4.5 MPU boot issues
Non-booting Bally/Stern MPU boards fall into a couple different categories. If the board's LED is good and turns off at bootup, you're part of the way there. Refer to the flash sequence section to figure out where your issue lies to narrow down the problem area.
You must have a good set of roms in a game as well as a known good working 6800 cpu chip. Suspect any sockets that have a brown color or closed frame - often the spring tension on these sockets is poor necessitating replacement. Be very careful soldering on the mpu boards as the traces and pads are very fragile.
A "cheat sheet" comprised of the signals and pinouts for the primary chips and all of the header connections used with a Bally -35 MPU board is available here. The PDF is specifically for a -35 board, however, Bally -17 and Stern M100 boards are similar. The obvious differences are the ROM jumper settings and connection J5 (J5 on a -35 board has one more pin).
4.5.1 LED locked on
Remove all ICs except for U6, U9, U11, and U2 if it is a Stern. Leon's test rom is recommended for tricky MPUs. This allows you to leave out U11 and check most address and data lines.
Short pins 40 and 39 momentarily on the cpu chip; after doing so, see if the LED goes off. If it does, concentrate your efforts on repairing the reset circuit as it's not holding the reset low on powerup for the minimum required 50 milliseconds. If it doesn't, it's best to pull the board from the machine and put it on the bench, using the benchtop power supply details to keep working on the board outside the machine.
Ensure that there's not simply a problem with Q2 or the LED itself (i.e. the board isn't actually booting up anyway, just with the LED locked on). A logic probe on pin 18 of the U11 6821 PIA will tell you if the LED signal is changing.
An easy quick test for the reset section is putting your DMM on pin 40 of the cpu. It should read 5v, and in most cases if it does the reset is good. Attach a logic probe to the output of Q5 - on power on, you should see this circuit start low and then go high approximately 50 milliseconds later. If it goes high immediately, at a minimum replace Q5, then Q1. Make sure the 2 watt resistor is not leaning on Q5. The heat can kill the transistor.
Next put the probe on U9 pin 3, 36, and 37 to see if you get pulsing - these are all clock signals and you should see pulsing on all 3 pins. If you have an oscilloscope you can visually see the signals, or even on a multimeter, the voltage will show between 2.5-2.9 volts. If you see zero volts or 5 volts, your clock signal is bad. The CPU could be dragging down the clock. Remove U9 and see if the clock signal is now good. When clock signal is bad suspect U15 or U16. Also check C15 and C14. U15 is failed most often.
Check U9 Pin 2. This is the HALT line. You should see 5v, if not replace U9
Check U9 Pin 5. This is the VMA line. It is a pulsing signal and reads about 2.8v with a DMM. If incorrect first try a new U9. Next check U14D, U15C, U19B. U15 is the most likely IC to have failed.
Still locked on? After double checking all your jumper changes, roms, and rams in another board if possible, time to start suspecting other issues such as bad sockets, traces, or addressing chips. Bad sockets are quite common.
4.5.2 Reset Circuit
The purpose of the reset circuit is to ensure the +5 vdc is stable before allowing the system to start. At startup, the reset signal is held low via pull-down resistor R139 until the +12 vdc line rises above the zener diode VR1's value (8.2 or 9.1 volts depending on your board). At this time, the input voltage threshold for the +5 regulator on the solenoid driver board has been met with some headroom. Q1 starts conducting, turning on Q5, which provides the actual reset signal. All of this happens in approximately 50 milliseconds.
Most of the reset circuit is in the corrosion zone (lower left corner of the mpu board) and consists of most of the components commonly included in "repair kits". If the reset circuit is not working, the LED will not turn off. If Q1 or Q5 fails, conducting all the time, the cpu chip will never come out of reset.
The reset circuit continuously monitors the +12 volt line; if it falls below VR1's threshold voltage, the game will reset. If a Bally/Stern game is resetting, concentrate on why the +12 volts is dropping below 8.2/9.1 volts. (VR1 sets the threshold at which the reset signal turns on, and also when it would turn off in a rebooting scenario.)
Note that Q1 and Q5 actually form a switching power regulator for the reset signal only. The normal +5 supplied to the rest of the board is derived entirely from the solenoid driver board's regulator. The reset signal regulated power is used to charge the battery, provide the reset signal, and to power the 5101 chip(s) on the board. Normal +5 vdc is blocked from entering the reset circuit by the 1N4148 diode CR7. Diode CR5 (1N4148) blocks the battery voltage from powering the entire board via CR7.
4.5.3 Bally / Stern MPU Board LED Never Lights or is Locked On
If the LED never lights, either the +12v (TP2, J4 pin 12) is missing or the LED is bad. By default, the LED is ON until the software tells U11 to turn it OFF.
If the LED lights solid, there's some digging to do. First off, if the board has any corrosion damage at all, it needs to be cleaned and neutralized before attempting any repairs. While shot-gunning components might fix the board, it won't be 100% reliable if the corrosion isn't addressed.
Make sure you are getting a solid power path from the rectifier board through the solenoid driver / power regulator board to the MPU board. A locked on LED can be caused by a poor connection anywhere in this chain. Sometimes reseating connector J4 on the MPU board (lower left) will 'clean' a connector well enough to make a better connection. While this may be a short term fix, be aware that any connector, which seems to work better after being reseated, really should be repinned and have its header pins replaced.
Next, put your DMM or a logic probe on pin 40 of the 6800 cpu chip (U9). Power on the board. You should see the voltage remain low / at zero, and then approx 1/10 of a second later, rise to about +5 vdc (high). This is the reset signal, which originates from the components in the lower left corner of the board, and is sent throughout the board to U9, U10, and U11. What the reset section (called the 'valid power detector') does is not allow the MPU to boot until the +12 volts are stable over the value of ZR1 (a zener diode, usually either 8.2 or 9.1 volts). This delay ensures that the +5 voltage is stable enough to run the MPU board reliably (the +5 volts is derived from the +12 volts on the solenoid driver / power regulator board).
The 6800 CPU chip will not 'unlock' and start program execution until it sees a transition from a low (0 volts) to high (~5 volts) signal. This is the purpose of the power on reset delay. The reset delay and signal must be present at all three of the reset inputs at U9 (pin 40), U10 (pin 34), and U11 (pin 34). If the signal starts out immediately at a high level, the MPU will not start to boot until the transition takes place. If you have a locked on MPU, you can take a screwdriver or your meter probe and short pins 39 and 40 together for a brief moment on U9. If the game starts to boot after doing this, it's a safe assumption that the reset circuit is to blame. Shorting the pins together simulates what the reset circuit does.
If you need to rebuild the reset circuit, full kits are available from specialty suppliers such as Great Plains Electronics or Big Daddy Enterprises. The kits include all replacement components in the corrosion zone. Bare bones component replacements are Q1 (2n3904 or 2n4401) and Q5 (2n4403 or 2n3906), but it is a good idea to go ahead and replace all the parts that come in the kit. Replace components one at a time to ensure that you do not mix any up. Note there are some components that are polarized in their installation, which include VR1, CR5, Q1, and Q5. Look carefully at the board to see if there are traces on the top and bottom of a component. The continuity at a through hole can be compromised due to alkaline corrosion. Therefore, it is recommended to solder these components from the top and bottom of the board to ensure that a good connection is maintained.
The following is the list of parts for the reset section that should be replaced. Parts listed with more than one type are equivalent and can be substituted freely. It is also possible that the inductors L1 and L2 need to be replaced as well, however this is very rare. If there is heavy corrosion on them they should be replaced.
Q1 - 2N3904/2N4401 (lower left area)
Q2 - 2N3904/2N4401 (near LED)
Q5 - 2N3906/2N4403 (lower left area)
VR1 - zener 1N9598/1N4738A
CR44 - 1N4004 rectifier diode
CR5, CR7 - 1N4148 switching diode
CR8 - LED
C1, C2 - 820pF ceramic capacitor
C3 - 0.01uF ceramic capacitor
C5 - 4.7uF tantalum capacitor
C13, C80 - 0.01uF ceramic capacitor
Resistors: (1/4 w unless noted)
R1, R3, R24, R28 - 8.2k
R2 - 120k
R11 - 82 ohm/2 watt
R12 - 270 ohm
R16 - 2k
R16 - 2.2k (stern mpu-200 only)
R17 - 150k
R29 - 470 ohm
R107 - 3.3k
R112 - 1k
R134 - 4.7k
R140 - 20k
If your reset circuit is operating as designed, yet the LED is still locked on, next step is to pull all the chips from the board except for U9, U11, and U6 (leave all chips U1-U6 installed on Stern MPU boards. Only U6 is required on Bally boards to perform the initial LED turn off). It helps to have a known working U6 from a Bally game to use as a test chip for this purpose. Be aware that you need to know/have the board jumpered for the correct type of chip you're inserting.
See if the board starts and turns off the LED with just the chips above installed - if it does, add these chips back in this order to see which might be bad: U10 PIA, U1-U5 program chips, U7 6810 ram, U8 5101 ram. Often a bad ram/rom can cause the entire system to lock up. Bad chip sockets can be a factor as well; the early Bally -17 boards have a closed type brown socket that's especially prone to failure.
Double check the jumpers to ensure they match the ROM chips you have available, and change the ROM chips to known working ones for testing. A final thing to check if the machine won't boot is the clock circuit. It is fairly robust, and far more common that the reset circuit itself or chip sockets are the issue. To check the clock circuit you need a logic probe or oscilloscope. A multimeter might show the average voltage on a clock circuit, or it might just show meaningless constantly changing numbers. The clock signal is pin 3 on the CPU chip, and the shifted clock signal goes to pins 36 and 37. The frequency is about 500 kilohertz for Bally -17, -35 and Stern MPU-100 boards, and approximately 850 kilohertz for the Stern MPU-200 board.
If the clock signal is missing, pull U9 first to make sure the CPU chip isn't damaged, and test it again.
It can be frustrating to track down a locked on LED problem, but breaking the problem up and testing each section individually helps. Just remember that if the LED turns on and then off, most of the battle is won. The board has booted far enough that the software was able to start and turn off the LED. Proceed onto the LED flash testing to determine what needs to be fixed beyond that.
4.5.4 Bally / Stern MPU Board LED Flash Sequence
Upon start up, Bally and Stern boards have an LED that flashes. The LED is used to convey the results of specific tests conducted on various parts of the system. This section explains what is being tested and how, according to information from the Bally "FO-561-2 Theory of Operation rev. 5-1982" manual, and the Stern manual "Theory of Operation, Stern's Microprocessor Controlled Solid-State Games".
When power is first applied to the MPU board, the LED by default is ON. The very first set of valid instructions in every Bally / Stern game is to turn the LED off. This is more of a flicker than a flash, so is not counted as a flash in the 7 flash sequence.
126.96.36.199 Quick summary
Flicker: MPU reset good, program booted.
1st Flash: ROM Checksums OK
2nd Flash: U7 6810 ram OK
3rd Flash: U8 5101 ram OK (U8 & U13 on mpu-200)
4th Flash: U10 PIA OK (see details for caveats)
5th Flash: U11 PIA OK (see details for caveats)
6th Flash: U12 555 Display interrupt timer OK
7th Flash: Zero crossing interrupt detector OK (solenoid voltage present)
188.8.131.52 LED flicker:
The LED flicker tells you the CPU chip was able to start a valid program stored in the EPROMs, and that the reset circuit itself is good.
184.108.40.206 First flash:
After the program is running, it performs a checksum of all programmed chips U1-U6. Most Bally games' programming is split between an operating system chip U6 and game ROM chip(s) at U1/U2. Stern games were programmed a little looser: operating system and game code are freely interspersed.
Bally checksums are calculated by summing each byte, and discarding any carries. Most games check their code in $0400 blocks, so it would be possible to determine down to the chip which chip failed this checksum. (A 2716/9316 chip has hex $0800 space available in it - 2732 sized images are $1000 in size. The smallest chip used was a 474 PROM which has $0200 bytes available) However, to do so would require a way to read the X register from the 6800 CPU chip at the time of checksum failure. So if you do not get the first flash, it is best to replace the U6 chip first, then move onto the other chips - U5, U2, and U1 (if present).
Stern checksums are calculated similarly, but not in chunks. The entire program space is summed and must equal $00 for the first flash to occur.
Regardless of whether the manufacturer is Bally or Stern, after the checksum is passed, the first flash occurs.
220.127.116.11 Second flash:
Next, the program tests the 6810 RAM chip at U7, by writing the data $00 to each memory location contained in the RAM ($00-$7F). It then reads back each location to ensure that $00 is returned. It increments the data to $01 and repeats this test. This continues until the data read back is $FF (256), which is the maximum value of any one byte stored in the RAM. It then increments the memory location being tested, and repeats the $00-$FF data storage test. If any of the tested locations return an unexpected result, the program stops, and alerts you to a problem with U7 (since you got the first ROM checksum flash, but not the 2nd U7 OK flash).
18.104.22.168 Third flash:
Now, the program tests the non-volatile 5101 RAM chip at U8 (U8 AND U13 on Stern MPU-200 boards). The 5101(s) store(s) bookkeeping data, game parameters, high scores, replay levels, etc. The program tests this RAM ($200-$2FF) by reading the original nibble / byte (see sidebar), and saves it in a temporary location. Then, it stores a test pattern in the location similar to the U7 test. After the byte successfully passes the test, the original data is returned to the location, and the program loops onto the next byte.
LEARN MORE: How does a 128 byte 5101 RAM occupy 256 memory locations?
If you look at a pinout of the 5101 memory, you will notice it is a 128 byte device. Yet, it is addressed by the MPU via 256 memory locations ($200-$2FF). This is because the 5101 is actually a 256 nibble device - a nibble is a half-byte (4 bits). So data stored to a 5101 in a pinball machine actually only stores half of the data byte being sent to it. Which half it stores is dependent on the board design. Bally and Stern use the upper nibble for storage, and Williams used the lower nibble. Stern MPU-200 boards have an additional 5101 at U13. This stores the lower nibble in conjunction with U8 storing the upper nibble of a byte saved to $200-$2FF. This allows MPU-200 games to store more data, and avoid doing some fancy processing by getting the data in and out of the non-volatile ram area.
For example, here's some pseudo-machine code for what happens:
LOAD #$24 (the data you want to store is 24)
STORE $231 (you want to store the data 24 at memory location $231)
READ $231 (you want to read back the data you just stored)
The data returned is not #$24 as expected, but rather #$2F. The lower nibble was never stored, because the 5101 memory does not store data as bytes but rather as nibbles. To store #$24 properly would require splitting the byte into its nibbles '2' and '4'. The 2 would be stored in one memory location, while the 4 would be shifted, and stored in another memory location.
Showing the byte as binary might be helpful to visualize what's involved. The hex #$24 in binary is %00100100. Split into nibbles is %0010 (the 2) and %0100 (the 4). The upper nibble is the one the 5101 is able to store directly, but the position in the byte of the lower nibble prevents it from being stored. A shift operation is performed 4 times on the byte to reposition the lower nibble as the upper nibble, which enables it to be stored to the 5101. Each shift moves the binary pattern to the left one bit - here's the full sequence:
This gives the #$4 in the high nibble. All byte data has to be split this way to be saved, and recombined upon reading from boards (Bally -17 and -35, Stern MPU-100) with a single 5101 RAM chip. You can see why Stern added the second 5101 RAM to their boards. It makes programming much easier!
22.214.171.124 Continuity Chart for U8 5101 RAM in a Bally AS2518-17, -35 or Stern MPU-100 Board
IC & Pin# Data Address Continuity Points
U8 pin 1 (A3) U7 pin 20, U6 pin 5, U9 pin 12
U8 pin 2 (A2) U7 pin 21, U6 pin 6, U9 pin 11
U8 pin 3 (A1) U7 pin 22, U6 pin 7, U9 pin 10, U11 pin 35
U8 pin 4 (A0) U7 pin 23, U6 pin 8, U9 pin 9, U11 pin 36
U8 pin 5 (A5) U7 pin 18, U6 pin 3, U9 pin 14
U8 pin 6 (A6) U7 pin 17, U6 pin 2, U9 pin 15
U8 pin 7 (A7) U7 pin 15, U6 pin 1, U9 pin 16, U11 pin 24
U8 pin 8 GROUND
U8 pins 9&10 (D10&D00) U7 pin 6, U6 pin 14, U9 pin 29, U11 pin 29 (* pins 9 & 10 siamesed together)
U8 pins 11&12 (D11&D01) U7 pin 7, U6 pin 15, U9 pin 28, U11 pin 28 (* pins 11 & 12 siamesed together)
U8 pins 13&14 (D12& D02) U7 pin 8, U6 pin 16, U9 pin 27, U11 pin 27 (*pins 13 &14 siamesed together)
U8 pins 15&16 (D13&D03) U7 pin 9, U6 pin 17, U9 pin 26, U11 pin 26 (* pins 15&16 siamesed together)
U8 pin 17 (CE2) Q5 Right 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, U9 pin 34, U11 pin 21, 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
126.96.36.199 U8 & U13 5101 RAM Continuity Chart for Stern MPU-200 ONLY
IC & Pin# Data Address Continuity Points
U8 pin 1 (A3) U13 pin 1, U7 pin 20, U6 pin 5, U9 pin 12
U8 pin 2 (A2) U13 pin 2, U7 pin 21, U6 pin 6, U9 pin 11
U8 pin 3 (A1) U13 pin 3, U7 pin 22, U6 pin 7, U9 pin 10, U11 pin 35
U8 pin 4 (A0) U13 pin 4, U7 pin 23, U6 pin 8, U9 pin 9, U11 pin 36
U8 pin 5 (A5) U13 pin 5, U7 pin 18, U6 pin 3, U9 pin 14
U8 pin 6 (A6) U13 pin 6, U7 pin 17, U6 pin 2, U9 pin 15
U8 pin 7 (A7) U13 pin 7, U7 pin 15, U6 pin 1, U9 pin 16, U11 pin 24
U8 pin 8 U13 pin 8, GROUND
U8 pins 9&10 (D10&D00) U13 pins 9 & 10, U7 pin 6, U6 pin 14, U9 pin 29, U11 pin 29 (* pins 9 & 10, siamesed together)
U8 pins 11&12 (D11&D01) U13 pins 11&12, U7 pin 7, U6 pin 15, U9 pin 28, U11 pin 28 (* pins 11 & 12 siamesed together)
U13 pins 11&12 (D11&D01) U7 pin 3, U6 pin 10, U9 pin 32, U11 pin 32
U8 pins13&14 (D12&D02) U13 pins 13 & 14, U7 pin 8, U6 pin 16, U9 pin 27, U11 pin 27 (*pins 13 &14 siamesed together)
U13 pins 13&14 (D12&D02) U7 pin 4, U6 pin 11, U9 pin 31, U11 pin 31
U8 pin 15&16 (D13&D03) U7 pin 9, U6 pin 17, U9 pin 26, U11 pin 26 (* pins 15&16 siamesed together)
U13 pins 15&16 (D13&D03) U7 pin 5, U6 pin 13, U9 pin 30, U11 pin 30 (* pins 15&16 siamesed together)
U8 pin 17 (CE2) U13 pin 17, Q5 Right Upper leg, U9 pin 40, U11 pin 34
U8 pin 18 (OD) U13 pin 18, U14 pin 9
U8 pin 19 (CE1) U13 pin 19, U17 pin 8,
U8 pin 20 (R/W) U13 pin 20, U7 pin 16, U9 pin 34, U11 pin 21, U14 pin 10
U8 pin 21 (A4) U13 pin 21, U7 pin 19, U6 pin 4, U11 pin 22
U8 pin 22 (Vcc) U13 pin 22, C13 Left Leg, R12 upper Leg, CR5 Lower Leg
188.8.131.52 Issues with 5101 RAM chips
The socket at U8 or U13 is very close to the corrosion effects of the alkali of a leaking battery. This is often the cause of boards not showing a 3rd flash. There are traces on the component and solder side of the board, and the socket can hide corrosion to these traces. In turn, the corrosion can affect the 5101 chip's contact with the socket pins.
Unless the board is unusually clean, a rebuild of a non-working board should include replacing the socket at U8 (and U13, for MPU-200). The chip is an oddball configuration with 22 pins which makes machine pin sockets hard to find. Use 2 strips of machine sockets, so that you can solder both above and below the board. Inspect the traces closely on the top; it is best to avoid soldering on the top of a board unless you have to as it makes it extremely difficult to desolder the sockets in the future. Take extreme care not to make a solder bridge of the traces on the component side, as it is easy to do. Performing a continuity test to adjacent socket legs after soldering is a good practice. Also, note that some of the pins are shorted together. (see above)
Failure to achieve the third flash can also be attributed to the following: If the U8 and U13 (MPU-200) show correct continuity as charted above, and the 5101 RAM is known to be good, the problem could lie with the U19 (4011) chip. On Stern MPU-200, replace U14 (4572) instead. Finally, although rare, the 6800 CPU at U9 could be bad.
The speed of the 5101 RAM chip can make a difference in the functioning of an otherwise good MPU board. If a slow chip is put into a good MPU, strange behavior such as inconsistent boot ups, incorrect score display behavior, among other things can result from the RAM not being able to keep up with the demand from the CPU. Below is a list of 5101 RAM chips and their speed. A lower number means a faster chip. A Stern MPU-200 board needs 2 chips of at least 450ns, while the Bally AS 2518-17 or -35 and the Stern MPU-100 can work correctly with a slower chip of 650ns. The stern mpu-200 board should have matching speeds for the 2 rams as they are selected and accessed simultaneously. A faster chip is fine for a replacement, but there will not be a performance enhancement.
PCD5101P (Philips manu) 150ns
5101-8 800ns (too slow for any board)
The 5101 RAM chip is especially sensitive to static discharge that will damage it, so take extra precaution in handling. The memory RAM replacement chip sold by Tom Callahan at pin-logic.com uses a 6116 and special adapter, and makes a good replacement. Pin-logic also sells the static ram chips that do not require batteries for backup; although he advertises that these do not work in the Stern MPU-200 board, 2 of them did work on a test board. Another source for varying types of 5101 and 6116 replacements is a blog site WarpZoneArcade. There is information how to roll-your-own 5101 replacement adapters. Or, visit its accompanying site, PinForge, and purchase pre-made RAM adapters.
184.108.40.206 Fourth and Fifth flashes:
Next, the program tests each of the 2 6820/6821 peripheral interface adapters (PIAs) at U10 and U11, starting with U10. The PIAs are set to a known state, then data is stored and read back from them to verify their registers are functioning properly. It is important to note that it is not possible for the PIA to be 100% tested with this test, as external data would have to be fed in to do so. However, the test will at least test the internal registers. It would be possible for a PIA to pass the self-test, but still not work properly with external inputs.
Assuming the PIAs pass, the fourth (U10) and fifth (U11) LED flashes occur. The LED itself is connected to the U11 PIA. So if the LED is locked on, U11 might be bad. It's worth letting a locked on LED board 'sit' for a minute or so to see if the game boots all the way up without flashing each test step. This is an example of how a PIA can pass self test but still be bad. The LED control pin has no feedback as to if the LED is in fact flashing.
Because the PIA tests all possible values in all possible registers, anything connected to the registers will be activated very briefly as well. This manifests itself most obviously via the flipper relay which on most games will click during the U11 PIA test. Holding the flipper buttons in will usually result in a flip or half-flip during this test. This could be helpful if purchasing a machine/repairing one in a noisy environment and you want to see how involved you might getting before removing the backglass to visually see the LED diagnostic flashing.
220.127.116.11 Sixth flash:
The sixth flash waits for an external input on U11 pin 40 from the display interrupt generator circuit, which occurs 320 times a second. If you're missing the sixth flash, there may be a problem with either the U12 circuit, OR the input pin on the PIA. A logic probe, oscilloscope, or a multimeter on pin 40 can help you determine which is at fault. A logic probe will pulse if the display circuit is operating; the scope will show you the signal's waveform; and the multimeter should settle on a voltage somewhat between 0-5 volts.
One definite reason for lack of a sixth flash is a poor connection on the ground side of C16, the film capacitor. If there isn't a sixth flash, but all other tests and measurements are good, it's a safe bet that U11 is bad and needs to be replaced.
Note that the sixth flash does NOT check for the proper frequency of operation of the display generation circuitry. As long as there is a pulsing signal (technically, ONE state change), the test is marked good and the program allowed to continue.
18.104.22.168 Seventh flash:
The last flash waits for an external input on U10 pin 18 from the zero crossing detector circuit, which occurs 120 times a second (as the AC waveform passes or "crosses" 0 volts). Diagnosis of issues with the 7th flash are similar to the 6th flash. You can measure the input to pin 18 to determine if the signal is present or not. A signal present, but no flash could mean a bad U10 PIA. A missing signal usually points to missing solenoid voltage. The source of the zero crossing signal is derived from the solenoid voltage delivered from the rectifier board. Equally, if a signal is present, there may be an issue with the zero crossing detection circuit itself.
Note again that the seventh flash does NOT check for the proper frequency from the zero crossing detector. It simply checks for a pulsing signal, and only checks for ONE transition.
After the 7th flash, the program does some background setup: reads dip switches, enables the displays, attract modes, switch scanning etc. in a 'game over' mode, waiting for player input. The LED will sometimes be dimly glowing or even pulse as this happens, which is not a cause for alarm. You can rebuild the LED circuit around Q2 if this worries you, but it is harmless.
One interesting anomaly is that the 7th flash will not occur, if there is a bad 1N4004 diode (CR3) used on -32 / -50 sound boards. On these two sound boards, the +12VDC used on the boards is derived from the +43VDC solenoid voltage, and isolated by CR3.
Bally MPU with burned R113 and R16
If you have +43VDC solenoid voltage, but no 7th flash. Check R113 and R16 on the MPU. These two resistors take the 43v and form a voltage divider. If these resistors burn they can stop the 7th flash from happening. Test point 3 should show approximately 21.5 pulsating DC if the resistors are ok.
Still no 7th flash and you have 21.5 vDC on TP3, check CB1 which is on U10 pia pin 18. It is the logic side of the zero crossing and should be pulsing. If it is stuck low or stuck high replace u14.
4.5.5 Chip Sockets
Reliable socket connections are a requirement for any printed circuit card to work as designed. Old sockets, as discussed below, should be replaced. Use extreme care in desoldering the old sockets, the traces and pads on Bally Mpu-17 boards are easily lifted. It is possible to lift other boards' traces and pads as well, especially if any battery corrosion is in the area.
The chip sockets on old Bally and Stern boards (also most any board of this era) are long past their reliable lifetimes. They may work, but they may also cause intermittent connections that will have you chasing your tail tracking down odd problems with your game. Like the 40-pin interconnect used in Williams System 3-7 games, these sockets should always be replaced. On Bally/Stern MPUs, these include U2, U6 (more game ROMs if you don't combine the ROM images), U7 RAM, U8 5101 RAM (and U13 5101 RAM if a Stern MPU-200), U9 CPU, U10 and U11 6820/6821 Peripheral Interface Adapters.
Perhaps the most maligned socket brand, and rightfully so, is the Scanbe brand. In the picture below, you can see why. These 30+ year old sockets passed the point of reliability many years ago. Included in the Scanbe socket picture below are two pins pulled from a Scanbe socket. The pins were designed to grip the SIDES of the IC legs, unlike the design of modern sockets that grip the front and back faces of the IC leg. Get rid of them now.
Poor quality and old sockets found on many early pinball machine circuit boards
It's hard to believe that these "brown" sockets ever worked.
The spring strength of the pins in these AUGAT sockets eroded over time.
Closeup of the infamous "Scanbe" brand socket. "Scanbe" is molded into the top of the socket face.
One Method of Socket Removal (works for at least ScanBe and Augat sockets)
Prying off an OEM Augat socket. Be careful to not crush/damage traces as you pry.
Heat and remove each pin individually.
Clean up the through holes with your favorite solder sucker. Here, a Hakko 808 is being used.
this is a stub
The Motorola part number equivalencies that appear on the top of U9, U10, and U11 are...
SC44216P = 6800 microprocessor
SC44067P = 6820 PIA
4.5.7 Clock Signal
replacement 9602 board: http://www.homepin.com/9602.html
replace with a 6802: http://pinballeon.com/6802/e6802.htm
4.6 How to make a Benchtop power supply for the MPU board
ATX Power Supply connector modification
Bench Top Power Supply hooked up to MPU board
Being able to test the MPU board on the bench is a great advantage in trouble shooting. A benchtop power supply is quite easy and inexpensive to put together. All that is required is a computer supply and some mods to the connectors. An ATX or old AT computer power supply will supply the + 5 vdc and +12 VDC that is needed to run the first 6 flashes. The 7th and final flash requires +43 vdc and cannot be supplied by the ATX power supply, but the MPU can be "fooled" into thinking it is present.
An ordinary ATX computer can be had used or new for under $20. Exact wattage is unimportant. In order to be able to switch the power on and off, the 20 pin connector on a ATX supply must be modded. This is not necessary on an older AT supply. Because modern motherboards use a soft power on/off, the green pwr-ON at pin 14 must be tied to the black COM ground at pin 13 or 15 thru 17. This can be done easily done with a short length of wire and a .062 pin attached to each end and shoved into the 20 pin connector at pin 14 and 15. Or, simply cut the wires free from the connector and solder them together.<bt>
Take one of the molex 4 pin connectors which have 1 RED wire (+5 vdc), 1 YELLOW (+12 vdc) and 2 BLACK ground wires. Attach an alligator clip to the RED, YELLOW and one BLACK lead. A good tip is to use the color coded rubber boots to protect from accidental shorts and keep it easier to identify the leads at a glance. If you should accidentally short the +12 volts to +5 via you will kill all the ICs on the board!
To use the power supply, with the power OFF, clip the black lead to TP 4 at the top right of the board. Clip the RED +5 vdc lead to TP 5 at the bottom right, near the battery. Finally clip the YELLOW +12 vdc to TP 2 on the bottom left of the board. Make certain that this connector is not on or shorted to TP3 which is nearby. This error would damage the board, so double check this before turning on the power. Connect the power supply to 120vdc outlet and flip the switch to boot.
Upon power up, the brief flicker should be seen, then 6 (not 7) more flashes for a good board. The 7th flash cannot be achieved without +43vdc present. On some occasions, the LED will lock on and not go into the flash sequence. This COULD BE because the MPU is sensitive to the exact voltage supplied and will not boot with a computer supply. If the board successfully boots in the game, this could be the cause. This condition is rare, but possible. Also, if the LED locks on immediately, try doing a manual reset by briefly shorting together pins 39 and 40 of the U9 CPU with a small screwdriver. This forces the pin 40 to go low and begin the boot process. It is also possible to short the junction of resistors R1 and R3, on the right side of R1, to board GROUND with a jumper clip to accomplish the same effect.
4.7 Solenoid Driver Board Issues
4.7.1 Testing and Replacing Transistors
Before and After of Replaced Drive Transistor, Diode, and Resistor
To test a transistor set the DMM to diode test mode. With the game turned off, place the black lead on the metal tab of the transistor. Probe the two outer legs of the transistor with the red lead. The DMM should read between .4v and .6v (some DMM will show 4xx - 6xx). The center leg should be a dead short to the metal tab. If either outer leg reads anything outside of the .4v to .6v range, the transistor more than likely needs to be replaced. Use a TIP-102 as the replacement transistor.
If a transistor needs to be replaced it is usually a good idea to check / replace the 1n4004 diode and 330 ohm resistor associated with that transistor as these will often burn / short. Also check the diode on the associated coil, it is likely to be shorted as well.
The transistor output can also be tested by using the diagnostic LED on the MPU board. An alligator test lead and a component lead clipping or finishing nail are needed.
With the power off, attach one end of the alligator test lead to TP6 located on the MPU board. TP6 is located in the vicinity of of the diagnostic LED.
Next, attached the other end of the test lead to the component lead clipping or finishing nail.
Turn the game on, and place it into solenoid test or just allow it to remain in attract mode.
Carefully place the lead with the finishing nail into one of the many solenoid outputs on the solenoid driver board (J1, J2, or J5).
If the LED lights briefly as the solenoid test cycles, the transistor and header connection on the solenoid driver output is good.
If the LED does not light, suspect the following...
a failed transistor,
a cracked header joint,
a discontinuity between the transistor and the header pin,
the associated diode on SDB has failed,
the 3081 transistor array on the SDB,
the 74154 4 to 16 decoder on the SDB, or
failed connections on J4 of MPU and / or J4 of the SDB, or bad daisy-chain connection at J1 of the sound board (if a sound board is used)
If the LED remains on, this means that the transistor is shorted. The exceptions are the diagnostic button return connection (ground) located at J2 pin 7, and the coin lockout coil ground at J2 pin 8. The coin lockout is a high resistance coil, and can remain on without destroying its associated drive transistor.
The above test using the MPU diagnostic LED can be performed with the J1, J2, or J5 connectors connected or disconnected. WARNING: DO NOT place the lead on the J3 or J4 connectors on the solenoid driver board.
4.7.2 Replacing a failed Solenoid Transistor
If a coil has locked on, or any SE9302 transistor has failed the diode test with your multimeter, replace the following components on the Solenoid Driver board as a set. In addition, also test the playfield coil disconnected from the game wires, with the ohm setting. The coil should have a value higher than 3 ohms up to around 15 ohms. Lower than 3 ohms is a dead short and will heat up and burn up components if the fuse does not blow. The coil diodes should also be replaced if the coil has locked on. Since the diode must be disconnected from the lugs to test it, it just makes sense to clip it off, replace the diodes, and not bother to test them. 1N4004 diodes or better, (1N4005,6 or 7) should be used. You cannot test the diode in place. The diode band (cathode) goes to the power lug of the coil, usually a double wire connection, because the coils are daisy chained. If it is a single wire, look at another nearby coil and note the color of the double wire.
Replace any suspect SE9302 with a TIP 102 transistor. They are used in many pinball machines, so having a small supply of them is highly recommended. The TIP 102 is a more robust transistor and can handle more current. On the SDB replace the associated diode, and resistor for any driver transistor you replace. Check the associated pin for signs of burning and replace the header pins and connector pins if they have turned brown or look corroded. If a diode and/or resistor has actually burnt, check for continuity and make sure the circuit board traces have not also burnt up. This is a common problem when someone has over-fused the game.
4.7.3 Solenoid Transistor Mapping
Bally Solenoid Transistor Mapping
Stern Solenoid Transistor Mapping
To view a chart of what specific transistor controls a specific coil on a particular Bally game (transistor mapping), click the image at left.
To view a chart of what specific transistor controls a specific coil on a particular classic Stern game (transistor mapping), click the image on the right.
4.7.4 Solenoid Driver Board Transistor, Resistor, Diode & associated Connector Chart
TIP 102 1N4004 Diode 330 ohm Resistor SDB Connector
Q1 CR1 R9 J1-pin 2, J2-pin 9
Q2 CR2 R6 J1-3, J2-4
Q3 CR3 R16 J2-5, J3-4
Q4 CR4 R18 J1-5
Q5 CR5 R10 J2-10
Q6 CR6 R12 J2-11
Q7 CR7 R14 J2-12
Q8 CR8 R20 J5-10
Q9 CR9 R26 J5-9
Q10 CR10 R28 J5-15
Q11 CR11 R32 J5-14
Q12 CR12 R30 J5-13
Q13 CR13 R22 J5-12
Q14 CR14 R24 J5-11
Q15 CR15 R39 Flipper Relay
Q16 CR16 R34 J2-6, J3-7, J5-8
Q17 CR17 R42 J5-7
Q18 CR18 R45 J2-15, J3-9, J5-3
Q19 CR19 R47 J2-8
TIP 102 1N4004 diode 330 ohm Resistor SDB Connector
4.7.5 Solenoid Driver Upgrades
Visual of Solenoid Driver Upgrades
There are several upgrades which can be performed to reduce stress on connectors.
The two 5v test points on the SDB can be tied together. On the solder side of the driver board jump TP1 and TP3 together. Be careful to apply this to the correct test points. This eliminates a possible failure point on J3 pins 13 and 25. Note that TP1 is for the raw +5vdc coming directly from the regulator. TP3 is for the voltage that feeds all the +5vdc circuits in the driver section. If there's a problem in the driver section, it can be isolated by removing this jumper, and removing pin 13 or 25 from J3.
Next the C23 capacitor needs a common ground with the rest of the SDB. Jump the negative side of C23 to a convenient ground trace. On most boards there is a ground trace in the vicinity of the C23 negative lead. Other boards might require a longer jumper. Scrape the solder mask off the trace to get a good solder surface to solder the jumper. If replacing the C23 capacitor, (a good idea, they can be 30+ years old at this point), it may be possible to leave the negative lead "long" on the new capacitor. Then, bend the leg down to the ground trace, and solder it in both the mounting hole and the trace.
The C26 capacitor needs a ground upgrade similar C23 mentioned above. Tie the negative side of C26 to the ground trace located at the perimeter of the board. This connects both C23's and C26's negative side to a common ground on the SDB. The goal of these modifications is to provide redundant grounds with the least resistance possible.
4.7.6 How to Rebuild the Bally/Stern Solenoid Driver Board
The Solenoid Driver board is critical to the operation of the electronics of the game. The Solenoid Driver Board (SDB) comes in several types for Bally and Stern, and are completely interchangeable for any game of the 6800 MPU type. The SDB supplies the game with the voltages for the MPU, coils and high voltage for the displays. Recommendations are given below for replacing components to ensure proper operation and reliability. In all cases, do the ground modifications as shown in another part of this Wiki.
Minimum Recommendation for a Working board
At a minimum, replace both large capacitors on a working or non-working board, especially if they look original. Original caps are often metallic blue or metallic silver, and are at least 30 years old and due to fail, owing to the electrolytic chemicals in the capacitor drying up. The capacitor at C 23 has a factory value of 11,000 uf and 20 volts, but these values are not easily found these days. An electrolytic capacitor of between 11,000uf and 16,000 uf and 25 volts or greater can be safely substituted. A screw terminal capacitor makes for a easy installation, but a snap cap with leads can be used as well. Recent prices for screw terminal caps have increased greatly so this may be a factor in your decision.
The high voltage capacitor at C26 has an original spec of 160 uf and 350 volts and is an axial electrolytic capacitor. Once again, this part is difficult to encounter with these exact specs, but an axial or radial cap from 150uf to 180 uf and at least 350 volts can substitute. 400 v or 450v caps can be had in a radial format and size that will fit. With a radial cap you will have to make leads that bend back to the negative (-) solder pad.
Preferred recommendation for a working board
Clearly fractured .156 header pins on a Bally single sided driver board
Besides the above, replace ALL the .100 and .156 molex header pins and the connector terminals in the nylon housings with Molex TRIFURCON Phosphor Bronze tin plated crimp terminals (Molex Part# 08-52-0113 for 18-20 ga. Wire, and 08-52-0125 for 22-26 ga.) for the .156 connectors, and Molex .100 tin plated Phosphor Bronze crimp terminals, (Molex Part#08-52-0123). Do NOT skimp on this step and just do one or the other. The receptacles can be re-used if not burnt. If replacing the receptacles, the locking or non- locking ramp type are fine. Often, the connector at J5 does not cover the last few pins. I don’t like this personally, although no harm can be done if the key is in place.
Check every resistor that is usually covered by the plastic shield between the two big heat sinks for correct value and no sign of burn. Replace any that are suspect. Use the diode function on your multimeter to check the zener diode at VR1 and the 1N4004 diode. The zener diode can be difficult to source. See below for recommendation.
Gold Standard for Working or NON-WORKING board
It doesn’t make too much sense trying to trace down the exact problem and replacing only the bad components on a NON-WORKING board. Better to replace every possible bad component and start fresh. To ensure long life and proper operation of a working board, in ADDITION TO THE ABOVE, it is recommended that the components below are replaced regardless. Below is a list of components that effect the HV and +5 volt logic circuits on the SDB. Replace them all! * The Stern Revision J board is rather different, & a component list for that board follows. It can be identified by the extensive silk screening of each component and its function as seen in the 4th photo above, and by the HUGE 2 Watt and (2) large 1 Watt resistors. It is most often found on late date Stern games like Flight 2000 or Viper, etc.
Component replacement list for Bally AS2518-16 or -22 and Stern SDU100 SDBs, except * revision J
R51 22k ohm 1/2Watt
R52 390 ohm 1/4 Watt
R54 8.2k ohm 1/4 Watt
R55 1.2k ohm 1/4 Watt
R56 82k ohm 1/2 Watt
R35 100k ohm 1 Watt
CR21 1N4004 400PIV 1 amp or better (1N4007, etc.)
VR1 Zener diode 1N5275A 140 volts, 1/2 Watt. Can Sub 1N5275B or NTE 5099A
Q21 2N3584 250volts, 2 amp, TO-66 NPN
Q22&23 2N3440 250 volts 1 amp TO-39 NPN
Q20 LM323K (original 78H05KC or LAS1405)
C27&28 .01 uf 400 vdc metal polyester capacitor
RT1 25k ohm potentiometer (2 types) a 15mm black one, Piher Part# PT15LH06-253A2020 or a 6mm blue one BournsPart#3306P-1-253
Stern SDU100 revision J
R35 100k ohm 1 Watt
R51 33k ohm 2 Watt
R52 390 ohm 1/4 W
R53 2.4k ohm 1/4 W
R54 8.2k ohm 1/4 W
R55 1.2k ohm 1/4 W
R56 82k ohm 1/2 Watt
R73 3.3 ohm 1/4 W
R74 470 ohm 1/4 W
R75 100k ohm 1 Watt
CR22 1N4004 or better
Q24 2N3904 Transistor
Other diodes, transistors and capacitors as above for AS2518-18
4.7.7 Modify the Display Fuse holder
Fuse adapter for the SDB allowing use of a common fuse
Some versions of the solenoid driver board have a small fuse present for the displays. This fuse is a difficult to find, expensive, type 8AG 3/16 amp fast blow fuse. It would be best when rebuilding the SDB to replace the fuse holder with new clips and convert it to a more commonly available 3AG type. The original value of 3/16 amp can be substituted with a more common 1/4 amp fast blow fuse without worry of over fusing the circuit, due to inherent variations of tolerance by the manufacturer.
Perform the following to replace the existing 8AG fuse holder:
Bend up or clip off the upper solder lead of the clip
Move the TOP fuse holder up one hole
Solder a 20 ga. wire from the base of the clip back down to the lower hole.
Be careful not to solder the clips with the inner "ears" facing in the wrong direction, or the work will have to be redone. It may be helpful to put an old 3AG fuse in the clips to aid in spacing, and assure that the "ears" are positioned correctly. The fuse will also hold the clips in place while soldering. See the photo above. Of course, an inline fuse holder can also be used instead, but looks less professional.