View Full Version : Commodore PET 2001 problems...

November 26th, 2009, 09:43 AM
After wanting one for ages, I've finally picked up a Commodore PET 2001 :cool:

Unfortunately, it doesn't work...

The previous owner said that it was working a couple of years ago, but he then put it in his garage... and now it boots weirdly.

I do get characters on the screen, but only one single line in the centre.

Also, if I switch the PET on, the screen will flicker and then go off. If it quickly flick it ON-OFF-ON it will stay on.

Any ideas? I have tried removing/replacing all socketed chips, but no change.

Can anyone tell me the voltage pinouts on the PSU (Euro 240V).

Also, can anyone tell me the exact model of PET? It's lost it's spec label on the back. It's also badges as as "CBM 2001" which I haven't seen before.





November 26th, 2009, 10:14 AM
Hm, you get proper characters on screen. I would check the Basic ROM chips to start with, but perhaps also the Kernal. The problem is those are 6540 chips, each one different. Somewhere I may still have a couple of 6540 chips, but I would need to look up which revisions each one are. If you can not obtain correct chips, Jim Brain has developed a 6540 adapter.


List of chip numbers:
http://zimmers.net/anonftp/pub/cbm/firmware/computers/pet/PET-parts.txt (http://zimmers.net/anonftp/pub/cbm/firmware/computers/pet/PET-parts.txt)

Firmware ROM files:

November 26th, 2009, 07:25 PM
Boy, I sure hate to disagree with Anders of all people, but that sure looks like a 16K dynamic RAM PET (AKA 3016?) using 3x2332 and 1x2316 ROM chips which can be replaced with 2532 and 2716 EPROMs if necessary; Jim's 6540 adapter will not work (and isn't needed) in that board.

Am I right, Anders?

An interesting symptom though; can't say I've ever seen anything like it. At a fast look I thought it was a vertical drive problem, but it really is a single line of text so it probably is indeed a logic problem.

The usual first step is to carefully slightly pry up all socketed chips and push them back in firmly; same with the connectors.

What kind of equipment do you have? If you have an EPROM burner or some other way of reading the EPROMs then the second step would be to read them and compare to known good images.

If that doesn't find anything then we'll actually have to think about the logic and try to figure out what might cause that symptom.

Here's hoping it's something simple! Good luck!


nige the hippy
November 27th, 2009, 01:07 AM
Can't remember if you have a scope, but is the reset line doing it's thing? (a storage scope helps!)

Just as a dampener on things, a sharp mz80-k I have has got 3 separate unrelated faults so far, and I don't think it's been messed with.

1) usual psu caps
2) a stuck ttl output
3) a memory fault

I think time & cosmic rays might have something to do with it!

November 27th, 2009, 01:46 AM
No problem, I'm a layman in terms of PET who just happened to come across a lot of hardware during the past few years. I based my assumption on that newer boards may or may not fit into the 2001 with small keyboard. At least I have learned the power supply has different connectors so the PSU from a 2001N/3000/4000/8000 machine can not be used on an original 2001. We tried that earlier this year, and even went as far as looking up schematics to figure out how to rewire the power connector.

Then again, Mike may be perfectly right above, that it really is an upgraded (?) 2001N board inside which makes it far easier to install new ROMs. The suggestion to try and dump the ROMs is a good one. I did something similar before when I sent a few chips to another fellow on the forum: first pry chips from spare parts boards, then read them with my EPROM programmer and compare to downloaded ROM dumps.

Most of the links I gave still hold true, and Dongfeng can steer clear from the 2001 service manual and instead look for the 2001N and newer schematics available one directory further above.

Actually the ROM list I posted a link to says e.g. 901447-24 is a 2316-011 ROM Basic 2 (E000-E7FF) which should settle the case that I was wrong above.

I'll wait for Dongfeng to try and dump the ROMs. I think I'm short on 2532's but may still have some original spare ROMs if you're really lucky.

November 27th, 2009, 06:20 PM
Thanks for the replies so far!

Interesting about the model, maybe my PET does indeed have an upgraded motherboard. The dates on the chips date from very late 1979 to very early 1980. When did they stop using the chicklet keyboard? That might explain why the spec sticker has been removed from the back... or it could be Commodore using up old cases. I have also yet to see another one branded as "CBM 2001 Series" so could be something to do with that? It also has the green-text monitor... Hmmm...

Sadly, I don't have a scope or an eprom burner... I've always wanted one so maybe is the time to get an eprom burner! I haven't touched a 'scope for years and have completely forgotten how to use one. I have a multi-meter though :D

Could I do it with something like this: http://search.ebay.co.uk/200399776235 ?

Dwight Elvey
November 28th, 2009, 06:08 AM
It is most likely the address decoder that is used to
address the video RAM that has failed. The fact that
the characters look fine would indicate that the
analog part of the video is working OK.
As for the screen going blank, that may be something
else wrong.
I'm not a Pet man so this are just guesses on my part.

February 27th, 2011, 02:51 AM
I've the same computer model. The label on the front says "CBM 2001 series" and the one on the back says "2001-8". The manual I've got is for CBM 3032 series and the references inside it are for 2001 series, a quite strange. The motherboard is the same, except for the 8 RAM chip sockets soldered on the first row, so my system has 32K. I think the strangeness of this model are due to the fact that at the time Commodore was changing rapidly the release of internal cards, correcting bugs in the firmware, and also it could not export its computers in Europe as PET, because of the issue with Philips. The CBM 2001 was the result and may be that these specimens had been the last before the 3000 series itself, with standard keyboard and external tape recorder. So I think we have a CBM 3016 / 3032 indeed. Comparing the real machine with the one emulated with VICE, this hypothesis seems to be confirmed. Regarding the malfunctions, I would look at the kernel ROM and the logic that drives the video. This model doesn't have the video chip yet.