k ohm, near the J4 connector) is open. This resistor takes the 43 volt DC coil power to the zero-crossing detector circuitry and normally runs hot. Hence it may eventually go open, so check it if only 5 flashes from the LED are seen. Thanks to Ray J. for this tip.

Techno Guide: The U9 CPU chip monitors PIA2, port CA1 (U11). If transitions from high to low are detected, the CPU decides the Display Interrupt Generator is working. If U12, a 555 timer, or any associated circuit component fails, the CPU will not flash the LED the sixth time.

* Note on Baby Pacman and Granny and the Gators, this flash step is skipped and not tested.

Seventh (Last) Flash Completed**:
The Fakers Guide: Getting the seventh flash means the CPU board has passed all self-tests. After the 7th flash, the LED should have a very dim glow to it (this "dimness" occurs because the game is strobing the switches), and the game should be in 'attract' mode. No last LED flash generally means there's no +43 volts DC for the solenoids (power transformer fuse F4 is probably blown), PIA U10 (6821) or its socket are bad, or U14 is bad.

Remember the 7th flash is looking for 43 volts (or 6 volts AC in the case of Baby Pac's last flash) on the MPU board. Note this test for voltage is performed on Baby Pac too (the prior flash step is skipped, giving Baby Pac a total of 6 LED flashes). This is done thru PIA U10's CB1 port. The PIA is looking for the zero-cross of the 43 volts (or 6 volts for Baby Pac). If 43 volts is not present (rectifier board F4 fuse blown), the last flash will not occur.

If the game locks_up after the 7th flash, try unplugging the sound card and reboot. If it then works check the isolation power diode CR3 on the sound card - sometimes this bad diode on the sound board can lower the 43 volts enough to stop the MPU board's 7th flash (Star Trek comes to mind). Also Check the solenoid driver board cap C23 is in good condition. There is also a chance the 555 timer chip and/or C17 could be bad on the MPU board.

The 43 (or 6) volts can be "faked" and not supplied (for example if you are booting the MPU board on the work bench), and the last LED flash can happen. Just connect the top leg of resistor R23 (leg closest to chip U12) to the top leg of resistor R17 (leg closest to TP3) using an alligator test lead.

Techno Guide: The U9 CPU chip monitors PIA1 port CB1 (U10). If transistion from high to low are detected, the CPU decides that the zero crossing detector is working. If U14 fails and the CB1 line is stuck high or low, the test will also fail. The zero crossing detector circuit input is the +43 volts DC line that is used for the solenoids. If the fuse in that line (F4 on the power transformer module) is blown when the game is turned on, the CPU will not flash the LED the seventh time.

** Note on Baby Pacman and Granny and the Gators there is no 7th LED flash.** These games only had six flashes instead of seven. The easiest way to determine if the MPU board is a -133 designed for Baby Pac/Gators and 6 flashes is to look at the J4 connector (lower left corner) around pin 19. Between L1 and L2 (the very large inductors that look like 2 watt resistors) is either a diode (cr52) or a resistor (r113). If it's a diode, this is a -133 MPU board and there will only be 6 flashes. If it's a resistor, it is a -35 or -17 MPU board and there should be 7 flashes. Converting the -133 MPU to a -35 MPU is very easy do. Just replace the 1N4148 diode at CR52 with a 2k ohm 1/4 watt resistor. New EPROMs will need to be burned too of course For more information on the -133 MPU board see here.

Game Initialization.
The U9 CPU chip now initializes the two PIA's U10 and U11, assigning to each port its role as either an input or an output, as required. It then clears out U7 (6810 RAM). Now the CPU takes a picture of the settings of fixed switches S1 to S32 on the MPU board. It stores this "picture" in memory in chip U7. The CPU next jumps to a routine which turns on the "Game Over" feature light, lights the "Ball in Play" light, and the "Credit Indicator" light if there are credits stored in memory. It resets the drop targets and activates the saucer kickers or any kicker associated with a playfield device that can trap the ball and keep it out of the outhole. It then energizes the coin lockout solenoid to allow the game to accept coins (unless the credit maximum was met). Playfield and backbox feature lights associated with and appropriate to animation effects are turned on. With the game tested and initialized, the CPU now divides its time between monitoring momentary switches for closure (coin switch, credit button) and updating displays (lamps and score registers).

Problems/Solutions with the Seventh Flash.
Of course first verify that the MPU board is *not* a -133 model with Baby Pacman or Granny and the Gator EPROMs (hint: if U2 and U6 are a 2732 and 2532, suspect a -133 MPU board). A few paragraphs above describes how to tell if it is a -133 MPU board.

If fuse F4 is blown on the solenoid board, the seventh flash will not occur. But what if there is a problem on the playfield which is forcing this fuse to blow (stopping the final LED flash on the MPU board, hence stopping your MPU diagnostics)?

The easiest way to deal with this is to remove the solenoid connectors from the solenoid driver board to the playfield. These connectors are on the left side of the solenoid driver board, and the one connector at the bottom right of the board too. This should allow the F4 fuse to be replaced, and the completion of the MPU booting process.

To help find the playfield coil that is causing the F4 fuse to blow, replace the connectors one at a time on the solenoid driver board (with the game off), and reboot the game. This will help issolate the bad coil.

Another thing to try: remove the under-the-playfield coil fuse, and replace fuse F4. If F4 does not blow, then one of the coils under the playfield is somehow shorted or staying energized (and blowing the solenoid driver F4 fuse). If fuse F4 still blows, there is either a problem with the backbox knocker, or the cabinet coin door lockout coil, the solenoid bridge rectifier (on the rectifier board), or the rectifier board's varister.

Also try removing connectors J1 and J3 from the rectifier board (this moves the solenoid power back a step further, not allowing it to get any further than the rectifier board). Replace fuse F4 on the solenoid driver board, and turn the game on. If the fuse still blows, the solenoid bridge rectifier (on the rectifier board) or the rectifier board's varister is probably at fault.

If it is a coil under the playfield, check the coils to see which one energizes when the game is powered on. Or disconnect a wire on each solenoid, and re-attach each wire, one at a time, until the fuse blows. At this point it could be the coil, coil diode, or coil driver transistor at fault.

Still No Seventh Flash - Other Things to Check.
First check TP3 and make sure there is 21 volts DC (of course this assumes fuse F4 is not blown on the solenoid board). If there is still no seventh flash, here are some other things to check. Remember all the components mentioned below are in the battery corrosion area.

Check resistor R17 (150k). If this resistor goes open, or is not making good contact the circuit board, there will be no 7th flash. This resistor commonly fails.
Check resistor R16 (2k).
Check resistor R18 (1.5 meg).
Check diodes CR52 and CR49 (1N4148 or 1N914).
Check the IRQ line (U9 pin 4 to U10,U11 pins 37,38) and resistor R134 (4.7k) which ties the IRQ to +5 volts.

Faking the 7th Flash.
If you are diagnosing a Bally MPU on the bench and don't have a 43 volt power supply, there is a way to fake the MPU board into doing the 7th flash. This may also be useful if all other things have been tried and there's still no seventh flash.

To fake out the MPU board, use an alligator jumper wire and connect the top of resistor R23 (leg closest to U12) and the top of resistor R17 (leg closest to TP3). This makes the MPU think the 43 volts is present, and will allow the board to fully boot (7 flashes). This happens because the 43 volts is reduced to 5 volts and then converted into an "impulses train". This signal comes from pin U10 pin 18 and then the MPU monitors this impulse train signal. If the signal is present the MPU thinks the 43 volts must be Ok. By making the connection between the tops of R17 and R23 we simulate the presence of the 43 volt impulses train.

Problems After the 7th flash (game sitll doesn't work).
In a normal boot-up sequence, after the seventh LED flash, the LED glows very dim (some LEDs show this more than others). Occassionally, a game will do the seven flashes, but won't go into "attract" (game over) mode.

Stern MPU-200 boards in particular can have problems where the seventh flash is seen, but the game does not work. Of course the first thing to check (especially on multiball games) is if all the pinballs are installed and the ball trough switches are working. If there is a single dip switch on anywhere, the MPU boots (seven flashes), then some score digits on all displays come on for about five seconds, then all displays go out (display blanking). Slam switch (and its capacitor) are not the problem!

So what could be the problem? The answer is usually a bad 5101 RAM chip. Especially on Stern MPU boards, the 5101 RAM chip(s) need to be the proper speed (100ns, unlike Bally which can use a 300ns 5101 RAM chip), or the game will not work (even though it passed the 5101 self-test at power-on).

MPU Boots Fine, but after Turning Off and Immediately back On, The MPU board is Locked.
Game is turned on and works fine. Then the game is turned off, and within a few minutes, turned back on. But the MPU board's LED is locked on, and will not (flicker) boot. The reset section of the MPU board has been rebuilt, as described in this document.

The first thing to suspect is MPU battery corrosion in the reset circuit. Inspect the board for any damage due to corrosion. The problem may be the reset line stays high from the battery power. When the battery is discharged enough, the game will restart. This can happen from the two 8.2k resistors (R1, R3) at the bottom of the MPU board. Other resistors in the reset section should be checked too (R2, R112, R120, R140, R139, R138, R140, R12, R11).

Seven Flashes Constantly Repeats.
The MPU boots and gets to the 7th flash, then start all over and repeats the seven flashes, then starts all over and does it over and over (like the game is constantly trying to reboot itself over and over). This can happen from a bad 6810 RAM at U7, or possibly a bad socket or broken trace at U7. Also there can be a broken trace or bad socket at the U11 PIA especially around pins 35 to 40. Since U7 and U11 are in the "battery corrosion zone", look for broken/corroded traces or bad chip sockets in this area.

But besides that, if all -32 MPU board DIP switch are off, the game goes in self-test mode (toggling alternatively every solenoid, flashing controlled lamps and testing single digit at a time on every display) after the seventh flash.

3c. When things don't work: Test EPROM & Fixing a Dead MPU board.

+5 and +12 volts.
Valid Clock signal chips U15 and U16 (usually not socketed).
Test EPROM in U6 (either 2716 or 2732).
6800 CPU chip in U9.

That's about all that is needed (all other socketed chips can be removed to minimize problems, and can be added one at a time after the board is running the Test EPROM to confirm their validity, or all the chips can be left installed if desired). The reset section of the board doesn't even need to work - instead we can do a "manual reset" by jumping TP5 (+5 volts) to CPU U9 pin 40 to get a the processor running.

The Test EPROM only needs the 6800 CPU at U9 and the reset/clock circuits. It is not dependent on memory chips U7 (6810) and U8 (5101), and the game-specific EPROMs (U1, U2, U5, U6). And the PIA chips at U10 and U11 don't need to be installed (even though the Test EPROM will try and test these two PIAs).

The Test EPROM has a short program which runs from EPROM U6 and controls the two PIA chips at U10 and U11. Basically it sets the U10/U11 PIA chip outputs PA0-PA7 and PB0-PB7 and CA2/CB2 high and low, over and over at one second intervals. While it does this it flashes the built-in MPU LED at one second intervals and a Test LED which we will connected to CPU U9 pin 15. This gives us the opportunity to test the PIA outputs using a Test LED or logic probe or even a DMM. The Test EPROM does not care if the chips are working or not, and it does not care about the other memory chips on the board either. It will try and run this test over and over regardless, unlike the built-in Bally/Stern test which will stops as soon as it finds a problem. This allows us to get the MPU board working with a mimimum of chips (and a minimum of problems).

Remember it is always easier to tackle a small set of problems, instead of trying to find all problems at once. If a MPU board can run Leon's Test EPROM, you are well on your way to fixing the MPU board. It is always the initial step of making a MPU board boot on its own that is most difficult. Problems past this tend to be easier to tackle.

Download the Test EPROM files here and burn them into a 2716 EPROM:
2716 format Test EPROM.

Using the Test EPROM.
If using the Test EPROM "on the bench" outside of the pinball machine, connect +5 and +12 volt (43 volt is not needed) as shown in the Making a MPU Test Fixture section of this document. This involves using a computer power supply and connecting +5 volts to TP5, +12 volts to TP2, and ground to TP4. I highly recommend this approach. If using the Test EPROM in the game, remove connector J4 on the Solenoid Driver board to kill power to the solenoids.

With the power off, remove MPU board chips U1, U2, U5, and U6 (the game ROMs), also U7 (6810) and U8 (5101). Note all these chips do not have to be removed, but it's a good idea to minimum potential problems by removing them. Also remove the two 6821 PIA chips at U10 and U11 (again not necessary, but a good idea).

Because the Test EPROM is a 2716 EPROM, the MPU board need to be jumpered for 2716 or 2732 EPROMs. Or if you have a MPU board with its original black 9316 ROM chips, the easiest thing to do is to make an adapter as shown in the 9316/2716 Adaptor section of this document (note only one of the two sockets is needed, and the test clip does *not* need to be connected - see the description below). Instead of making MPU board jumper changes (which could be mistakenly done wrong, making the MPU board more difficult to fix), using a 9316/2716 adaptor socket is a much better idea. If the MPU board is jumpered for 2732 EPROMs that is fine - this 2716 Test EPROM will plug right in without any jumper changes and work.

The 9316/2716 adaptor for this application takes two good quality (machine pin) 24 pin sockets, and sandwiches them together. There are a couple pins that need to be cut and jumper wires added:

On one 24 pin sockets, jump pins 21 and 24 together with some wire.
On the above modified socket, cut pins 18 and 21 short so they won't plug into anything.
Plug the modified sockets into an unmodified socket. Make sure pins 18 and 21 do not contact the unmodified sockets. To be double sure this happens, you can cut pins 18 and 21 off the bottom socket too.
Plug the 2716 Test EPROM for U6 into the modified socket sandwich.
Plug the socket sandwich and EPROM into the MPU board at positions U6.

The Test LED.
In addition to the built-in LED on the MPU board, we also need to make a "Test LED". Even though the on-board MPU board LED is present, without PIA U11 the on-board LED will not work. So to test a MPU board with a minimum of chips, a "Test LED" is needed to indicate if the Test EPROM is running (without the Test LED, there is no easy way to tell).

The Test LED will be connected to address line 6 of the 6800 CPU chip (U9 pin 15 or MPU connector J5 pin 16). Take a LED and a 150 ohm resistor connected to the "flat side" of the LED leg. The resistor side of the LED is connected to ground (TP4), and the non-resistor side of the LED is connected to U9 pin 15 or J5 pin 16. Use micro-clips to attach the Test LED.

Using the Test EPROM.
To get started, install the Test EPROM in the MPU board at socket U6. If the board originally used black 9316 ROMs, an adaptor will be needed for the 2716 EPROM (do *not* change the jumpers on the board!) Remove all socketed components from the board except for the U9 6800 CPU, the U6 Test EPROM, and U15 and U16 (which are normally not socketed anyway). Do not have U10 and U11 (the 6821 PIAs) installed at this time. Also do not have the U7 (6810) and U8 (5101) RAMs installed.

Power up the MPU board (ideally on the work bench using a computer power supply for +5 volts to TP5, +12 volts to TP2, and ground to TP4. The Test LED will immediately start flashing about once a second. If the Test LED flashes in rhythm then the reset section of the MPU board and the Clock circuit are working (a very good sign). Note the on-board MPU mounted LED will be locked on if PIA U11 is missing or bad.

If the Test LED is not flashing: Probably the reset section of the MPU board is dead or the MPU board is not jumpered correctly for the U6 EPROM installed. To check the U9 reset, with the board power on, use a DMM and check for +5 volts at U9 pin 40 (this is the reset pin of the 6800 CPU chip). If less than 2 volts, the reset is being held low, and the reset circuit of the MPU board is not working (the reset circuit should hold U9 pin 40 low for about 50 miliseconds, and then bring U9 pin 40 high to 5 volts and keep it high - this will start the CPU chip). If U9 pin 40 shows 5 volts, then the reset section is working.

If U9 pin 40 is low and you want to simulate the reset circuit with a "manual reset", power the board on. Then use a jumper wire and jump TP5 (+5 volts) to CPU U9 pin 40. This will make the CPU chip's reset line "high", and it should start the CPU running (assuming the CPU chip itself is good, as are all the address and data lines to the U6 ROM). If a "manual reset" does not wake up the CPU, try another 6800 chip at U9. Still no LED flashing? Checking the signals of the U9 CPU chip:

Check for a clock signal at U9 pin 3. A DMM will show about 2.4 volts, or use a logic probe or an oscope for a clock signal. No clock, then check U15 and U16.
Check for the second clock signal at U9 pin 37. This should also show about 2.4 volts.
Check U9 pin 2 for +5 volts.
Check U9 pin 4 for +5 volts.
Check U9 pin 5 for around 2.8 volts.
Check U9 pin 34 for a positive signal with very short negatives spikes (use a logic probe for this).
Check the address line signals U9 pin 9 to pin 24 using a logic probe, except A8,A9,A10 and A13 (U9 pins 17,18,19,23), as these four address lines will be 0 volts.

If any of these signals are bad (and the U9 6800 was replaced with a known good one), trace the bad signal to its source and repair it using the schematics. A last possibility is that the ROM selection of U6 doesn't work. Check the following pins of U6:

U6 pin 18 without an adapter 2.5 volts (with an adapter 0 volts).
U6 pin 21 is 3.6 volts (if no 9316/2716 adaptor) or +5 volts (with a 9316/2716 adaptor).
U6 pin 20 is ground with very short positives spikes, only seen with a logic probe or scope.

If all these signals are good, there should be a flashing Test LED (again, with PIA chip U11 removed, the on-board MPU LED will be locked on). Once the LED is flashing, replace the PIA chips U10 and U11. With U11 installed, the on-board MPU board mounted LED should flash in rhythm (but opposite) with the Test LED.

Testing the PIAs Using the Test LED.
Using the Test LED, all outputs of PIA chips at U10 and U11 can be tested. The Test LED will need to be connected as shown above, which is slightly different than the U9 Test LED connection. The non-resistor leg of the Test LED goes to +5 volts (TP5), and the resistor leg of the LED is used to probe the pins of the U10/U11 PIA chips. Note with PIA U11 installed, the MPU board mounted LED will not blink too (but this LED really does not help with testing). Using the Test LED, probe the two PIAs U10 and U11:

PIA Pins 2-17 high then low (Test LED on and off), alternating every second.
PIA Pins 19,39 high then low (Test LED on and off), alternating every second.
PIA Pins 26 to 33 are the data lines, and should be pulsing (use a logic probe for these pins).

If one PIA output does not flash, temporarily connect it (short it) with the output pin next to it. This will cause one of two things - Both outputs "dance": this means the first output is dead and the 6821 must be replaced. If both outputs do not "dance": there is an short somewhere on the suspect output line. If this is the case, disconnect the output line by bending the PIA pin up and out of the socket, and connect the Tester LED to the extracted pin. If the LED dances, the 6821 PIA is OK and the short is somewhere on the MPU board. If the extracted PIA pin does not "dance", then the 6821 is dead and must be replaced.

Finishing Up.
If all the PIA ouput signals are there, replace all the other chips (except the original ROM chip U6). The replaced chips should have no influence on the test (repeat the testing of the U10 and U11 PIAs using the Tester LED). If something is different there is something wrong with the selection of the newly installed chips since it disturbs the test EPROM program.

Memory Test using the Test EPROM.
To execute this additional memory test, have all the MPU board chips installed and the Test EPROM installed at U6. Also have the Test LED connected to U9 pin 15 or J5 pin 16, as done before the PIA tests. Power the MPU board up. Now press the red button which is on top of the MPU board. The memory test will then start and give you one flashes on the Test LED, and then a continual flashing of the Test LED as the Test EPROM goes back to testing the PIA chips.

Note there are some restrictions to the Memory test. Although Bally and Stern cpu boards can be interchanged, there is a major difference between the -17/-35 Bally MPU board and the Stern M-200 MPU board. The Stern M-200 has an extra 5101 memory chip extra at U13. To make both types of boards use the same memory test, Leon used some tricks in programming the test EPROM. If this wasn't done, and you're testing a Stern board without U13 (or with a bad U13 chip), it would be reported as OK, because a regular Bally board doesn't contain a U13 chip.

When the Test EPROM is started, the Test LED will start to light in rhythm, and this is the regular PIA test. Now push the MPU board mounted red button. The Test LED will turn off and after one second it should flash a first time. This means the test of U7 6810 has succeeded. Again a second later the Test LED should start flashing continually (indicating a good "second flash"). The test has run on U8 (5101) and U13 (5101) and has succeeded. But if the LED flashes this second time 5 seconds after the first flash, this means the test has succeeded but *only* for U8.

So on a Bally board both flashes have to be about 5 seconds apart, meaning both U7 (6810) and U8 (5101) are good. On a Stern M200 board the second flash (continual flashing) has to be just one second after the first flash, meaning U7 (6810) and U8 (5101) and U13 (5101) are all good. But if the second flash comes 5 seconds later, then the M200's U8 is fine but U13 is bad (you can test this by removing U13, or swapping U8 and U13). So when testing, if U7 or U8 is defective there'll be no flash, or only one flash. If you have a Stern M200 board and two flashes (continual flashes) are seen with 5 seconds between them, then U13 is not installed or U13 is bad. This way one test EPROM can be used for either a Stern M-200 or a Bally -17/-35 MPU board.

Another restriction of this memory test (not the PIA test) is that it cannot use a 9316/2716 adaptor socket. When using the adaptor socket it's not possible to use U7 6810 (check the repair manual for Bally Stern MPU boards). Because if only two flashes are seen with the game ROMs installed, there's is a problem with memory chips. But unfortunately the Bally test does not say which memory chip (U7 or U8). So if you don't want to adapt your MPU board to use EPROMs, you cannot use the Leon Test EPROM for the memory test (but the PIA test works fine).

All Done.
Now that all checks out, install the original U6 ROM, and count how many times the MPU board's LED flashes (should be 6 or 7 times). If less than this, see the LED flash count section of this document.

3d. When things don't work: Game ROMs, EPROMs, and Jumpers - the Basics

Bally has provided us with about 35 different "jumper" locations on the -35 (and -133) MPU board, and about ten on the -17 MPU. This was done to handle all the different game ROM (read only memory) chip configuations, for all the different games. Depending on supply and demand for certain ROM chips, Bally pretty much had all the bases covered. They could use any combination of ROM types and sizes to fit their needs.

Important: Before you Change any Jumpers!!
It is EXTREMELY important that you have a working MPU board before you change any jumper locations! If you MPU board currently has ROMs in it, get it working first before you play with the jumpers. If you have your MPU board jumpered incorrectly for the game ROMs installed, the diagnostic LED light will stay on and the board will not power-up. So it is absolutely important that the jumpers are correct for the ROMs installed. Get your MPU board working first before proceeding.

Which MPU board Is It: -17 or -35 or -133? (How to Tell!)
Of course Bally silkscreen the MPU part number (AS-2518-17 or AS-2518-35 or AS-2518-133) right on the MPU board. But sometimes due to battery corrosion, etc, the part number can not be read. It is important to know which board is being worked on, as the game ROM/EPROM jumpers are different.

Basically there are two flavors of MPU board: the -17 or -35 (the -133 is really a -35 board with R113 changed to a diode CR52). The easiest way to tell which board is to examine connector J5. On a -17 MPU board, this connector will have 32 pins (including the removed "key" pin). On a -35/-133 MPU board, J5 will have 33 pins (including the removed "key" pin).

Bally/Williams' Jumper Information.
Though I personally don't recommend or use it, here is a link to the information that (recently) Williams posted about the original Bally board jumpers. You may find this information helpful (though personally I don't). It lives at ballyrm1.htm.

Masked ROMs versus EPROMs.
If you are repairing a game, you may be replacing the original masked (black) ROMs. These ROMs, as they get old, break very easily. In particular their pins tarnish and/or break off. The black tarnish is silver oxide, which is caused by oxygen and moisture in the air, and dis-similar metal electrolytic action between the silver plated pins and the tin plated sockets. The black oxide doesn't conduct well, and this can make the ROMs stop working correctly. They are not available any more, but they can be replaced with EPROMs.

EPROMs are basically the same as masked ROMs, but have a small cyrstal window on top of the chip. If UV (ultra-violet) light is shined through this window, the chip can be erased and re-programmed. An EPROM programmer is necessary to re-program an EPROM. These attach to your computer (though there are some stand alone ones too), and can "burn" a blank EPROM with a game's program code. After the EPROMs are burned, a sticker is placed over their window to prevent light from erasing them. EPROMs are the best replacement for your old and tired black masked ROMs.

Chances are if you are doing any repairs to a Bally MPU board, and you need new game ROMs, you will be installing EPROMs. EPROMs are available blank from many sources. But you need to have them "burned" first before you can use them.

The two chips on the left are the black "masked" ROMs.
All masked ROMs used in these Bally games are known as "9316"
EPROMs can be 2516, 2716, 2732, or 2532. The last two numbers
identifies the size of the ROM (16k or 32k bits of program data).
Masked ROMs are only usable in the game they were designed
for. Note how their legs are tarnished black. The two chips
on the right are EPROMs. Note the "window" which allows
them to be erased, and re-programmed. Also note how the
EPROM's legs are not tarnished. You can just read the
notation on the one EPROM as "2532".

The Program Code that Lives inside the ROMs.

The Jumpers.
Each kind and size of ROM chip has its own set of enable and address requirements. To allow a board to use any type or size ROM that Bally had in stock, jumpers were used on the MPU board. It also provides a way to take a MPU board out of a newer game, and put it in an older game (and only have to change the jumpers and the game ROMs).

The wide array of jumpers Bally provides is very confusing. Some games use just one ROM at U6. Others use three ROMs at U6, U2, and U1. The size and type of these individual chips can change too.

What Type and Size are My ROM chips?
Before you can change a MPU board's jumpers to another game (or for new EPROMs), you need to know the type and size of chips you are installing. If you are installed Masked ROMs, these are all 9316 (16k bits of data) or 9332 (32k bits of data). If you are installing EPROMs, you need to read the markings on the EPROM. They will say what type and size of EPROM you have. They will be 2516, 2716 (both 2516 and 2716 are the same type of 16k bit EPROM), 2532, or 2732 (32k bit EPROMs). Note the jumpers for 2532 and 2732 EPROMs are different, as these chips have different pin assignment (but the jumpers for 2532 EPROMs and 9332 ROMs are the same, which is the ONLY sharing of masked ROM and EPROM jumpers). This is unlike 2516 EPROMs and 2716 EPROMs, which are identical and have identical jumpers.

Texas Instruments TMS2716 Warning.
Do *not* attempt to use TMS2716 EPROMs in any Bally/Stern MPU or sound board. The TMS2716 EPROM requires -5, +5 and 12 volts to operate and are not pin compatible with "normal" 2716 EPROMs. Also most EPROM programmers can not program a TMS2716. Note that TMS2516 *are* compatible with standard 2716 EPROMs and work fine.

What if I don't have an EPROM Programmer?
You can also buy the above EPROM chips already programmed in 2732 format (or any other size) from a variety of sources for a nominal fee.

What Do the Masked ROMs Chip Numbers Mean?
If the game in question uses original black masked ROMs, there are numbers on the ROMs that denote information about the chip. This information is can be seen by clicking ballyro2.htm, where the ROM numbers are sort numerically. Another page located at ballyrom.htm shows the same information, but sorted by game name. Also shown are the original jumper settings. In addition, the chart at roms.htm shows the Bally ROM part numbers and checksums.

Note the higher -xx number means a later revision. In the case of the U6 chips, that revision is important, as generally a game can not be run with the "wrong" U6 part. But for the U1/U2 chips, it usually means a later version of the same game (and all the chips must be of the "correct" revision to match the others, though there's not always a good way to know what that is).

As an example, Lost World uses "720-28" for U6. But then the other chips should all be "729-xx", with the latest revision being "729-33" at U1, and "729-48" at U2. If the game has an earlier version, they would have some other numbers after "729-". But note there can be no mixing of old and new versions!

Do I have the Right Jumpers/ROMs Installed?
As stated above, if the jumpers are set incorrectly for the ROMs installed in the MPU board, the MPU LED will be locked-on. This happens because the ROM code is what turns the LED off - if the MPU can't access and run the ROM code, then the LED will never turn off and start its flash sequence.

Game ROM Software.
If you have an EPROM programmer, the game ROM software is needed for programming into EPROM(s). The best solution is to upgrade your current -17 or -35 MPU board to use 2732 EPROMs. But since the first eight Bally games from 1977 to 1978 use a -17 MPU board (which is fairly easy to convert to two 2716 EPROMs), a ZIP file just for those games is 2716 format is provided (bly2716.zip). Note starting with the first -35 game (Lost World) Bally used three 9316 ROMs, which are better utilizied in two 2732 EPROM format.

All 1977-1985 Bally software in 2732 format: Bly2732.zip.
All 1977-1978 Bally software for -17 MPU board in 2716 format: Bly2716.zip.
All Stern EPROM software: sternROM.zip.
Bally Pacman MPU and Vidiot board software: babypac.zip.

* Go to the Bally Repair Guide Part 1
* Go to the Bally Repair Guide Part 3
* Go to the Pin Fix-It Index
* Go to Marvin's Marvelous Mechanical Museum at http://marvin3m.com