• Please review our updated Terms and Rules here

CBM PET 3032 STRANGE BOOT

Right,

So, from my understanding (and I bow to Nivag's knowledge here) that the RAM is internal to the Nivag card (not the PET) and the ROM I am not sure...

The character 'twinkling' are the faults in the video circuitry that is still being used.

So, to go back to my original question, have you read Nivag's manual on the card and what is being tested?

Dave
 
>>> Yes i think that this board test ram and rom ics with its integrated directly.....

Nivag's board can do the following (it may do more, but this is the general gist of what I understand) depending upon the link settings:

1. A NOP generator.
2. It can be used to replace the PET's RAM - or use the PET's RAM.
3. It can be used to replace the PET's ROM - or use the PET's ROM.

When replacing the PET's ROM - it can either simulate BASIC (or a number of different BASICs) or can implement my PETTESTER code, or a number of other test routines.

It all depends upon how the blue links are configured.

You have taken a photograph of how the links are set - but what does the setting actually do?

Dave
 
>>> Yes i think that this board test ram and rom ics with its integrated directly.....

Nivag's board can do the following (it may do more, but this is the general gist of what I understand) depending upon the link settings:

1. A NOP generator.
2. It can be used to replace the PET's RAM - or use the PET's RAM.
3. It can be used to replace the PET's ROM - or use the PET's ROM.

When replacing the PET's ROM - it can either simulate BASIC (or a number of different BASICs) or can implement my PETTESTER code, or a number of other test routines.

It all depends upon how the blue links are configured.

You have taken a photograph of how the links are set - but what does the setting actually do?

Dave
Links are setting like in picture Dave
 
Yes, but I don't know what those link settings actually do! I can make an educated guess at some of them - but there is documentation to tell us what the links actually do.

Dave
 
Hi. Sorry I have not been on this forum for a few days as my attention is elsewhere...

So I think the aim should be to get the tests to run for a prolonged period... If they start to run and then things wobble or fail then we need to establish the cause.

So... What happens if you just leave it on for say 4 hours?
 
So, your default configuration of Nivag's board is that the RAM that is being tested is on Nivag's board, and the PETTESTER code is running out of Nivag's board (not the PET ROM), so the only thing that is being tested is the video circuitry (and that is exhibiting a fault).

Dave
 
So, your default configuration of Nivag's board is that the RAM that is being tested is on Nivag's board, and the PETTESTER code is running out of Nivag's board (not the PET ROM), so the only thing that is being tested is the video circuitry (and that is exhibiting a fault).

Dave
That's right. I ship it in a configuration that has a good chance of working. The onboard RAM is enabled by the RAM jumper being in the ON position.

1678873354442.png

which it appears to be. You should have 32K of happy RAM with that state.

PS
Just move the jumper and press the RESET button to go to PET RAM, but first need to have it rock solid in this configuration.
 
So, back to the video circuitry again - we have to work out where the 'sparklies' are coming from.

Again, you need to use your oscilloscope and look for signals around the video RAM (F7 and F8) that are 'randomly' changing intermixed with the 'normal' signals. There is no other substitution for the oscilloscope and your brain...

Get it to the PETTESTER's DRAM test screen (where the video signals should be relatively stable) and poke around F7, F8 and then E11 pin 9 (the video signal from the parallel to serial shift register).

I would suggest using the VERT DRIVE signal J7/3 or G10/11 as a trigger on the second channel of your oscilloscope. This should (hopefully) give you a stable signal - and we are looking for signals fairly close to the VERT DRIVE signal activity (video pixels at the top of the screen).

Does you oscilloscope have a x10 timebase setting? Try without this first - and then with it (if you can't observe any 'sparkiles') and use the X-shift control to scroll along the trace (if your oscilloscope works in the way I think it should).

Dave
 
Last edited:
The video is in post #480.

This is with your RAM and my PETTESTER code working on your nice board.

There are some 'disturbances in the force' that I am putting down to an IC (or two) that is marginal - or starting to fail - in the video circuity somewhere.

Previously, we removed the video RAM and used bits of wire to 'hardwire' a fixed character to appear on the output of the what would have been the data lines of the video RAM. We did this with the 6502 CPU removed - thus guaranteeing no video RAM accesses. This is a good diagnostic trick in my book...

When we did that, we appeared to get a stable display, thus demonstrating that the latch, character generator, parallel to serial shift register and the remaining TTL logic after that to the monitor were 'good' - and that the problem must be occurring either with the video RAM itself or ahead of it (i.e. video RAM address multiplexers or the state machine generating the SAn signals for displaying the characters).

Another possibility (when we tried the 'bits of wire trick') is that we didn't leave it long enough (if it is a temperature-dependent effect) and so we are confusing ourselves. Perhaps we ought to repeat this test and see what happens over a longer period of time?

I hope this makes sense to condense 490 posts into the current position?

Dave
 
That's right. I ship it in a configuration that has a good chance of working. The onboard RAM is enabled by the RAM jumper being in the ON position.

View attachment 1254195

which it appears to be. You should have 32K of happy RAM with that state.

PS
Just move the jumper and press the RESET button to go to PET RAM, but first need to have it rock solid in this configuration.
Nivag thanks so much for this useful board!

With the boar left in this configuration, the computer no longer freezes. However, after about 5 minutes, the characters start shaking or moving...there is something in the video system that fails when it warms up I think...
 
The video is in post #480.

This is with your RAM and my PETTESTER code working on your nice board.

There are some 'disturbances in the force' that I am putting down to an IC (or two) that is marginal - or starting to fail - in the video circuity somewhere.

Previously, we removed the video RAM and used bits of wire to 'hardwire' a fixed character to appear on the output of the what would have been the data lines of the video RAM. We did this with the 6502 CPU removed - thus guaranteeing no video RAM accesses. This is a good diagnostic trick in my book...

When we did that, we appeared to get a stable display, thus demonstrating that the latch, character generator, parallel to serial shift register and the remaining TTL logic after that to the monitor were 'good' - and that the problem must be occurring either with the video RAM itself or ahead of it (i.e. video RAM address multiplexers or the state machine generating the SAn signals for displaying the characters).

Another possibility (when we tried the 'bits of wire trick') is that we didn't leave it long enough (if it is a temperature-dependent effect) and so we are confusing ourselves. Perhaps we ought to repeat this test and see what happens over a longer period of time?

I hope this makes sense to condense 490 posts into the current position?

Dave
In your opinion why with the Nivag card the computer no longer freezes? Is there a problem with ram or rom?
 
I often use a similar board. If it works with RAM & ROM mapped out, its a simple task of changing the jumpers to map ROM in then RAM in and see what happens.

Looking at the board, the top left link maps the PET RAM in or out. Moving the link from middle to top to middle to bottom should put the PET ram back in circuit (this can be done with the power on, then press reset on the board). If it works, the RAM is ok, if it doesn't you have a problem with the RAM.

Same for the ROMS, in the middle are 4 links for various memory location, set them to middle to bottom to enable the PET's ROMS (just remember where they were and what works before you move them)

The next row of links I believe select which ROM image you use. Leave that one be for now.
 
Last edited:
>>> However, after about 5 minutes, the characters start shaking or moving...there is something in the video system that fails when it warms up I think...

I have explained exactly this in my posts #490 and #492 if you read them...

>>> Is there a problem with ram or rom?

One or the other (or both).

As Gary has stated, I would recommend changing the RAM link on Nivag's board (so you are using the PETTESTER on Nivag's board and the PETs DRAM) and see if that fails. We can then work on why the DRAM is failing (safe in the knowledge that we know that the ROM is good - as it is running from Nivag's card).

However, I would look at fixing the video fault FIRST.

Dave
 
However, I would look at fixing the video fault FIRST.

Dave
I agree, as I mentioned before that is the fault you can actually see, in the video, so it should be possible to trace backwards with the scope, through the video generator circuit, starting at the parallel to serial shift register, to find out where and why that problem is happening. It does always pay to fix any visible faults as the first order of business.
 
Video First.

We need to be clear on the symptoms. I would like to see a video. Moving, Shaking and Sparkles are all different things.

I can think of reasons for some of those but we REALLY need to know what we are looking at!

Ignore... all other issues for now... Video first.

"However, I would look at fixing the video fault FIRST."... seconded.
 
I don't know why but the video was removed from the dailymotion.com site..... as soon as I get back from work I'll redo the video....
 
Back
Top