(Topic ID: 251495)

Black Knight: PIA issues (was: MPU Lockup)


By kevinclark

33 days ago



Topic Stats

  • 80 posts
  • 9 Pinsiders participating
  • Latest reply 16 hours ago by pincoder
  • Topic is favorited by 1 Pinsider

You

Linked Games

Topic Gallery

There have been 6 images uploaded to this topic. (View topic image gallery).

test (resized).png
0FBBD37B-A2C8-4945-B7B6-7B9B2478A65E.jpeg
8344D897-0AFC-4A74-B6EE-972235C5ED56.jpeg
8C8AD6B1-1FFB-49B7-ABED-4989914AD417.jpeg
7D2256FC-750C-4476-8BC6-60088FDF4915.jpeg
IMG_6273 (resized).jpeg

There are 80 posts in this topic. You are on page 1 of 2.
#1 33 days ago

I’m having a rough night. I had gotten my BK to playable and had started handling cosmetic things. But when I went to start it up to play a celebratory round, one of the upper flippers went out again. I adjusted the switches and still no love, so I shorted the coil and got no response. So I shorted the other upper coil to sanity check and got a weak response. Huh. Then I smelled the smoke. I’m still not sure what went wrong.

I seem to have fried my MPU board. Taking out the driver board had no effect. I got the MPU on the bench and confirmed the reset circuit is showing 5V on pin 40 of the MPU and the address bus pins appear to have signal. From what I’ve read it sounds like my next best tool is to put in a test rom, but I don’t have one.

Do I have another alternative? If not, how should I pick a burner? Is eBay my best bet? Any recommendations?

#2 32 days ago

I like the Minipro TL866CS. It's cheap and it does work. The II version no longer does >16volts though, so you might have to be picky about your replacement eproms, and you have to build an adapter for 2532's which stock your BK probably uses. I paid between $36 and $75 for my 3 (As soon as they discontinued the CS in favor of the II I wanted a couple backups) depending on various adapters that I'll probably never use.

There are lots of burner options though if you go up to a bit over a hundred I keep meaning to get the GX-4Q that a lot of people use but just haven't gotten around to it yet. I have an ancient Dataio which is terrible for a lot of reasons (parallel port only, dos only, eprom definitions fixed/no flex in parameters, and very SLOW.... which might have been fine back in 1997 when it was new, but today just doesn't cut it)

If you have a logic probe you can do some very basic testing of the CPU board, take the roms out and you should see activity all across the address bus in faster patterns as you go A15-A0 - without a rom the game will just cycle through all the addresses if it's booting. Disconnect the driver board and just leave the power plugged into the mpu board, or take it out and test it on the bench. If you don't even have that, there's something seriously fried.

#3 31 days ago

Thank you! I’ve ordered a gq-4x4 and some spare 2532 and 2716. The CPU seems to be running fine, so I hope I just blew out some ROM.

#4 29 days ago

Ok! I was able to run the roms from pincoder and I'm pretty sure I've got a dead PIA. I loaded up the solenoid test, but I'm not sure how to check the whether the flipper relay is energized. The flippers don't do anything at all. And in retrospect, I understand how I fried this thing but not exactly what I fried. I was moving too quickly and attached a test clip to the horizontal metal bar as a ground to test the coil. I'm only now seeing the switch attached. Soooo.

Could someone tell me what it is? Is checking the flipper relay a moot point? Should I just try to track down a new PIA?

IMG_6273 (resized).jpeg
#5 29 days ago

Also, my board's PIAs have the williams logo on them and the id: "SA-8972 8107". Though the schematics says MO 6821. I should replace with the latter, yeah?

#6 29 days ago

Yes, the williams' part is just a 6821 custom marked for them when they were buying 1000s of them at a time in the 80s.

#7 28 days ago

Looks like the horizontal bar was the ball-roll-tilt. So, I’m assuming I put way too much voltage (coil level) through the switch matrix via the green-brown wires, I would expect the PIA to be fried, right? Anything else I should check while I’m waiting for replacements to come in?

#8 28 days ago
Quoted from kevinclark:

Looks like the horizontal bar was the ball-roll-tilt. So, I’m assuming I put way too much voltage (coil level) through the switch matrix via the green-brown wires, I would expect the PIA to be fried, right? Anything else I should check while I’m waiting for replacements to come in?

there is 4049 and 7406 chips between the switch matrix and the pia so those will die before the pia when you put high voltage on a switch matrix wire. I guess its possible for them to short input to output and toast the PIA too.

The Williams and Motorola logo are really similar. Turn one upside down and it looks like the other. The PIAs are actually motorola with a custom part number for williams. They are 6821s and any brand works (ef6821, mc6821, s6821, hd6821 etc.)

#9 27 days ago

I think I've been chasing my tail. I'm seeing identical behavior on the board with or without IC17 inserted. I'm not sure why the test rom isn't getting started.

It look like the address busses are cycling and I've got continuity from IC17's pins to the data buffer to the CPU. I've got continuity from the CPU's address pins to ICs 3 and 4 and from those to IC17.

The EPROM burns and does verify ok. I'm not currently using an external power adapter for my burner (gq-4x) but my system claims 500mA are available and the DC power supply people mention only supplies 200mA.

I've checked continuity from the legs of the EPROM when inserted into the socket to the solder pads below the socket.

I have the driver board connected to the MPU, but was seeing the same behavior without it. My test bench setup is an ATX power supply with an adapter board and I've connected ground, 12v, and 5v to IJ2. Two leds light up but no digit.

What am I missing re: requirements for reading and running from the ROM?

Edit: Leon’s guide says only 5V should be needed plus a jumper from C23 to R3, but when I try that the whole thing turns off, including my power supply adapter board, after about 1s. If I add 12v this doesn’t happen.

#10 26 days ago

Here are my more detailed notes per IC incase something stands out:

IC1
* Clock signal on 38 a little high (1.8v vs 1.5. Probably fine?)
* 39 checks out. 40 is showing reset voltage (4.5v)
* Pin 2 and 3 at 5v voltage (4.3 coming from power supply, shows up there)
* Pin 5 showing 3v correctly

IC 4, 3, 8, 9 all look ok, though voltage is coming with a slight bump coming out the buffer
* Consistent bump though

IC 18:
* Pin 2-8 don't show up on the logic probe but show < 2v (1.7-1.9) with dmm
* 9-17 show ~5v

IC 36:
* Pin 2-8 show ~ 2.3v, but show nothing on the logic probe
* Pin 9 is 0.5v
* 10 5v, 11 1v, 12-14 5v, 15-16 0.5v, 17 5v, 18 0.5v, 19-20 5v

#11 26 days ago

I’ve pulled off the driver board to simplify things and see no change. Removing all the non-IC17 ROMs also had no change.

#12 26 days ago

Leon mentions IC13 and 16 need their CS pin (8) to be at 5v on startup. I'm seeing mine at 3.5v. The actual power being supplied by my 5v pin on the power supply is closer to 4.6 and I see that at TP 9. Seems like that might be one of my issues.

Edit: Course, half the PIA pins being at voltages right in the middle, where it's not high or low, seems like an issue too. But I'm not sure if that's because the ROM isn't running or *causing* the ROM not to run. Having bad PIAs sound like they'd keep the digit from displaying though?

And I've verified the data bus looks like it's getting real signals. But nothing is coming out of the PIA I outputs and going to the digit. So it sure seems like it's at least one major issue. I've got PIAs coming in and I'll put in a socket and swap it out and see what happens unless someone has a better idea for a next step.

#13 24 days ago

I’ve replaced PIA I on the MPU and now I’m getting flashing lights from the test ROM. Still no Diagnostic Segment Display.

#14 24 days ago

I *do* get the bottom right segment of the diagnostic display lighting up during the memory test portion of leon's rom. Going to try pinside next now that programs are atleast loading.

#15 24 days ago

Ok, pincoder tests 1-4 all run fine except that the 7 segment diagnostic led doesn't light all the way.

@pincoder Got a tip on what I need to run down to figure out what part of that is busted?

#16 24 days ago

If I check IC 34's pins based on this doc: http://www.blackknightpinball.com/sys7-mpu-leds-info.htm

It looks like I'm getting an 8 when I hit the diagnostic button. Seeing as it's on the bench with no driver board, that sounds good? Seems like I just need to replace the display itself?

#17 20 days ago

Progress report: while I have the board on the table I swapped out IC19 with nvram. I’m now getting 0 on the diagnostic if I’m reading the output from the 7447 correctly. The display doesn’t show anything anymore, though before it was only showing segment G (the cross on the 8), so it’d make sense that it shows nothing for 0.

On the assumption the display (MAN72A) is bad, I pulled it out in prep for a new one. But I was curious about why some segments failed so I hooked it up on my 3.3V with a 150ohm resistor and I’m getting every segment lighting up just fine. That makes me wonder if the voltage or resistance is off on the connections.

#18 20 days ago

Welp. It's my 7447. Inputs look good but no outputs.

#19 19 days ago

What inputs are you seeing on the 7447 (pins 1,2,6,7)

When you say "no outputs", you are reading neither high or low on the outputs?

#20 18 days ago

I took the 7447 out in prep for a new one, but I was seeing high on all the inputs (previously was seeing inputs corresponding to 8 according to that doc - inverse binary was a weird choice). Outputs don’t show up either way in my logic probe. If I’m recalling correctly they were in no man’s land for a signal and not enough Vf to light the LEDs.

#21 17 days ago

Replaced the 7447 and now pincoder’s test rom #1 counts up on the display just fine and the black knight flipper rom shows 0 on the diagnostic display.

Main displays are still blank. I checked the voltage and it’s about 90 at 4J6 which is expected since I replaced the Zener diodes when I rebuilt the power supply. Next I’m going to check the BCD inputs by using the pincoder display test (rom 5) unless someone has a better idea. That rom showed nothing when I used it before the 7447 replacement, but I didn’t poke around very long and it looked like the program might not be running. I wasn’t seeing signal changes on IC18’s PB pins which afaict go to bcd outputs, and I’m not sure how the rom could be running without that happening.

#22 16 days ago
Quoted from kevinclark:

Replaced the 7447 and now pincoder’s test rom #1 counts up on the display just fine and the black knight flipper rom shows 0 on the diagnostic display.
Main displays are still blank. I checked the voltage and it’s about 90 at 4J6 which is expected since I replaced the Zener diodes when I rebuilt the power supply. Next I’m going to check the BCD inputs by using the pincoder display test (rom 5) unless someone has a better idea. That rom showed nothing when I used it before the 7447 replacement, but I didn’t poke around very long and it looked like the program might not be running. I wasn’t seeing signal changes on IC18’s PB pins which afaict go to bcd outputs, and I’m not sure how the rom could be running without that happening.

Nice work on finding the faulty 7447. For the display issue, as I'm sure you can see from the schematics, IC6 is responsible for the display strobe outputs. Check pins 20-23 for pulses from PIA1 pins 2,3,4,5. They should be binary counting from 0-15 over and over and that causes IC6 to round-robin the display strobe outputs.

For System 7, The BCD outputs represent a 2 digit pair to be shown for any given strobe output. Player 1 and 3 displays are strobe together, as are player 2 and 4. The same goes for MATCH and CREDITS. So, the upper BCD bits (1J5 pins 1-4) are for Player 1,2 and MATCH, while the lower BCD bits (1J5 pins 5,7,8,9) are for Player 3,4, and CREDITS.

If you are missing activity on one or more of these pins try changing PIA1 (IC18).

If you are seeing activity on the strobe outputs and the BCD outputs, and the BLANKING test rom passes then all of the required signals are being sent to the master display board. If the displays are still blank then you should focus on the master display board.

#23 16 days ago
Quoted from pincoder:

Nice work on finding the faulty 7447. For the display issue, as I'm sure you can see from the schematics, IC6 is responsible for the display strobe outputs. Check pins 20-23 for pulses from PIA1 pins 2,3,4,5. They should be binary counting from 0-15 over and over and that causes IC6 to round-robin the display strobe outputs.
For System 7, The BCD outputs represent a 2 digit pair to be shown for any given strobe output. Player 1 and 3 displays are strobe together, as are player 2 and 4. The same goes for MATCH and CREDITS. So, the upper BCD bits (1J5 pins 1-4) are for Player 1,2 and MATCH, while the lower BCD bits (1J5 pins 5,7,8,9) are for Player 3,4, and CREDITS.
If you are missing activity on one or more of these pins try changing PIA1 (IC18).
If you are seeing activity on the strobe outputs and the BCD outputs, and the BLANKING test rom passes then all of the required signals are being sent to the master display board. If the displays are still blank then you should focus on the master display board.

pincoder Thank you! That's extremely helpful. I was having trouble finding a write up on how the display signaling worked.

I'm seeing signals on the PIA and on IC 6 with my logic probe. No oscilloscope, so I can't check details but I'm seeing movement. The cycling is very clear in IJ5, so I'm going to double check the blanking test rom and the move to the display board.

#24 16 days ago

Confirming that blanking checks out at TP4.

I reinstalled my mpu and driver board and put test rom 5 back in. It came up with a flash of the leds and the diagnostic display which then went blank. But I wasn't able to see any cyling on IJ5. I touched the chips lightly to see if something was loose and when I got to IC 18 the leds and diagnostic came up. Now IJ5 is showing high across the board. TP4 shows low. I'm not seeing pulses from PIA's 2,3,4. Behavior persists across restarts. IC 1 shows pulsing on the address and data bus.

Huh.

#25 16 days ago

I took off the driver board and now I'm getting a response from the rom that's in (which at this point is Leon's). I guess I'll leave it out and debug the displays before figuring out what's wrong with the driver.

#26 16 days ago

And with the displays test rom I'm getting digits cycling! Player 4 display is flakier than I remember, but seems to be doing better as it warms up.

Progress!

#27 16 days ago

I've established that without a driver board, Leon's test rom runs correctly. When I add the driver though, I get no blinking (just lights on and 0 on display). Leon's guide (http://pinwiki.com/wiki/index.php?title=Leon_Borre_WMS_System_7_Repair) seems to indicate that the good PIAs should show signal while the bad PIAs shouldn't, but when the driver board is added I'm seeing no signal on even the MPU PIA outputs. They're locked high on IC 18 and vary, but don't pulse, on IC 36.

Does anyone know what expected behavior should be in this situation?

#28 16 days ago

Have you replaced the 40-pin connector between the driver and mpu boards?

#29 16 days ago
Quoted from kevinclark:

pincoder Thank you! That's extremely helpful. I was having trouble finding a write up on how the display signaling worked.
I'm seeing signals on the PIA and on IC 6 with my logic probe. No oscilloscope, so I can't check details but I'm seeing movement. The cycling is very clear in IJ5, so I'm going to double check the blanking test rom and the move to the display board.

Your welcome. It's nice to get good feedback on these test ROMs

Quoted from kevinclark:

Confirming that blanking checks out at TP4.
I reinstalled my mpu and driver board and put test rom 5 back in. It came up with a flash of the leds and the diagnostic display which then went blank. But I wasn't able to see any cyling on IJ5. I touched the chips lightly to see if something was loose and when I got to IC 18 the leds and diagnostic came up. Now IJ5 is showing high across the board. TP4 shows low. I'm not seeing pulses from PIA's 2,3,4. Behavior persists across restarts. IC 1 shows pulsing on the address and data bus.
Huh.

Sounds like the MPU is crashing altogether when the driver board is attached. That's not a nice one to debug because the cause is tough to nail down. Typically one or more of the PIA and/or logic chips on the driver board is locking up the data and/or address bus. For that I'd suggest using the 01b-bus and 01c-transceivers ROMs. Comparing those results with and without the driver board connected may help. If you have to start pulling chips from the board to eliminate the culprit, remove one PIA at a time from the driver board (desolder, remove, install socket, solder, re-install original chip) and re-run the test. Of course you have to be careful when removing the old chip not to damage it from excessive heat, bending, etc. Using a heat gun to remove the chips works great (practice on some old boards if you have any handy).

As for leon's ROM, I haven't any experience with it so I can't comment on it's behaviour.

#30 16 days ago
Quoted from frunch:

Have you replaced the 40-pin connector between the driver and mpu boards?

frunch I haven’t, but it looks like whoever worked on it last did. I’m seeing extra long mating pins and the coloring doesn’t look old enough to be original. I’ve got parts on hand to replace it if I need, but I suspect, given how I zapped it, I have chip issues on the driver board. Possible connection too though.

#31 16 days ago
Quoted from pincoder:

Typically one or more of the PIA and/or logic chips on the driver board is locking up the data and/or address bus. For that I'd suggest using the 01b-bus and 01c-transceivers ROMs.

pincoder I’ll give those a shot. Other theory I ran into is that a bad PIA can muck with the reset voltage and keep things from moving.

#32 15 days ago
Quoted from kevinclark:

pincoder I’ll give those a shot. Other theory I ran into is that a bad PIA can muck with the reset voltage and keep things from moving.

That's an interesting theory. None of the chips have the ability to pull the reset low (They're input only on the reset pin). It'll be interesting to see if that's the case.

Does the board pass the 01a-leds test with the driver board connected? That test does not initialize the other PIAs and ignores them. the later tests initialize all of them even if they arent being used for the particular test.

#33 15 days ago
Quoted from pincoder:

That's an interesting theory. None of the chips have the ability to pull the reset low (They're input only on the reset pin). It'll be interesting to see if that's the case.

Pulled the idea from here: http://pinwiki.com/wiki/index.php?title=Williams_System_3_-_7#CPU_will_not_start

Kind of just fishing in the dark for ideas to test though.

Quoted from pincoder:

Does the board pass the 01a-leds test with the driver board connected? That test does not initialize the other PIAs and ignores them. the later tests initialize all of them even if they arent being used for the particular test.

pincoder 01a-leds doesn't appear to start when the driver board is connected to the mpu. The lights and segment display flash once and then stay off. I've seen it turn on after I've had it on in that state a little while, but I haven't yet figured out what changes. Just happened again actually. I wasn't touching the machine, it was sitting on (with the MPU completely connected and the driver board with only the 40 pin connected) with my logic probe connected to TPs 10 and 9. I heard a click and now the leds and diagnostic display are on. If I turn it off and back on the display and leds are on immediately, as if I'd filled up a cap that hasn't had time to discharge. I haven't figured out how long it takes to get out of that state, but if I let it sit long enough I've seen it go back to the other behavior (flash and then blank). This kind of behavior is what makes me wonder about a voltage issue.

I'm going to take it back to my bench and poke around with the other roms and my logic probe. Starting to wonder if I should get an oscilloscope so it's easier for me to log and see what's going on.

#34 15 days ago

Two minutes appears to be enough time to get it to go back to LEDs off after the initial flash.

#35 15 days ago

Here's a quick investigation of 01a-leds with and without the driver board. Will run 01b-bus asap. Notable is that the voltage drops pretty significantly once I plug the driver board in. Unclear if that's because of my PSU (a mini ATX with an adapter board) or indicative of something else:

TP voltages with logic probe attached but no driver board, LED diagnostics running:

Input from PSU: 12V nominal 12.9V actual, 5V nominal 4.4V actual
TP2: 3.3V
TP3: 0.08
TP4: 0.01V
TP5: 4.3V
TP6: 2.2V
TP7: 4.4V
TP8: 4.4V
TP9: 4.4V
TP10: 0

2P1 1,2: 4.4V

Address and Data buffers on 2P1 pulse. Reset pin (19) at 4.3V. Blanking pin (37) at 0.01V.

With driver board connected and program apparently not running and LEDs locked on and 0 on diag display:

Input from PSU: 12V nominal 13V actual, 5V nominal 3.9V actual

TP2: 2.7V
TP3: 0.08
TP4: 0.14V
TP5: 3.7V
TP6: 1.9V
TP7: 3.8V
TP8: 3.7V
TP9: 3.8V
TP10: 0

2P1 1,2: 3.7V

Address and data buffers pulse. Reset pin (19) at 3.7V. Blanking pin (37) at 0.14V

#36 15 days ago

@pincoder Here's results of 01b-bus with driver board connected. Will collect without connection when I next have time. Looks like the driver board might be keeping it from finding the rom?

Differences from stated system 6 6808 results (though I've got a system 7):

5 VMA: HIGH
7 NC: low - 0.01V
9-20 A0-A11: pulses
22-33 A12-15, D7-0: pulses
34: HIGH
36: low
38: pulses
39: pulses
40: HIGH

#37 14 days ago

Re: reset signal: That's an interesting case you pointed out (where the PIA dragged the reset signal down). Apparently it IS possible (with a bad PIA). I was referring to the design of the board.. The only component that should have the ability to pull reset low is the reset circuit and normally functioning chips on this board don't have the capacity to do so on their own (their reset pins are input only). Interesting that if a chip is fried in just the right way it can have the ability to drag it low on its own. You learn something new everyday!

It's worth mentioning that upon normal power up, the LEDs do flash and the zero shows up on the LED segment display. This occurs before the CPU starts executing any instructions. The first bit of code in any ROM should be to initialize the PIAs and that's when you see the LEDs and segment go out. The zero means nothing really. It is not the result of some test being performed in any of the pincoder ROMs, and I would also suspect is true for the Williams ROMs. It's just how IC33 and IC34 power up in this circuit.

Re: two minutes: Just to be clear, when you say "display" you are referring to the segment display on the MPU board, not any of the master or slave displays behind the backglass, correct?

From your test results above, it looks like your reset signal is high, but is it a STEADY high? Make sure your logic probe is set to PULSE and connect it a few seconds after powerup.

You should not be seeing activity on the address and data bus lines when using using 01b-bus (at least not after the first few microseconds after the reset goes high). Since you are, I'd say the CPU is constantly resetting and trying to run the ROM. Once it successfully runs the 01b-bus ROM the CPU goes into a wait state (not executing any more code) so the pins on the CPU (including address and data lines) should go to their states shown in the 01b-bus documentation (basically idle).

So, your board is either continually resetting because of:

1) Insufficient power
2) The reset circuit itself
3) A PIA is in fact pulling the reset circuit low.

In the case of 1) I have used a cheap 400w ATX power supply on the bench in the past. You should be fine. You should connect as many of the GND, 12V, and 5V pins of 1J2 as you can for more current flow. When you connect the driver board you should also supply power to 2J8 in the same fashion.

In the case of 2) the reset circuit is taking too long to finish its job (start low, wait ~0.5 seconds, go high)

In the case of 3) it could be possible that as the (bad) PIA warms up it lets go of the reset line, finally allowing the CPU to start executing code.

I'm not sure what the click you're hearing is. Hold your finger on the flipper relay after you power it up and see if that's it. With 01b-bus nothing is initialzed at boot - it does one READ from address 0000 and then goes into the wait state. This means that the flipper relay should do nothing at all under this test.

Edit: I've updated the documentation for the 01b-bus test so if this post still leaves you a little unclear download the beta release and review the document:

http://pincoder.ca/ccn/roms/pincoder_roms_2019.10.03.1703_beta.zip

#38 14 days ago
Quoted from pincoder:

It's worth mentioning that upon normal power up, the LEDs do flash and the zero shows up on the LED segment display. This occurs before the CPU starts executing any instructions. The first bit of code in any ROM should be to initialize the PIAs and that's when you see the LEDs and segment go out. The zero means nothing really. It is not the result of some test being performed in any of the pincoder ROMs, and I would also suspect is true for the Williams ROMs. It's just how IC33 and IC34 power up in this circuit.

With the Williams ROMs, the number will stop on the test or initialization that fails:

0 Test Passed (only if display goes blank)
1 IC13 RAM Faulty
2 IC16 RAM Faulty
3 IC17 ROM 2 Faulty
4 IC17 ROM 2 Faulty
5 IC20 ROM I Faulty
6 IC14 Game ROM 1 Faulty
7 IC26 Game ROM 0 Faulty
8 IC19 CMOS RAM or memory protect circuit faulty
9 Coin-door closed, memory protect circuit faulty, or IC19 CMOS RAM Faulty

If 0 is left on the display, the CPU is not running. If 0 flashes for an instant and the display goes blank, it is running.

#39 14 days ago
Quoted from Schwaggs:

With the Williams ROMs, the number will stop on the test or initialization that fails:
0 Test Passed (only if display goes blank)
1 IC13 RAM Faulty
2 IC16 RAM Faulty
3 IC17 ROM 2 Faulty
4 IC17 ROM 2 Faulty
5 IC20 ROM I Faulty
6 IC14 Game ROM 1 Faulty
7 IC26 Game ROM 0 Faulty
8 IC19 CMOS RAM or memory protect circuit faulty
9 Coin-door closed, memory protect circuit faulty, or IC19 CMOS RAM Faulty
If 0 is left on the display, the CPU is not running. If 0 flashes for and instant and the display goes blank, it is running.

Yes, if the CPU is able to boot the williams ROM it will then run a williams test and display any faults that it finds. However, on-board system 3-7 diagnostic tests are not very reliable.

#40 12 days ago
Quoted from pincoder:

Re: two minutes: Just to be clear, when you say "display" you are referring to the segment display on the MPU board, not any of the master or slave displays behind the backglass, correct?

pincoder That's correct.

Quoted from pincoder:

From your test results above, it looks like your reset signal is high, but is it a STEADY high? Make sure your logic probe is set to PULSE and connect it a few seconds after powerup.

I don't have an option to pulse or not pulse, but the pulse light is on and while I'm seeing pulsing on other pins (address and data bus for example) I'm not seeing it on reset (either IC 1 pin 40 or TP 8).

Quoted from pincoder:

So, your board is either continually resetting because of:
1) Insufficient power
2) The reset circuit itself
3) A PIA is in fact pulling the reset circuit low.
In the case of 1) I have used a cheap 400w ATX power supply on the bench in the past. You should be fine. You should connect as many of the GND, 12V, and 5V pins of 1J2 as you can for more current flow. When you connect the driver board you should also supply power to 2J8 in the same fashion.

My ATX supply might be a special case - I pulled it out of a NAS and it's made specifically for it but has an ATX connector and seems to give normal voltages (though I'm missing -12V). I've got power applied to 2J8 similarly and I see no change. I'm seeing about 4V actual today from 5V nominal with the driver board connected. Counter argument to the power issue is that I see the same behavior when connected with the actual power board in the backbox and since I just replaced the caps and I've seen it give proper voltages, I can't imagine it would be insufficient power. I'm going to call 1 not very likely for now.

Quoted from pincoder:

In the case of 2) the reset circuit is taking too long to finish its job (start low, wait ~0.5 seconds, go high)

If I flip power on while my logic probe is on reset I see it start low and almost immediately jump high.

Quoted from pincoder:

In the case of 3) it could be possible that as the (bad) PIA warms up it lets go of the reset line, finally allowing the CPU to start executing code.

I've got 3 additional PIAs on hand and two of the chips on the driver board are already socketed, so I'm going to swap a couple out and see if that makes a difference.

#41 12 days ago

Reset looks fine on pin 40 of the solenoid matrix PIA but low on the switch and lamp matrix PIAs, so I went to swap those out. I pulled out IC 11 on the driver board and decided to test it by plugging it into the known working MPU (without driver board) in place of the good PIA in IC 18. MPU no longer booted. I swapped in the good PIA just to make sure things were still working and get no lights. TP 9 and 10 were measuring half a volt. My 12V contact on the power supply was only measuring ~7V. I turned it off and reseated the good IC 18. Still no lights. I'm seeing 5V on pin 1 of MPU IC 1.

Did the bad PIA short something?

#42 12 days ago

Huh. Moved my ground connection from 1J2-1 to 1J2-2 and now it’s booting again.

Edit: Looks like those headers need to be replaced. I'm seeing cracks underneath.

#43 12 days ago

pincoder IC11 is swapped and 01a-leds is now running with the driver board attached.

Dunno if it was the reset getting pulled low, but the PIA on the driver board was keeping the CPU in a restart loop.

#44 11 days ago

Plugged my boards back in and I'm able to get to audit mode! Digits test runs fine. That's the good.

The bad is:

* The big resistors on the bottom right of the driver board are running crazy hot and seem to have burnt the board. I can smell it when they turn on. I've got IRF9Z34Ns on hand and am going to replace the TIP42s and resistors, but seems like there might be an underlying issue there.

* Sound test has no response, though sound diagnostic button works fine.

* Solenoid test has no response. Test 25 or coming back to 1 makes a clicking sound like a relay flipped. Maybe IC 5 PIA bad too.

* Switch test replies with 38 37 35 34 33. So column 5 seems like it's having some issues, but not across the board.

* Lamp test has no response. Thinking IC 10 PIA is bad.

#45 11 days ago

Disconnected the driver board and now I'm not getting a boot again (*facepalm*). Reset was low. I jumpered it against 5V and then it boots and runs 01a-led.

#46 10 days ago

Today, after replacing the TIP42s, reset seems to be fine.

#47 10 days ago

Oh, and IC10 PIA tested fine in the IC18 spot last night (after I jumpered reset), so that theory for the lamps is out. But it seems possible the transistors I just swapped out had fried and were the issue. Could explain the resistor issue too. Will test later.

#48 10 days ago
Quoted from kevinclark:

Plugged my boards back in and I'm able to get to audit mode! Digits test runs fine. That's the good.
The bad is:
* The big resistors on the bottom right of the driver board are running crazy hot and seem to have burnt the board. I can smell it when they turn on. I've got IRF9Z34Ns on hand and am going to replace the TIP42s and resistors, but seems like there might be an underlying issue there.
* Sound test has no response, though sound diagnostic button works fine.
* Solenoid test has no response. Test 25 or coming back to 1 makes a clicking sound like a relay flipped. Maybe IC 5 PIA bad too.
* Switch test replies with 38 37 35 34 33. So column 5 seems like it's having some issues, but not across the board.
* Lamp test has no response. Thinking IC 10 PIA is bad.

Those resistors do run crazy hot, almost 200F when everything is working properly. It gets worse if you still have incandescent bulbs in the machine. The MOSFETs will drop that down to a little over ambient.

I lost track with what you have done. Have you replaced the 40 pin, ROM and CPU sockets, header pins around the boards? Those are typical problem spots that I always replace before killing myself chasing other symptoms.

#49 10 days ago

There are quite a few intermittent problems here. I agree with schwaggs that the connectors, solder joints etc all be addressed first. ROM's can't detect stuff like that so good lighting and a great magnifying glass are good to have.

In the latest beta release of the pincoder ROMs there is 11b-sound. It's similar to the existing 11-sound ROM (now called 11a-sound) except that it triggers each sound three times a few seconds apart all on its own. So you don't need to touch anything except the power switch to turn it on. This allows you to test without touching the machine, to see if it can run steady on it's own (because you're not moving wires or pushing buttons etc).

You could run this ROM and if all goes as expected, start gently pushing or wiggling chips/connectors/wires around where you think you may have a connection issue. Could be of some use here if you are unable to visually find problem solder joints etc.

#50 9 days ago

schwaggs pincoder I very much appreciate the advice! I wouldn't have gotten this far without your help. Virtually everything was working before I shorted it out, so it's a little sad I'm going back to basics but not that unexpected.

Just to review, here's what's been done so far:

* Power board has had caps and resistors replaced per Vid's bulletproofing guide. This happened before the short.
* On MPU: 7447 replaced (IC 34), PIA I replaced (IC 18), CMOS RAM replaced with NVRAM, IJ2 reflowed as a stopgap until header replaced.
* On Driver: PIA II (IC 11) replaced, TIP 42s replaced with IRF9Z34Ns and resistors replaced with jumpers in the lamp column drive section.
* PIA IV tests good when placed in MPU's IC 18 socket. PIA II is not socketed and hasn't been removed for testing yet.

The headers appear to have all been replaced whoever cleaned it up before it wound up at my parents' house. Use since then was very very light, likely in the low 10s of hours. But the solder job doesn't look great. Downright ugly in some places. I'm wondering if desoldering and re-soldering would be as useful as replacing the headers. I've got a hakko 301, so it's easy to do and the hardware doesn't appear worn or burnt. Though it isn't the square pins. If anyone has a strong opinion that I should replace the hardware entirely, I'd appreciate hearing it. If it was original I'd definitely do it, but my gut says this is mostly fine except for the solder job.

If I'm replacing headers, these are all .156"?

The *connectors* all appear to be original IDC. I wouldn't be surprised if they've gotten a bit banged up in this shuffle and expect I might have to replace them. pinrepair.com has a list of tools for handling connectors but it seems a little bloated and I can't figure out which things are actually needed and which are best bang for the buck. I don't mind spending a little more to get a good crimp that I don't have to find a re-work later. GPE's site seems to advise away from grabbing tools for ejecting pins which makes me question the pinrepair recs entirely. Would very much appreciate advice here.

pincoder I'm going to spend some time working through the ROMs and will make sure to take a look at 11b and the beta release. Thank you thank you! I'm a coder by trade and have done a fair amount of low level stuff. Let me know if there's anything I can do to help in addition to testing out releases. At this point I've used so many of your ROMs that I'm considering swapping out IC 17's socket with a ZIF so I stop mangling pins. I'm really ready for that adapter board to exist so I can put it all on one chip.

Promoted items from the Pinside Marketplace
$ 20.00
Electronics
Yorktown Parts and Equip
From: $ 18.00
Apparel - Men
Pinside Shop
There are 80 posts in this topic. You are on page 1 of 2.

Hey there! Got a moment?

Great to see you're enjoying Pinside! Did you know Pinside is able to run thanks to donations from our visitors? Please donate to Pinside, support the site and get anext to your username to show for it! Donate to Pinside