• Please review our updated Terms and Rules here

Dead PET 8032 - any suggestions?

gubbish said:
One more thing, and so far the only easily quantifiable difference in the signals I've observed at the CRTC between good/bad PET 8032s is the clock phase at pin 21 of the CRTC chip. In the working PET, the clock signal at 21 has the opposite phase from the signal at pin 23 (E, which I believe is an enable signal for reads/writes to/from the CRTC registers). In the bad PET, pin 21 has the same phase as the signal at 23. In fact what I'm seeing at pin 23 on the good PET doesn't match with the schematic I'm looking at. In the schematic it says that that pin 23 is connected to BA2 (buffered address 2) - on the good PET these pins are not connected, and in fact pin 23 is connected to /CLK1, inverted from the main clock signal by the inverter at UE4 pin 10.
Click to expand...
I suspect that there's a misprint in that schematic (assuming you're talking about 8032081-10), but I also think you might have miscounted the pins.

Both schematics show /CLK1 connected to pin 21 (not pin 23) of the CRTC.

The old schematic indeed says that pin 23 is connected to BA2, whereas the later version (8032087-10) shows it connected to Bclk2 (UD15-2), which makes more sense. On the other hand the old schematic does not show Bclk2 connected to the CRTC, so it would be interesting to know where its pin 23 is really getting its signal on that board and perhaps explain the phase difference. Any jumper wires by any chance where someone may have connected it to the wrong place? Did this PET ever work while you've had it?

I assume that you're getting no horizontal or vertical sync signals on pins 39 & 40?
 
Last edited:
You're right about miscounting, I'm having a bit of "pin fatigue" so the pin numbers are starting to jumble together..
I have the old schematic that indicates BA2 is connected to 23. It was pin 21 that was actually connected to /CLK1 on UE4 pin 10, on the good pet 8032. Pin 23 appeared to have the same signal as non-inverted CLK1. I checked continuity between pin 23 and phi-0, clock pin 37 on the 6502, and saw a low but non-zero resistance between them.

On the bad PET, pin 21 has zero resistance with pin 37 on the CPU, so it appears to be directly connected to CLK1. I haven't been able to trace pin 23 yet, but it appears to have the same signal as CLK1. It does not appear to be connected to BA2, UD13 pin 7.
There are jumpers on the non-working board. Suspiciously enough, one jumper wire is connected to pin 10 on UE4 - where /CLK1 lives. I did confirm that the connections made by the jumpers matched what I saw on the schematic, but this may be wrong because I've been looking at the old schematic. This is a 8032030 board.
No horizontal or vertical sync from 39 and 40 on the CRTC. Actually no sign of activity at all from the CRTC. Data is presented to it and the chip select is happening, but nothing on the address pins from the CRTC, no sync or anything else. That's what is leading me to suspect that chip.

This PET has never worked while I've had it. It was one of the PETs given away by the old computer store near Berkeley.
 
gubbish said:
In the bad PET, pin 21 has the same phase as the signal at 23.
Click to expand...

Your bad PET is the old 8032030 type, is it not? While the working one is the much newer universal 8032090 type.

There are typos in the schematic of the 030 as Mike mentioned. Pin 23 says BA2 but must be the B°2 Clock while pin 21 says CLK1 which would explain why the phases are the same, but I think it really should be /CLK1 as it is in the 8032087 schematic:

http://www.zimmers.net/anonftp/pub/cbm/schematics/computers/pet/univ2/8032087-10.gif

The fact that you say the two clocks are in phase in the broken computer is weird as one (pin 23) is for CPU read/write and the other (pin21) is for video refresh as you say. As Mike mentioned, check to see if someone was fooled by the typos and rewired the board with the wrong phases.
-Dave
 
Last edited:
gubbish said:
...No horizontal or vertical sync from 39 and 40 on the CRTC. Actually no sign of activity at all from the CRTC. Data is presented to it and the chip select is happening, but nothing on the address pins from the CRTC, no sync or anything else. That's what is leading me to suspect that chip.
Click to expand...
Well, it wouldn't hurt to check the source of the signal on pin 23; I would expect the CRTC signals to be more or less the same between the two board versions, and if it's really the wrong signal that could well prevent the CRTC from doing its thing.
This PET has never worked while I've had it. It was one of the PETs given away by the old computer store near Berkeley.
Click to expand...
Ah, that always makes things interesting, as Eudimorphodon can attest... ;-)
 
MikeS said:
Well, it wouldn't hurt to check the source of the signal on pin 23; I would expect the CRTC signals to be more or less the same between the two board versions, and if it's really the wrong signal that could well prevent the CRTC from doing its thing.
Click to expand...

Mike,
Yes, you are right, if somehow the clocks were modified with them being on the wrong phase, you could not correctly load the CRTC registers to get it kicked off.
 
gubbish said:
...I checked continuity between pin 23 and phi-0, clock pin 37 on the 6502, and saw a low but non-zero resistance between them.

On the bad PET, pin 21 has zero resistance with pin 37 on the CPU, so it appears to be directly connected to CLK1. I haven't been able to trace pin 23 yet, but it appears to have the same signal as CLK1. It does not appear to be connected to BA2, UD13 pin 7.
Click to expand...
I don't think any of these signals should go directly to the CPU, and pin 21 should be inverted with respect to the CPU's pin 37. Definitely sounds like a few signals crossed there somewhere; maybe check that out before you remove/replace any parts.
There are jumpers on the non-working board. Suspiciously enough, one jumper wire is connected to pin 10 on UE4 - where /CLK1 lives. I did confirm that the connections made by the jumpers matched what I saw on the schematic, but this may be wrong because I've been looking at the old schematic. This is a 8032030 board.
Click to expand...
Sounds like you're looking at the right set (8032081) but you might want to look at the relevant pages in the 8032087 set for comparison.
 
MikeS said:
Sounds like you're looking at the right set (8032081)
Click to expand...

Mike,
Now I'm getting confused as to which printed circuit boards gub has. I thought he mentioned the bad board had a top assembly of 8032030 which would be the first type of 8032 made, and the working one had a printed circuit assembly of 8032089 which must mean it has to have the very latest top assembly number of 8032090. So the schematics we should be looking at are the 8032029 and 8032087. Why is everything with Commodore so confusing? :)
-Dave
 
dave_m said:
Mike,
Now I'm getting confused as to which printed circuit boards gub has. I thought he mentioned the bad board had a top assembly of 8032030 which would be the first type of 8032 made, and the working one had a printed circuit assembly of 8032089 which must mean it has to have the very latest top assembly number of 8032090. So the schematics we should be looking at are the 8032029 and 8032087. Why is everything with Commodore so confusing? :)
-Dave
Click to expand...
You're right again as always of course, Dave; the CRTC section is identical right down to the misprint, but we should be looking at 8032029, not 8032081, and comparing to 8032087.
 
Mike,
Isn't this an interesting problem... I think it can only happen if some ham-fisted tech got a hold of the board in its distant past. But who knows...
-Dave
 
Hi again PET friends,

Well, finally I have some solid progress to report!
After much pacing and pondering, I decided to take the scary action of de-soldering the 6545 from the working 8032, and putting a socket in. It was either this, or wait several days while a replacement 6545 made its way across the country.. and I was too eager to try out my bad 6545 theory to wait that long.
So, after much hunching over of boards and desoldering, I had a liberated 6545 chip and a bit of back pain.
To my relief, I first put the 6545 back in the working board, and it started up just fine.
Then I replaced it with the hitachi HD46505SP from the bad board - and it did not start. No beep, no video.
So the 46505 was looking like at least part of the culprit. Put the 6545 and screen ROM in the bad board, power up - and LIFE!
Well, there are still some issues, but this is definitely a big step.
There is what looks like a basic screen and a big grid of '0' characters. The keyboard responds properly, but it seems like every other character typed comes out incorrect.
Note that this is in the board with the clock phase issue we discussed earlier, so that may have something to do with the screen corruption.
Taking a break from board fixing for a bit to recharge. Thanks again for all the help, and I'll post the picture of the current state of the computer in a second.
 
dave_m said:
Mike,
Isn't this an interesting problem... I think it can only happen if some ham-fisted tech got a hold of the board in its distant past. But who knows...
-Dave
Click to expand...
After hearing about Eudimorphodon's baby maybe some of these computers were used to train and test aspiring technicians, and these are from the ones who flunked... ;-)
 
Well done on progress!

I'm not sure what kind of video RAM your 8032 uses but weird characters on the screen but good keyboard response suggests a problem somewhere in the video decoding maybe video RAM? Does the 8032 use 2114s? I had to replace BOTH those chips in a 3032 board I had. The character generator IC is another possibility. Or a line getting stuck in one of the logic chips associated with that part of the circuitry.

Tez
 
Here are some pictures of the 8032 screen.
http://www.flickr.com/photos/25219890@N06/sets/72157627049451315/
Grid of 0's, some text corruption. It seems like in the columns without the 0s, typed characters show up correctly - in the 0 columns they are corrupted. But corrupted in the same way each time - for instance I typed several 'i' characters. On the clear columns, they showed up as 'i', on the 0 columns as 'y' every time.
I think the clock phase thing is likely to be at least partially at fault. Will look into that next.
 
tezza said:
Well done on progress!

I'm not sure what kind of video RAM your 8032 uses but weird characters on the screen but good keyboard response suggests a problem somewhere in the video decoding maybe video RAM? Does the 8032 use 2114s? I had to replace BOTH those chips in a 3032 board I had. The character generator IC is another possibility. Or a line getting stuck in one of the logic chips associated with that part of the circuitry.

Tez
Click to expand...

This board does use 2114s, looks like 4 in total. I happen to have 2 extra 2114s from a previous PET repair which was a more or less straightforward video ram replacement. Will check into this during my next PET repair session.
 
gubbish said:
This board does use 2114s, looks like 4 in total. I happen to have 2 extra 2114s from a previous PET repair which was a more or less straightforward video ram replacement. Will check into this during my next PET repair session.
Click to expand...
Looks like a 2114 all right, specifically bit 0 of UC7, pin 14 stuck high. Try piggybacking a good one.

Looks like that phase business was a red herring...
 
Last edited:
MikeS said:
Looks like a 2114 all right, specifically bit 0 of UC7, pin 14 stuck high.
Click to expand...

OK, pin 14 stuck high makes sense as a space character (ASCII 20H) is being shown as a zero (ASCII 30H), but it looks like only the even bytes are messed up, so wouldn't the bad RAM chip be C5?


Looks like that phase business was a red herring...
Click to expand...
Yes, and I am pissed. It was one of the best 'out-of-the-box' theories we ever came up with. I still can not come to grips with those two clock inputs to the 6545 having the same phase. It just should not work in the PET design. I want a full investigation. I am calling the IEEE police! :)
 
Just tried piggybacking a working 2114 chip to see if it cleared up the corrupted characters.
Tried first on UC7, and there was no change.
UC5 did the trick! Screen looks normal, keyboard types the right characters. I'll send some pictures in a bit.
So, looks like if I can find a replacement 6545 chip, and burn a new EPROM to replace the bad screen ROM, then this PET lives again!
Although I have been testing using the power supply and monitor from the working PET. I'll have to confirm that the keyboard/monitor from the previously broken PET work ok.
Thanks again for the help. I wouldn't have known where to start without it!
 
Removed the bad 2114 from the board, and replaced with a good one. Seems to be working well.
Also tested with the original power supply and monitor and those also work.
I ordered a few parts from this place,
http://www.unicornelectronics.com/index.htm
including a few 6545s, EEPROMS, and other parts that looked interesting or potentially useful. This place seems to have a lot of parts used in PETs so check them out if you need replacements. They do have a $25 minimum order so I had to pad out my order with a few things.
Also I will have a couple of extra 6545s after this, most likely, so if anyone needs one just let me know.
Thanks again!
 
You must log in or register to reply here.
Back
Top