Commodore PET 2001
 

Vintage computing really isn’t my thing. However I did dabble with ZX81s, Spectrums, VIC20s etc many years ago and at the time acquired a PET 2001 at a school fête of all places. I don’t currently have access to the machine due to current restrictions but from memory I found it to be complete, in good physical condition and apparently unmolested. It's been in dry storage all this time. It did boot but just displayed a locked screen full of random characters.

The machine’s current custodian is keen to see it disappear and I will of course oblige when circumstances allow. In the meantime I’d appreciate any thoughts on how best to proceed bearing in mind that I’ve no desire to keep it in the longer term. In truth I wouldn’t mind getting it working before disposal if this can realistic be achieved in terms of things like parts availability.

I appreciate that I’ve only provided scant information but the thoughts of anyone with experience/knowledge of the PET 2001 would be most welcome. In the end I’d like to pass it on to someone with an interest in vintage computing via the forum offered section.

Alan

Re: Commodore PET 2001
 

I should have said that this is the 8K version with chiclet keyboard and built in cassette player like the one in the attached photograph.

Alan

Re: Commodore PET 2001
 

Hi,

I have worked on a couple of these over the last few years.

The first had a faulty PIA chip that was holding down the reset line so the processor wouldn't run. I removed and replaced it and all was well.

The second had a faulty RAM chip if I remember correctly it was a 2114. Again replaced and it leapt into life.

Fortunately these machines don't have a fist-full of special chips so they are quite nice to fix. They also have a flip up lid and a bonnet stay so that can't be that bed.


All the best,

Andy.

Re: Commodore PET 2001
 

Many thanks for that Andy. The threat of this venerable old machine being sent for recycling has prompted me to take an interest in it at last. When I originally bought it life was hectic with young children around and of course guidance on possible repair just wasn't as easy to find as it is today. I'm aiming to try and find a way to get hold of it asap so that some proper investigations can be undertaken.

In the interim any further contributions would be much appreciated.

Alan

Re: Commodore PET 2001
 

I have a 2001-8 in the to-do queue for refurb. Yours sounds to me like its not resetting properly. Early PETs had a fault in the 555 timer circuit used to reset the CPU - mine was fixed by the retailer back in '78 when my Dad first bought it, so I don't know exactly what they did but I suspect they increased the value of the timing capacitor so it held the reset low for longer. The mask ROMS Commodore used ran hot and were unreliable (a lesson they evidently didn't learn for the VIC-20 or C64!) - a faulty ROM can cause the PET to crash on reset and not clear the screen. The good news is that if the display shows random characters you can check that the character ROM is OK, and it shows that the VDU circuitry is working.
Replacements are still being made for the 6522 VIA chips by the western design centre, who also make 6521 PIA chips, which are a pin-compatible substitute for the 6520 in the PET. I've looked over the data sheets and the difference between the 6520 and 6521 is mostly the drive capacity of the I/O pins.
Unfortunately in early versions of the 2001 commodore used its own 6540 ROMs and 6550 RAMs that are not compatible with any EPROM or RAM. Later versions used 2716/2732 compatible ROMS and 2114 RAMs. There was a ram/rom board and diagnostic board made by Tynmouth Software but they have shut down due to COVID and they don't seem to have released their designs publicly.
Schematics for all 2001-8/16/32 variants can be found here:
http://www.zimmers.net/anonftp/pub/c...001/index.html

Images of the ROMS should you need to blow roms are at:
http://www.zimmers.net/anonftp/pub/c...pet/index.html

I did design adapters to adapt EPROMS to fit 6540 sockets but have yet to have them manufactured or tested. I'm not sure if there is enough clearance around the ROM chips (my 2001-8 is in storage, also inaccessable at the moment....)

I would heartily encourage you to replace all the electrolytic capacitors especially in the monitor, as they are likely to be very tired now. I failed electrolytic in the monitor could damage the irreplaceable CRT. The main smoothing capacitor might be suspect too, and it might be worth wiring a bulb in series with it when first switching on the power supply in case it is very leaky and the transformer gets damaged.
From what I recall, my PET had a number of Tantalum capacitors too, which I intend to replace with new Electrolytics.

Re: Commodore PET 2001
 

Really useful detailed information, thank you. Plenty to think about and study in advance of diving in. I'd already decided to fire up the machine using a Variac once the initial physical inspection has been completed.

As I'm fairly sure that this is an early machine the potential problem with unobtainable RAM chips is a worry. I'd be interested to know if anyone is aware of possible work arounds?

Alan

Re: Commodore PET 2001
 

While the 6550 RAM is functionally equivalent to the 2114 RAM (i.e. its 1kx4) the 6550 has 4 chip select inputs (2 active low, 2 active high) and a Phi-2 clock input to synchronise the write timing so a direct pin adapter is not easy. In the case of my 2001-8 the RAM failed a few years after buying the computer (just out of warranty...!) so we took the opportunity to upgrade to a 32k dynamic ram card that plugged into the CPU socket with the CPU plugged into that. I suppose you could do that these days with a single 62256 RAM chip!

I did design a 6540 ROM -> 2716 adapter (which has 5 chip select inputs!) so its possible you could make similar adapters for 6550->2114 rams.

Re: Commodore PET 2001
 

Again more useful info. Can't really progress much further until I can get hold of the machine. Will report back in due course.

Alan

Re: Commodore PET 2001
 

I have a full set of 2114's for this. I would also recommend upgrading the roms to the later version.

Re: Commodore PET 2001
 

Yes, the basic 2 (Starts with "*** commodore basic ***") has serious bugs in the IEEE488 bus, which prevents the use of disk drives to save programs. The later Basic 4 (Starts with "### commodore basic ###") is much better.

Re: Commodore PET 2001
 

As of this evening the PET is with me after a lockdown compliant transfer. Haven't been able to do anything much with it yet other than clean off the worst of the dust and confirm that it is an early 2001-8 with 320008 main board complete with 6550 ROM and 6540 RAM chips. Everything looks very much original. Case and keyboard are in good condition apart from one or two minor scratches on top of the monitor.

I hope to continue investigations tomorrow and will aim to post some 'photos as well.

Alan

Re: Commodore PET 2001
 

Other way round - sorry. Inexperience showing.

Alan

Re: Commodore PET 2001
 

Re: Commodore PET 2001
 

A standard routine service on these was to remove and replace the chips one by one since the sockets used the standard "side swipe" contacts. Poor chip to socket contact was a constant cause of trouble as were general interconnects. We had a fleet of around 50 of these in various versions forty years ago, most faults were caused by poor contacts.
Neil

Re: Commodore PET 2001
 

Yes, I would recommend at very least "rocking" each socketed chip and pushing it firmly home. They have a tendancy to corrode and push themselves out of their sockets.

If the video display is stable then that means your video ram is probably OK, and the character generator ROM is clearly working well, so thats good.

You can manually reset the processor by grounding contact 27 on the big edge connector (the contact on the other side is ground, so I used a suitably bent paper clip for this purpose, but since its connected to the output of the gate I wouldn't keep it grounded for long).
If you have a logic probe or multimeter on pin 3 of the 555 it should go high for about 10s or so - this should be easy to see with a LED and resistor connected temporarily. TBH I'd probably replace the Tantalum 10uF capacitor as a matter of course....

Re: Commodore PET 2001
 

I knew this would be fun! Turned on the PET today to do some voltage checks and was greeted with the display shown in the attached 'photo. The display also flickered badly. The flicker was easily cured by cleaning the pins of J7 connecting the monitor to the motherboard. The black bar was gone when the machine was re-booted but reappeared at a later stage once the PET had time to cool. It now reliably reappears when the machine is started from cold and disappears when it's re-booted after running for a few minutes. Any helpful suggestions gratefully received!

As for voltage checks, all four 5V lines measure correctly. The voltage on pin 3 of the 555 timer rises to 4V or so shortly after switch on before dropping again after about a second which I think means that the circuit's working correctly. Someone please put me right if I've got that wrong. Resetting the 6502 processor by shorting pin 2 of the hex inverter has no effect by the way.

Setting aside the black bar problem my next step is to take out the motherboard so that I can then remove all the socketed ICs for pin cleaning and re-seating. I'll then attempt a start up without the PIA chips and the unnecessary 12 RAM chips which, in theory, should allow the system to boot with minimum memory. It'll be interesting to see what happens.

Alan

Re: Commodore PET 2001
 

Quick update on progress such as it is. Seem to have cured the black bar problem by cleaning the connectors inside the monitor. Fingers crossed.

Checked the output voltage on pin 2 of the inverter at start up and it stays low for a couple of seconds before rising to a steady 3.7V. This is roughly in line with the SNLS04 datasheet so I think the MPU's timed reset circuitry is working properly. For reference I'm attaching a cropped image of the timer circuit and PDFs of the schematic plus parts list for this particular PET 2001 version.

The only other thing I've noticed is that each time the machine is started the random character screen is very similar but not identical. Don't know whether or not this means anything. Anyway next step is to remove the motherboard and start pin/socket cleaning.

Alan

Re: Commodore PET 2001
 

Yes, Commodore often used RAM / ROM IC's (usually their own special versions, to keep the MOS company they owned in business) with extra chip selects to save on some external address decoding logic. But may sometimes be able to do some simple diode-OR etc. logic to combine a few into one, to be able to use a more standard Memory IC.

Although with some active-high and others active-low (and some gating with clock phase required), then some external logic is probably required - Maybe using a PAL/ GAL etc. that you can now get re-programmable ones cheap.

And with non-JEDEC std. pin-outs on their special RAM IC's etc, then you really need to make a small conversion board anyway, rather than stacked-sockets and intercepting the odd pin trick for ROM-to-EPROM etc. converters

Re: Commodore PET 2001
 

Yes, those tin-plated connectors have probably oxidised quite a bit over the decades, so may well need a bit of cleaning - either by moving the connector up and down a few times, or more thoroughly by cleaning pins with some very fine abrasive material / fibreglass pencil etc. Getting to the terminals in the housings, is a bit more tricky. And rather than having to release them, suitable contact-cleaner spray may work quite well. The same for the IC's in sockets.

If you're getting different random characters each time, then that's probably just due to random power-up state of the Video RAM that isn't being cleared - because the O/S ROM Program isn't being executed by the CPU, to initialise the RAM / put the right contents into it for the expected display.

And that can be due to the CPU itself (or its reset / clock / power rails not being quite right) or a fault in the ROM's / it's working-area RAM or power rails to these / address-decoding logic.
In the absence of known-good spares / a logic-IC tester or something similar that works to swap IC's with, then it's probably a case of 'scoping / logic probing all the data & address-lines to see if there's any stuck ones that might indicate a faulty IC on that line. (After checking all power rails are in spec. & clocks + reset are OK)

I've also got one of these PET's, that also does the same random characters display, that I need to get out of where I've stashed it one day, and fix.
IIRC, it too had the board with the 6540 / 6550's, but I think I found someone with an IC tester who managed to get it to test the RAM's
Or could build one for RAM's, as there are similar ones for 2114's on Github etc. that use an Arduino + shield and a bit of software to light a red or green led.
I did buy few spare PET boards I spotted at a Radio Rally once, but I think they were one of the other types of the 4 that CBM seemed to randomly fit whatever they had around to many of the PET models!


It's a shame Dave Curran has recently removed most of his Tynemouth Software website, to stop people making enquiries about these at present
- As his blog did have quite detailed articles on how he repaired various CBM PET's over the years including his own design IC Adapters - I'll have to see if I've got any saved-copies of these.

Re: Commodore PET 2001
 

It is awkward but, you can move around the site as it was last November at archive.org https://web.archive.org/web/20201101...oftware.co.uk/