• Please review our updated Terms and Rules here

CBM PET 3032 STRANGE BOOT

I have strange wave on h11 pin 11
 

Attachments

  • 1679156570169..jpg
    1679156570169..jpg
    1.2 MB · Views: 3
I'm going out to dinner with friends tonight so I won't be able to proceed with the tests, I'll continue tomorrow if that's okay with you! A thousand thanks ;)
 
When you turn the machine on tomorrow, post another video of what you see on the screen 'from cold' and then what happens when things warm up.

Dave
 
That's very interesting.

The behaviour seems slightly different from what we saw before.... In this case pixels seem to go missing but it is always the letter E. I believe in earlier cases there might have been extra pixels?

Very good test! Good job
 
You notice that the whole of the screen is not affected but it seems a band of destruction passes through. This is very like the effect you get when you are NOPPing as the read comes through and contends with rendering the screen.
 
You notice that the whole of the screen is not affected but it seems a band of destruction passes through. This is very like the effect you get when you are NOPPing as the read comes through and contends with rendering the screen.
Nivag, maybe i have a bad video logic?
 
Nivag, maybe i have a bad video logic?
That is EXACTLY what you have... The CPU should not be doing anything, as it is physically removed...

The only real way to solve this problem (in my opinion) is to thorougly freeze each chip in turn using freezer spray. Whatever it is, it is clearly temperature dependent...

Warm the PET up and get the video problem to appear on the display.

Make a plan of logic chips that you want to freeze. Make a small box out of cardboard for the purpose of constraining the freezer spray to one chip at a time. Move the box around from chip to chip following the plan - and thoroughly freeze the chip under examination - looking for changes to the display.

If that doesn't work, move on to the next chip on the list and so forth.

I would test the following devices on schematic 8 in the order specified:

H11, G11, F9, F10, E11 and G9.

Since we also suspect G5 on schematic 6, I would do that as well.

Tick each one off as you go.

If it is none of those, then we move further back into the timing logic next.

Dave
 
That is EXACTLY what you have... The CPU should not be doing anything, as it is physically removed...

The only real way to solve this problem (in my opinion) is to thorougly freeze each chip in turn using freezer spray. Whatever it is, it is clearly temperature dependent...

Warm the PET up and get the video problem to appear on the display.

Make a plan of logic chips that you want to freeze. Make a small box out of cardboard for the purpose of constraining the freezer spray to one chip at a time. Move the box around from chip to chip following the plan - and thoroughly freeze the chip under examination - looking for changes to the display.

If that doesn't work, move on to the next chip on the list and so forth.

I would test the following devices on schematic 8 in the order specified:

H11, G11, F9, F10, E11 and G9.

Since we also suspect G5 on schematic 6, I would do that as well.

Tick each one off as you go.

If it is none of those, then we move further back into the timing logic next.

Dave
I am desperate really now!
I ran out of coolant spray....and I think the one I had before wasn't good, it left the board greasy! I bought it at a pharmacy.
 
Pharmacy spray is no good. It is not cold enough. It is cold enough to freeze things on your skin. You require proper electronics freezer spray. That may be why you didn't see any difference last time.

The only other thing we can do in the meantime is to see if we can see anything untoward with the oscilloscope. But we are now in territory where we are pushing your oscilloscope capability and your skill set.

Dave
 
That looks OK. It has the right words and pictures.

You might want to order a replacement G5 device as well. I am sure we are going to have to replace that - so be prepared (as the boy scout's moto states).

Dave
 
It could be any of the ones mentioned in #591...

I'm going to make a guess here... G11

Although I'm not sure earlier tests agreed with this I think that the correct character being rendered means that most of the early logic is OK... so I'm inclined towards G10, H10, G11... and I'm going to tend towards G11.

@Deperato If you move a slightly damp finger along the sides of G11 does anything change?
 
It is almost as though an IC input has become disconnected from the physical IC pin and is picking up 'noise' from other parts of the IC. However, this only occurs when the temperature increases.

It starts off OK (as indicated by the screenfull of stable 'E' characters), but then deteriorates as the temperature increases.

I suspect then that one input doesn't work as it is supposed to, and starts to pickup noise. As this noise is not synchronised to the screen clocking, it appears as a band of disturbance moving throughout the frame.

Interestingly, this shouldn't be isolated to the uppercase letters - but should equally apply to the numerals - yet that is not what we were observing. Of course, the fault may have morphed as the device has aged...

I hate these types of faults...

This would be where your fingerprinting device would absolutely come into its own Nivag!

Dave
 
Is your character generator in a socket? It looks like it.

Are you absolutely sure you don't have a 'contact' problem between the pins of the character generator and the socket?

Are you absolutely sure the soldering and PCB tracks are OK to and from the character generator socket?

Have you tried a replacement character generator?

Dave
 
Back
Top