• Please review our updated Terms and Rules here

Commodore PET 2001-8 stuck on garbled screen

Interesting when I power cycle with the probe attached to pin 3 from the scope I get a memory reading of 7167 bytes free which I guess is what I am looking for.

DSC_1880.JPG
 
Last edited:
Ok and now with the probe off it's still showing 7167 bytes free. Strange. All I have done is probe G3.

I tried the first poke. Nothing seemed to happen.
Curser disapers and it just hangs until i interrupt it.DSC_1883.JPG
 
I think you have some problems with G3 (possibly).

That behaviour shouldn't happen (i.e. with the oscilloscope probe)...

I would suggest a NOP generator test and check G3 pins 1, 2, 3, 4, 5 and 6.

Dave
 
daver2 said:
I think you have some problems with G3 (possibly).

That behaviour shouldn't happen (i.e. with the oscilloscope probe)...

I would suggest a NOP generator test and check G3 pins 1, 2, 3, 4, 5 and 6.

Dave
Click to expand...
It's been at 8k now for a good half hour. I am leaving it off to cool down to see if I can repeat the fault. I have had equipment that is faulty behave differently at different temperatures. If only I had some freeze spray.
 
BTW we are obviously playing a bit fast and loose by poking random locations but in this case we will probably get away with it.

If you are interested in how a PET really works then read... http://www.primrosebank.net/computers/pet/documents/The_Commodore_PET_revealed.pdf
I remember reading it at school!

Specifically you should really tell BASIC not to use the area you are poking... I could have told you how... 40 years ago.
 
RetroGadgetMan said:
It's been at 8k now for a good half hour. I am leaving it off to cool down to see if I can repeat the fault. I have had equipment that is faulty behave differently at different temperatures. If only I had some freeze spray.
Click to expand...
Contact cleaner works as a defacto freeze spray, if you put a small pool of it on an IC top, the latent heat of evaporation cools the chip, you can also keep applying it to the one IC and it is one way of localizing temp sensitive chip or component. Sometimes it is better than freeze spray as you can apply it from a soaked cue tip in a more localized manner, the freeze spray spreads out more.
 
The difference between 4K working and 8K working is really only nSEL0 and nSEL1... this can manifest itself two ways.... either nSEL0 and/or nSEL1 isn't getting to pin 19 of I3 and/or J5 or the buffer logic is wrong due to G3 (and possibly the buffers).

Note that many of these devices are active low so a low where you don't expect it can turn things on and ruin your day.

Given we know the decoder G2 is mostly behaving a dodgy connection is the most likely IMHO.... second to that anaemic output of G3.
 
Check the soldering, especially any places that have traces on top of the PCB.
You know it has been replaced so it is quite likely that a feed thru was damaged.
Dwight
 
daver2 said:
I think you have some problems with G3 (possibly).

That behaviour shouldn't happen (i.e. with the oscilloscope probe)...

I would suggest a NOP generator test and check G3 pins 1, 2, 3, 4, 5 and 6.

Dave
Click to expand...
Ok it's back down to 4k now.

Here are videos of pins 1-6 of G3
Vid Mov_1960 being pin 1
And last image is what is on board.
 

Attachments

  • DSC_1966.JPG
    DSC_1966.JPG
    4.3 MB · Views: 1
  • MOV_1965.mp4
    6.7 MB
  • MOV_1964.mp4
    6.5 MB
  • MOV_1963.mp4
    6.6 MB
  • MOV_1962.mp4
    6.7 MB
  • MOV_1961.mp4
    6.7 MB
  • MOV_1960.mp4
    6.6 MB
1674318132065.png


G3 pin 1 is nSEL0
G3 pin 2 is nSEL1

Considering G3's gate to be an OR with inverting inputs (an exercise for the reader as to why that works)... it's inputs become SEL0 OR SEL1.... so it's output should be high when either SEL0 or SEL1 were active.

I would therefore expect the ON period of pin 3 to be double that of either pin 1 or pin 2 equivalent OFF period.

I'm not sure what I am seeing from those traces...
 
The basic problem with the 'one pin at a time' oscilloscope traces is that we can't observe whether the logic gate(s) to which the pins are connected to are actually functioning correctly or not.

We do (however) seem to observe /SEL0 and /SEL1 low-going pulses - so that is good information.

When /SEL0 goes low on G3 pin 1, G3 pin 3 should go high.

When /SEL1 goes low on G3 pin 2, G3 pin 3 should also go high.

If this is not the case, G3 is not functioning correctly.

The signal on G3 pin 6 looks strange however.

Dave
 
Ok, there looks to be a problem with your oscilloscope set-up.

Let's not bother with pin 1 but concentrate on pins 2 and 3.

Trigger the oscilloscope on pin 2 (the input pin connected to /SEL1).

Set the trigger mode to CHOP. I think you have it set to ALT at the moment.

What are you running for the CPU to give you these traces? A NOP generator or some firmware?

I am a little concerned that the timebase seems as slow as it is. What have you got it set to?

Dave
 
daver2 said:
Ok, there looks to be a problem with your oscilloscope set-up.

Let's not bother with pin 1 but concentrate on pins 2 and 3.

Trigger the oscilloscope on pin 2 (the input pin connected to /SEL1).

Set the trigger mode to CHOP. I think you have it set to ALT at the moment.

What are you running for the CPU to give you these traces? A NOP generator or some firmware?

I am a little concerned that the timebase seems as slow as it is. What have you got it set to?

Dave
Click to expand...
I am struggling to use two probes and take video or pics at the same time safely in fear of shorting pins. I don't have any wire clips for IC legs to attach the probes safely.
I have just had the board back out to inspect for any contamination from solder particles etc and given it a brush down.
Pressed G4 H8 and 9 that I replaced with sockets in firmly and now I have 8k back.
Hopefully that will be it for that ram saga.

Today I rebuilt the smashed up datasette. Put a new belt in cleaned it up internally. To my suprise it works to a fashion. But sadly it randomly chew tapes and the heads need aligning as its loading gibberish and errors from tapes. Capstan is fine as is pinch wheel. Tape rides down capstan.

I will look at testing the original rams again tomorrow.
Just to clarify, fill first two rows i 1 2 and j 1 with good ram and then one at a time for testing in j2.
If it boots wait for ram test to complete?

Ps tape connection on the back does not say searching when I press play. It does play though but nothing loads from a working datasette.
 
You must log in or register to reply here.
Back
Top