3rd Feb 2021, 10:10 pm | #441 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,817
|
Re: Non-working Commodore PET 3016
OK. Extra hand press-ganged into action.
Note I had to replace the 6502 as the original was starting to give intermittent results. So these test which look a lot cleaner to me have a replacement 6502, UB3 and UC3 ICs. https://drive.google.com/file/d/1jOS...ew?usp=sharing |
3rd Feb 2021, 10:32 pm | #442 |
Octode
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,364
|
Re: Non-working Commodore PET 3016
Looking spot on actually - it appears those buffers were faulty...
|
3rd Feb 2021, 11:09 pm | #443 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,817
|
Re: Non-working Commodore PET 3016
Progress then? Felt like it to me when I saw the scope output.
Last edited by ScottishColin; 3rd Feb 2021 at 11:27 pm. |
3rd Feb 2021, 11:59 pm | #444 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,577
|
Re: Non-working Commodore PET 3016
Absolutely. I take it that was all with the replacement CPU with all pins connected, the replacement buffers fitted and UD7 still out of circuit? Does insertion of UD7 still stop the CPU from running, as indicated by the presence or absence of activity on A0 (CPU pin 9)?
PET experts: What does the UD7 PROM contain? |
4th Feb 2021, 12:28 am | #445 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,817
|
Re: Non-working Commodore PET 3016
I only had the selection of pins in the socket as before. I'll try it with all pins tomorrow to see what I get.
And no UD7. I'll try that back in tomorrow to see what results I get. I found this re UD7: https://www.nightfallcrew.com/03/11/...onitor-repair/ |
4th Feb 2021, 12:50 am | #446 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,577
|
Re: Non-working Commodore PET 3016
It's interesting that he mentioned that a duff UD7 gave a no picture fault, although it's not clear whether he meant a blank raster or no life from the monitor at all.
It's a shame he didn't mention the actual EPROM device he used as a replacement - I don't know if UD7 is 4K or 8K - 8K eproms in a 24 pin package are quite uncommon, I think the Motorola MCM68764 is one of the types which are sometimes used to directly replace them, certainly in the VIC-20. |
4th Feb 2021, 1:20 am | #447 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,577
|
Re: Non-working Commodore PET 3016
I just found the file which corresponds to that PROM (numbered 901465-02) and it looks to be a 4K PROM.
Unfortunately the more common 24-pin 4K 2732 EPROM is not a good pinout match for the PROM with several pins in the wrong place. The harder to find 8K MCM68764 is almost a perfect match, the only anomaly being that pin 21, CE3 on the original PROM and tied to +5V on the PET diagram, is A12 on the MCM68764. All that means is that you would program the 4K of code into the upper half of the 8K MCM68764 - with pin 21 (A12) held high, the upper half of the device would be permanently selected - so I'm thinking that is the device which is probably used as a drop in replacement when these need to be replaced. However, we are not quite at the point of needing a replacement yet. |
4th Feb 2021, 12:15 pm | #448 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,817
|
Re: Non-working Commodore PET 3016
OK. Here's what I get this morning from pin 9 of the new 6502.
New 6502 with all sockets connected and all ROM chips inserted - I get a momentary reading at power on on the scope then lose the reading. New 6502 with all sockets connected and only UD6 removed (note not UD7) - I get a reading. If all ROMs are inserted, I can get a reading if pins 30,31 and 32 are removed from the 6502 socket. So the ROM situation isn't staying still - the only thing that is constant is the requirement for pins 30,31 and 32 to be removed from the new 6502 if I want a reading with all ROMs inserted. |
4th Feb 2021, 12:38 pm | #449 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,577
|
Re: Non-working Commodore PET 3016
The above sounds like the 6502 is reading one of the 'illegal' KIL (halt) instructions from the databus, but only when UD7 is in place. by disconnecting some of the databus pins you are altering what the 6502 reads from the data bus so that it no longer looks like a KIL instruction.
My understanding is that UD7 is one of the two BASIC 2 PROMs. Try this: Put UD7 in your breadboard, remembering to position it on the centre line with the pins on either side of the centre channel, and with nothing else plugged into the breadboard. Make resistance measurements from the PROM IC's pin 12 and pins 8,7,6,5,4,3,2,1,23,22,19,18 (black lead on 12) and from the PROM IC's pin 24 to the same pins, (red lead on pin 24). All resistances from 12 to those pins should be about the same, all resistances from pin 24 to those pins should be a different resistance but again all similar to each other. If that's the case try the resistances from pin 12 to pins 9,10,11...13,14,15,16,17 and from pin 24 to the same pins. |
4th Feb 2021, 12:39 pm | #450 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,577
|
Re: Non-working Commodore PET 3016
Do you have any of the common microprocessor toys, like an Arduino, available to you?
|
4th Feb 2021, 1:51 pm | #451 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,577
|
Re: Non-working Commodore PET 3016
Just re-read your #448. I think we are getting close to the point where we need to be able to verify the PROM contents (but you can do the same checks on the UD6 IC as I suggested for UD7).
Bad data bus buffers are another possible underlying cause for this. |
4th Feb 2021, 2:12 pm | #452 |
Pentode
Join Date: Jul 2014
Location: Isle of Man, UK.
Posts: 102
|
Re: Non-working Commodore PET 3016
Jebus reading all these posts about this PET give me nightmares. I have a working CBM 4032 and I am terrified it will fail one day(which it will).
Hope you can get it fixed. I have saved all the comments on here for posterity. If you need me to take any readings I can do so on mine, if that helps. |
4th Feb 2021, 2:27 pm | #453 | |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,817
|
Re: Non-working Commodore PET 3016
Quote:
In all seriousness, I'm learning a lot though from the 'teachers' on here. |
|
4th Feb 2021, 3:47 pm | #454 | |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,817
|
Re: Non-working Commodore PET 3016
Quote:
UD7 - red probe on pin 24, meter set to 200k 1 - 121.5 2 - 120.8 3 - 122.5 4 - 95.8 5 - 131.9 6 - 155.5 7 - 90.5 8 - 91.9 18 - 91.8 19 - 114.6 22 - 118.7 23 - 120.5 I stopped there on UD7 as the readings were so varied. Note also that the readings on each pin varied wildly from 90-160. I took the readings above as and when they stabilised. UD6 - no readings from black probe on pin 12 to any other pin on any resistance setting on the meter UD6 - red probe on pin 24, meter set to 200k 1 - 26.7 2 - 26.7 3 - 26.8 4 - 26.5 5 - 26.4 6 - 26.5 7 - 26.4 8 - 26.3 18 - 26.2 19 - 26.4 22 - 26.6 23 - 26.7 9 - 16.6 10 - 16.6 11 - 16.7 13 - 16.6 14 - 16.7 15 - 16.7 16 - 16.6 17 - 16.6 Colin. |
|
4th Feb 2021, 3:48 pm | #455 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,817
|
Re: Non-working Commodore PET 3016
|
4th Feb 2021, 4:20 pm | #456 |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,297
|
Re: Non-working Commodore PET 3016
It may be worth testing the data bus buffer using the same method as the address buffers. One probe on each side of the buffer to compare input and output, which will change for read and write to memory.
|
4th Feb 2021, 4:27 pm | #457 | |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,577
|
Re: Non-working Commodore PET 3016
UD6 looks so good it could win a beauty contest, but I'm not at all sure about UD7, given that they are probably exactly the same type of PROM device so we would expect that they would both give very similar results.
Quote:
I think I probably know the answer, but does that have 3.3V logic in / out or 5.0V in/ out? If it's 5V friendly you could write a PROM reader program - if it's 3V friendly you could still do that but you would need some level-shifting hardware (To change 3V levels to 5V and vice versa) in between the PROM and the Pico. The classic old Arduinos like the Uno are 5V devices so you can connect them directly to 5V devices like your PROM. (Which is why I asked about Arduinos). |
|
4th Feb 2021, 4:34 pm | #458 | |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,577
|
Re: Non-working Commodore PET 3016
Quote:
Will come up with some pin number pairs to check when I get home, if Mark doesn't beat me to it. Or maybe Colin can already work out which pairs of pins we would want compared (on the data bus buffers, UE9, UE10). |
|
4th Feb 2021, 4:54 pm | #459 |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,297
|
Re: Non-working Commodore PET 3016
I’m hoping we can tell which direction the buffer is driving by looking at the signals on both sides of the buffer. We should at least be able to see correct logic levels on both sides.
|
4th Feb 2021, 5:21 pm | #460 | ||
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,817
|
Re: Non-working Commodore PET 3016
Quote:
The pi (I have a few of them lying around )can in theory output 5v, but it's actually 5v minus the power the pi is drawing. Would it be easier all round if I bought an arduino board? And if so, which one would you recommend? Colin. |
||