• Please review our updated Terms and Rules here

Commodore pet 2001-8 no video from logic board

It didn't occur to me that there may be modern equivalents. Fingers crossed I will get my post this Saturday. It's going to be like Christmas morning due to the build up 😆
 
Ok so the chips finally arrived after taking two weeks to get to me due to the postal stikes. And they where here in the UK🙄. But good news we now have a full screen of garbled goodness! Yay!
So B1 was the culprit in the end. Well done to Nivag for the diagnosis.
So I guess I am looking at a faulty character set now? If someone could confirm this that would be great. I have ordered a replacement one already in an adapter just in case.
 

Attachments

  • DSC_1044.JPG
    DSC_1044.JPG
    2.6 MB · Views: 15
Yep, that looks like either a faulty character generator or possibly the video shift register or something driving the character generator.

Is the character generator IC in a socket? If so, we can remove it and stimulate the data outputs.

You always appear to have the left-hand side pixel of each character cell set. This might be the data bit to check to see if this is toggling from the character generator (indicating a shift register fault) or not.

Of course, we coukd have two (2) independent faults concurrently...

it is strange that we either have completely solid blocks or completely blank blocks though.

Also, you should be able to monitor the address line inputs to see if they are all toggling away.

Perhaps a Saturday afternoon looking at the address lines to, and the data lines from, the character generator with your oscilloscope to see what you have?

Dave
 
Last edited:
Yep, that looks like either a faulty character generator or possibly the video shift register or something driving the character generator.

Is the character generator IC in a socket? If so, we can remove it and stimulate the data outputs.

You always appear to have the left-hand side pixel of each character cell set. This might be the data bit to check to see if this is toggling from the character generator (indicating a shift register fault) or not.

Of course, we coukd have two (2) independent faults concurrently...

it is strange that we either have completely solid blocks or completely blank blocks though.

Also, you should be able to monitor the address line inputs to see if they are all toggling away.

Perhaps a Saturday afternoon looking at the address lines to, and the data lines from, the character generator with your oscilloscope to see what you have?

Dave
Great progress !
 
I would check for activity on UA2 (74166) pins 15 (shift/load) and 7 (clk). This is the video parallel to serial shift register.

Then check UA3 (character generator). Pins A0, A1 and A2 (check the schematics for the actual pin numbers) should have nice steady clocks on them.

Pins A3 through A9 should have latched data from the video RAM on them.

What are the states of A10 and the three (3) chip select pins? They should be levels, not clocks.

Then, check the data output Pins (D0..D7).

Dave
 
Perfect. I am out buying my Christmas tree...

There are a few things we need to know next... Is it the same pattern each time?

If you remove the character ROM does it stay the same?

Looks like the video timing is golden now.

Merry Christmas!
 
All I want for Christmas is a working PET 😀!

There could be a song there somewhere...

Dave
I don't want a lot for Christmas
There is just one thing to get
I don't care about the presents underneath the Christmas tree
I just want to fix my PET
More than you could ever guess
Please bring my PET back to me
All I want for Christmas is he.
 
Given that replacing B1 has got you a screenful I feel I can continue to help! Now there are a few things to look at... Again you can go forwards or backwards.... We are aiming at a garbled screen of rendered characters (since we are going to assume that the CPU side has many defects and is not running at this stage)... Now we know the counters and sync logic is working so it is buffers. RAM, shift register ROM,... Working forwards but being a contrarian I suggest first we look at C1 due to the blocky nature of the output ...

Warm up a mince pie and put your logic probe on all the pins of C1
 
Well that was short lived. Back down to a single line on the screen after trying out the character set eprom removal and placing back. 😔

I was getting noting out of the vertical pin 11 on D8 but pin 13 and 12 showed inputs so I replaced it but still the same.

Since new D8, Pin 13 and 12 both low pin 11 hi.
Not much on scope from J3 pic 2

Btw I now have a X1 probe so I have adjusted volts div to 1
 

Attachments

  • DSC_1050.JPG
    DSC_1050.JPG
    2.4 MB · Views: 2
  • DSC_1049.JPG
    DSC_1049.JPG
    793.7 KB · Views: 2
Bummer...

But not wholly unexpected either I am afraid...

Where one IC has died, sometimes you should expect more will occur as we exorcise the demons out if the machine!

Dave
 
With new D8 in socket

B6
Pin 1 lo
Pin 2 lo
Pin 3 hi
Pin 4 hi
Pin 5 hi
Pin 6 lo
Pin 7 lo
Pin 8 hi
Pin 9 lo
Pin 10 hi
Pin 11 lo
Pin 12 lo
Pin 13 hi
Pin 14 hi
No pulse on any
 
Interesting though how the state machine and the address generator was working and then stopped.

The possibilities could be a bad socket connection or an intermittent IC or a new IC failure. If the machine is left off for a while and re-powered, it would be worth checking if the V.drive signal transiently recovered before failure. Any sign of an intermittent or thermal problem, you could try some freeze spray on individual IC's.

As mentioned with the signal ultimately being generated by a recursive digital loop and one IC in it failing/stopping and the whole loop stops, it might be required to break the loop and inject a clock pulse from a generator to find out which IC has kicked the bucket.

But with any luck that might not be required.
 
OK so nothing much happening... back to basics...

Do you still have HSYNC?

What's happening at pin 2 pf C5?
I think Pin 2 of C5 must be ok in both its frequency & duty cycle, or there would be no horizontal line on the VDU.

(It is worth turning the brightness right down on the VDU while this fault persists, or unplug the VDU if the beam cannot be extinguished or set for a very low brightness)

Possibly pin 13 of C6 will be stuck low and its output low too, which inhibits C7 and kills the difficult to see clock pulses to the state machine. It is definitely worth checking all three pins of C6 , pin 13, 12 and 11, because if pin 12 of C6 got stuck low it would give the same result, killing the state machine, and the fault might not be in the recursive loop circuitry.
 
Last edited:
Next... D5 PIN 3... then D6 pin 12.... then D7 pin 12...

(Hugo's point is a good one... make sure B1 is well seated)
 
Last edited:
Back
Top