1978 Commodore PET project

[BrackenFix]

Originally Posted by SiriusHardware

If you could just let us know whether your board is (or is not) the one in the first PDF in AJ's post #14, we'll be in a better position to advise.

One of the ICs you had out when cleaning up the PCB would have been the character generator ROM which is one possible cause of your display problem - if you could just make sure none of the ROMs have any of their pins folded over? I have been using DIP ICs for almost 50 years, and even now I sometimes get a pin out of line when inserting an IC into a socket.

No problem, ill report back when I get home later.

Will check every IC that i removed too

[ajgriff]

I wouldn’t be in a rush to buy additional test equipment or to de-solder anything. There’s a lot that can be done before considering the need for those options.

As Sirius says let’s identify the correct documentation for this board as a basic starting point. Studying these documents will make life easier for all concerned as well as increasing the chances of a successful repair.

Alan

[ajgriff]

Originally Posted by BrackenFix

.... hoping it was just a dirty video RAM chip.

Referring to the photo of the board it looks like the two video RAM chips (probably 2114s) are soldered rather than socketed.

Alan

[BrackenFix]

Originally Posted by SiriusHardware

The circuit diagrams for several different variations of this model are here:-

http://www.zimmers.net/anonftp/pub/c...001/index.html

If you can find a 32...... board number on the mainboard that would be helpful, and also report which sort of RAMs and which sort of ROMs are fitted in the PCB, that will help us to narrow it down to the diagram which best matches your computer.

I agree with Mark that the initial problem seems as if it may be in the display area rather than the overall CPU area - it looks as though the CPU is running to the extent that it fills the screen memory with spaces and then the startup message, but both the spaces and the startup message are being displayed incorrectly.

If you had a scope then you could (in conjunction with a test EPROM) try to narrow down the area of the fault without disturbing the actual hardware too much.

Without a scope you have to resort to cruder methods which may ultimately include replacing the display RAM and its support ICs, starting with the ones which are the cheapest to buy and easiest to replace. (So usually, you would replace the display RAM last).

You may also have a secondary problem in the form of a mostly none working keyboard but that will be easier to diagnose and fix with a fully working display, so fixing the display is the priority.

If you can first let us know the PCB number and the types of ROM and RAM IC fitted, we can then begin to make suggestions about how to proceed.

Hi

The board is 320350 REV E

I need to learn what the RAM and ROM chips are so will report back

[BrackenFix]

System RAM - TMS 4108-25JDL-0 (x8)
Video RAM - MPS 2114 45L 1280 (x2)

MPS6520 0179
MPS 6522 1380
8237 SY6502A
MC6821P 309-408

[SiriusHardware]

Hm, that's a little bit confusing. We have diagrams for 320349 and 320351, but not for 320350. This from another forum:

"On the bottom of the board it says Part No. 320350 Rev E. On the top it says Assy No. 320351".

That makes it sound as though the part number for the PCB (only) is 320350 and the part number for the assembled PCB is 320351.

I think we'll assume your PCB corresponds to the diagram for 320351 (As AJ suggested) for the time being. Edit - just caught your update with the chip details.

[ajgriff]

I think the files I posted earlier are as close as you’ll get in terms of diagrams.

Alan

[SiriusHardware]

Although there is quite a lot already working on your machine, there is a standard approach to servicing, which starts with

- A visual examination to see whether there is anything obviously physically wrong (burnt or blown up components, damaged tracks, loose wires or connectors, etc). Since you've already gone over the PCB to clean it I would say that you have effectively already carried out this step, and presumably found nothing which leaps out.

- Next step is to carry out basic checks on all the supply voltages. Although the behaviour of the machine does suggest that the supply voltages are in the right ball park area, you should never assume so because a supply voltage off or out by more than a certain amount can cause some very odd behaviour so it's always worth taking a short time to verify these voltages.

Could you please measure the DC voltages on the component leads indicated on the attached sketch? In case it's not clear from the diagram, we are asking for measurements on the right hand ends of diodes CR10, CR11, CR12 and on the lowest lead of regulator VR6. If you need somewhere to put the black probe while making these measurements, the -Ve (right hand) end of C62, the large electrolytic with the cable tie around it, is connected to 0V.

[ajgriff]

Originally Posted by BrackenFix

System RAM - TMS 4108-25JDL-0 (x8)
Video RAM - MPS 2114 45L 1280 (x2)

MPS6520 0179
MPS 6522 1380
8237 SY6502A
MC6821P 309-408

The RAM chips are exactly as expected. The 6502 CPU and the 6522 VIA look to be the original components. However, one of the two 6520 PIAs has been replaced with a later equivalent (6821) so someone has worked on the board before which is useful to know.

By the way, when reseating the ROMs it's important to return them to their original positions.

Alan

[BrackenFix]

I have just realised that 2 of the legs on chip 901465-02 are damaged. I'm 99% I was very careful when removing it.

It is different to all the others as is purple with gold legs

[ajgriff]

With reference to the board layout diagram which chip is it?

Alan

[SiriusHardware]

Regarding the damaged ROM, fixing or replacing that is now the priority because the system won't work at all without a working set of ROMs. One way to try to rescue it is to put it into a conventional 24 pin IC socket and solder what's left of the damaged pins to the contacts on the top side of the socket, then plug the IC+socket into the original mainboard socket. If it is only the narrow part of the IC pin which has fallen off, cut a couple of complete pins off any spare or scrap IC and solder them to what's left of the broken IC pins. In the long term you will probably need to get a Hitachi or Texas 2532 EPROM programmed as a replacement. You won't be able to carry out any of the steps below until you have repaired or replaced that ROM.

Assuming the supply voltages turn out OK, possibly the next thing to do is to measure the voltages on the video RAM data pins.

While a meter will only show the average voltage, that may be enough to show whether one bit is stuck high or stuck low because the chips are permanently selected and in read mode for the vast majority of the time. Even better if any bit identified as possibly stuck coincides with the bit Mark flagged up as potentially stuck.

After that, in the absence of a scope it may be necessary to remove the video RAMs UF7, UF8, and socket and swap them over to see if that changes the nature of the fault. This is something which can be tried without buying any new parts (with the possible exception of IC sockets).

If that exonerates the video RAM ICs (ie, no change when swapped) we can remove the video RAM and force various states on the SD0-SD7 lines to verify the operation of the video latch UF9 and the character generator ROM UF10.

The above suggestions assume that BrackenFix feels able to desolder ICs from double sided PCBs without damaging either the IC or the PCB.

Ordinarily, we would pursue a more diagnostic approach first but in the absence of a scope we are already almost at the point where the most pragmatic approach is to chip-swap. I'm assuming (BrackenFix) that you don't have a logic probe otherwise you probably would have mentioned it.

[SiriusHardware]

With reference to the board layout diagram which chip is it?

I believe it will be UD7 (Basic 2, ROM 2 of 2).

[BrackenFix]

I believe it is UD7

[BrackenFix]

Ok, I have bought a socket to try and salvage the chip.

Sadly, I don't have a logic probe

[SiriusHardware]

Don't worry if you can't save it, there are a couple of us here, myself included, who have a few 2532 EPROMs and the means to program them so the machine need not be out of action for long.

We may have to ask you to remove and socket a few other ICs in due course so it might not be a bad idea to stock up with a few common sizes of conventional IC socket, 20 pin for the buffer / latch ICs, 18 pin for the video RAM ICs, 14 and 16 pin for general logic ICs.

[SiriusHardware]

Incidentally, you may wonder why I specified a conventional socket for the 'carrier' for the injured ROM rather than a turned-pin socket. We have found in the past that inserting the relatively thick, round pins of a turned pin socket into the sockets in Commodore mainboards spreads the contacts in the mainboard socket and makes them unreliable, hence the suggestion to use a conventional socket which has thin, flat pins.

[BrackenFix]

Originally Posted by SiriusHardware

Don't worry if you can't save it, there are a couple of us here, myself included, who have a few 2532 EPROMs and the means to program them so the machine need not be out of action for long.

We may have to ask you to remove and socket a few other ICs in due course so it might not be a bad idea to stock up with a few common sizes of conventional IC socket, 20 pin for the buffer / latch ICs, 18 pin for the video RAM ICs, 14 and 16 pin for general logic ICs.

Luckily, I bought loads of different sizes.

I got the traditional carriers and not turned pins.

The thin legs broke off, there is the larger part of the leg is still there so I should be able to solder them.

[Mark1960]

With conventional modern IC sockets you might still need to take care before inserting them into conventional retro IC sockets. The modern style tend to contact the outer and inner surface of the IC pins. The retro IC sockets tend to contact on the edges of the IC pins. Some of the modern style have a wider flat pin to solder into the board and these can open up the retro style sockets past the point where they maintain contact force.

[SiriusHardware]

The thin legs broke off, there is the larger part of the leg is still there

In that case you might be able to rescue them without using a socket, by soldering 'donor' pins from another IC to what's left of the original pins. The attached crude sketch illustrates how, with the new / replacement pin highlighted in red soldered to the inside of the upper half of the original pin - this makes for a fairly invisible repair when the chip is inserted back into a socket.

Of course you can also solder the replacement pin to the outside of what is left of the original pin but then the repair is a bit more more obvious and not quite so strong.