• Please review our updated Terms and Rules here

PET 2001 Fault

C

Call Me Nige

Member
Joined
Oct 1, 2024
Messages
11
I have picked up a PET 2001 from a fellow local arcade collector, he acquired it a while back, fixed it and added a Romulator, but it no longer works (and he practically gave it away so I can't ask him to fix it again).

The symptoms are a full screen of the same character
Pet Fault 1.jpg
Which varies, usually R, graphic square or Q
Pet Fault 2.jpg

I have done the basic checks - the regulators are outputting a solid 5v, swapped out the 6502 for another made no difference, removed the Romulator and put the 6502 in the socket no difference, cleaned De-oxit'ed and reseated all socketed ICs again no difference.

Any pointers would be gratefully received. Photos of the PCB below, cosmetically the PET is in amazing condition and the monitor is still sharp, just the small matter of the computer itself not working.

Pet Fault 3.jpg Pet Fault 4.jpg
 
Your problem is video related. I suspect either the video RAM or the video RAM data buffers. There are other causes as well - data latch between the video RAM and the character generator ROM (if there is one) depending upon which version of the 2001 you have.

The ROMulator looks like bitfixer's. This should contain a copy of my PETTESTER code.

Make a note of the current settings of the DIP switches and then read the document here: https://github.com/bitfixer/bf-romulator for how to set-up the switches to run my PETTESTER.

What do you observe on the screen when you run my PETTESTER code?

For my own information (so I can lookup the correct schematics) what is the number on the logic board?

Dave
 
It is Bitfixer's ROMulator, unfortunately none of the tests changed anything, I tried all 3, still the same pattern on screen.

The board revision is 320008.

I forgot to mention two other things if that gives any clues - the red LED is not on and none of the ICs are hot to the touch.
 
The RED LED (on the PET logic board) is not usually on...

I assume you have a couple of 6550 RAM devices at locations C3 and C4 on the board? Are they in sockets?

What is at A2? Is that a 6540 character generator ROM?

Do you have 8 off 100 Ohm resistors (or near value) in your 'box of spares'?

Dave
 
The video RAM is present at C3 and C4 and socketed, so I swapped them around and also with 6550s from the system RAM, other than a variation in the character displayed across the whole screen it made no difference.

The 6540 is present at A2.

Yes, I have plenty of 100 ohm resistors.

Which chip(s) is the video RAM data buffer, would it be worth me changing that first?

Thanks for all the help so far by the way.
 
It behaved exactly as the test set out, with the 8x resistors on C3 to C4 tied to ground the screen showed '@' in all positions.

So I guess the character generator ROM is OK, which is good as that's custom.

The fault in your link was different, any suggestions where to go next?

I'm guessing the 6550s are OK as swapping them around made no difference, it could be the multiplexers at D3 and D4?
 
Before you remove the resistors, I would just move some of them from 0V to +5V.

If you do this, then the character on the screen should change from an '@' to something else that is related to the PETSCII code for the display data bus you have just forced with the resistors.

It could be the multiplexers D2, D3 and D4.

However, I am suspecting that the SAn signals are not changing, and what is being displayed on the screen is one character cell spread across the whole screen. The character itself will be pseudo-random based upon the address that the SAn lines settle upon and the contents of the 6550 SRAM on power-up. Just a thought.

You can check this by looking at the signals around F6, D6 and D7.

Alternatively, it could be a control signal (e.g. the clock or R/nW) to the 6550 video RAM. Check the pins around the 6550 video RAM.

Dave
 
You could try using an MCL65+ in place of the 6502 to peek and poke memory and peripherals from the CPU's socket.
 
daver2 said:
However, I am suspecting that the SAn signals are not changing, and what is being displayed on the screen is one character cell spread across the whole screen. The character itself will be pseudo-random based upon the address that the SAn lines settle upon and the contents of the 6550 SRAM on power-up. Just a thought.

You can check this by looking at the signals around F6, D6 and D7.
Click to expand...
daver2, I don't think so. If the D6 and D7 counters were struck or random, the vertical drive timing sequencer would not be working correctly and we would not be getting a raster scan.
 
Hi dave_m. Long time no hear.

I was thinking specifically of the SAn signals (after the multiplexers) not the D6 and D7 counters. Although how ALL of the SAn signals can be static I am having a hard time thinking about that as a fault.

However, having the buffered clock dissappear on another thread, that poses an interesting fault scenario when you have clocked 6550 video RAMs!

Dave
 
I've finally found the time to hook up the scope today, and although I can't be sure exactly what I'm looking for, there is what appears to be solid 5v peak to peak logic showing on all pins on E6, D6 and D7.

Except for pins 2 (input) and 5 (output) on E6 and pins 4 (input), 5 (output) tied to 6 (clk 2) and 8 (clk 1) on D6, all of which are interconnected in a way beyond my understanding. On these there is some sort of signal but it's a mess with hanging voltages, almost look like a sine wave as they rise and fall. D7 on the other hand shows 5v peak to peak logic on the equivalent pins, no issues there.

Tomorrow I will check the SAN signals on D2, D3 and D4. The output SA2 on D3 is linked to the same circuit above, maybe this is the cause, although given its changing output then surely the display would also change, even if showing garbage?
 
Call Me Nige said:
I'm guessing the 6550s are OK as swapping them around made no difference, it could be the multiplexers at D3 and D4?
Click to expand...
Nige,
Don’t assume you have any good data RAM chips. It is possible that the only working 6550 RAM chips on your board were the two video chips.

When a RAM/ROM board is needed to make a PET work on an original PET 2001 with its rare 6540 and 6550 chips, and it is intended for resale, the repair tech may elect to keep any working memory chips for his next repair.

So now if the video RAM chips have given up the ghost, you may need to replace them with an adapter using newer static RAM. But that is for later.

Keep troubleshooting as daver2 says, the D2, D3 and D4 LS157 multiplexers as they may be the problem.

I’ll see if I can find a link to a video RAM adapter if needed.
-dave_m
 
I'll check my parts bin tonight, I think I have some more of those ram adapter pcbs left, and I'm pretty sure I've got the srams laying around.
 
dabone said:
I'll check my parts bin tonight, I think I have some more of those ram adapter pcbs left, and I'm pretty sure I've got the srams laying around.
Click to expand...
Yes, I remember. I bought a RAM video adapter from you many years ago when I was hoping to buy and fix an original PET 2001.
Alas, I never got one, and now I am too old to be collecting more classic computers!
 
I'm not sure it's the 6550 video RAM, the problem seems to be before it even gets that far. I can try swapping out them out for system RAM, I only tried a couple but can be more thorough.

After some more probing around with the scope (a cheap Chinese handheld but should be enough for the job) I might be onto something, all of the faulty lines are linked, as per the lines in yellow below. These aren't showing normal 5v logic, they're a total mess with signals up and down in a waveform, some of them hanging around 3v.

The problems seem to start at D5 and all on flip-flop 1, I need to do some more probing around it, to be honest I really don't understand what is going on with this whole circuit and signals looping back around. All other SAn signals appear to be fine

Pet fault.jpg
 
Call Me Nige said:
The problems seem to start at D5 and all on flip-flop 1, I need to do some more probing around it, to be honest I really don't understand what is going on with this whole circuit and signals looping back around. All other SAn signals appear to be fine
Click to expand...
They are probably a little noisy but these signals are good TTL level signals. If not, you would not be getting the good raster scan to create the 25 lines of 40 characters (1000 characters) you are seeing.

Flip-flop D5 and counters D6 and D7 form a 10 bit counter that is doing double duty. It is incrementing through the 1000 addresses necessary to output the video RAM data, and it is being used to help detect the proper count to increment a state sequencer to the next state. See dual flop-flop B6 at the top of the 2001 schematic sheet 3.

These two flip-flops form a 4 state sequencer that creates the vertical drive waveform along with the vertical blanking periods. Sounds simple, but look at the complicated clocking circuit! However this circuit is all working (be grateful).

In the 4000/8000 version of the PET, this circuit was replaced with the 6545 CRT Controller.

You are right about signals ‘looping back around’. This circuit has caused a lot of grief to PET computer repair people.
dave_m
 
You must log in or register to reply here.
Back
Top