CBM PET 3032 STRANGE BOOT

[daver2]

Ok, so see if it 'crashes'...

So you have already changed F3, F5 and F6 - so that rules those devices out.

What we are left with is either a bad connection somewhere (still) or we haven't waited long enough when we removed the video RAMs (and used pieces of wire to force known characters onto the screen).

If you use your oscilloscope probe on the pins of F9, does it make the problem better or worse?

If you have a mostly static screen (as you do with the PETTESTER DRAM test) you should be able to observe the 'noise' if you know what to look for. Unfortunately, this takes experience that you may not have, but give it a go...

Look at the address lines of the character generator (F10) to see if there is a regular pattern superimposed with a bit of noise.

Dave

 

[Desperado]

Ok, so see if it 'crashes'...

Yes...crash after 3 o 4 minutes....
Time to waiting for Diagnostic board...

 

[Desperado]

If you use your oscilloscope probe on the pins of F9, does it make the problem better or worse?

same....

 

[daver2]

Yes...crash after 3 o 4 minutes....
Time to waiting for Diagnostic board...

You need to be careful of the terminology here...

Did the machine 'crash' or did something happen to the video display characters?

Check the CPU SYNC pin (pin 7) and the outputs from D2 to see what is going on with the CPU.

>>> Look at the address lines of the character generator (F10) to see if there is a regular pattern superimposed with a bit of noise.

And the above?

Dave

 

[Desperado]

it's impossible to find a damage with this machine...every time it's always different... I hope the Nivag card can help me

Time to change ls373 on F9 maybe?

 

[daver2]

That device should have been tested as OK ages ago when we were poking wires into the video RAM sockets... That's why I did that test.

>>> Look at the address lines of the character generator (F10) to see if there is a regular pattern superimposed with a bit of noise.

Have you tried the above?

Dave

 

[Desperado]

That device should have been tested as OK ages ago when we were poking wires into the video RAM sockets... That's why I did that test.

>>> Look at the address lines of the character generator (F10) to see if there is a regular pattern superimposed with a bit of noise.

Have you tried the above?

Dave

I am desperate!! How can i do to test f10 address lines?

 

[daver2]

Put your oscilloscope probe on each address line of F10 and look at the trace for signs of instability...

Look at the schematic for the pin numbers of the address lines!

You should be able to understand this process by now without me specifying to you the pin numbers...

Every ROM has address lines 'going in' and data lines 'coming out'.

Dave

 

[Hugo Holden]

Since there is no fault in the video SRAM IC's (which surprises me), it raises the possibility that both of the faults are caused by the same problem. If there is some intermittent noise or high frequency instability on the data and/or address lines, it probably would only be a matter of time before the CPU got sent to no man's s land and stopped executing code. Probably when the cause of this is found, both faults will go away. Since the random changes are present on the screen characters, at least that is visible, I would focus on finding the cause of that, if you can track down the cause, the other problem might go away too. Sometimes the 5V voltage regulators can go unstable at high frequencies if the capacitors on their outputs are not in good order. It could pay to tack some additional ones in parallel as an experiment.

 

[Desperado]

Since there is no fault in the video SRAM IC's (which surprises me), it raises the possibility that both of the faults are caused by the same problem. If there is some intermittent noise or high frequency instability on the data and/or address lines, it probably would only be a matter of time before the CPU got sent to no man's s land and stopped executing code. Probably when the cause of this is found, both faults will go away. Since the random changes are present on the screen characters, at least that is visible, I would focus on finding the cause of that, if you can track down the cause, the other problem might go away too. Sometimes the 5V voltage regulators can go unstable at high frequencies if the capacitors on their outputs are not in good order. It could pay to tack some additional ones in parallel as an experiment.

Hi Hugo, yes i can try this...i need to looking for these little capacitors so...Thanks

 

[Desperado]

Put your oscilloscope probe on each address line of F10 and look at the trace for signs of instability...

Dave, A0 to A10? Correct?

 

[daver2]

I think so, but I am just about to go out so I can't check no until tomorrow.

Dave

 

[dave_m]

Hi Desperado,
Yes on the character ROM itself, and on the schematic wires they are named RA7,8 and 9 and then LSD0 through LSD6.
ROM address A10 is connected to 'Graphic' which I think should be always HIGH after power up on the 3032.
-dave_m

 

[Desperado]

I am desperate!!! i'll waiting for Nivag s' board at this point

 

[daver2]

Well, none of us can speed up the postal system I am afraid...

Did you look at the character generator address lines with your oscilloscope?

Dave

 

[Desperado]

Well, none of us can speed up the postal system I am afraid...

Did you look at the character generator address lines with your oscilloscope?

Dave

yes but I don't understand what I have to observe...maybe I can try to change the character rom?

 

[daver2]

>>> maybe I can try to change the character rom?

That's not why we are looking at the address lines. But, if it is in a socket (and you have a spare to exchange it with), please do - to rule out something else.

>>> yes but I don't understand what I have to observe...

We are looking for something out of the normal. Unfortunately, if you don't know what normal is supposed to look like, you have a problem...

Do you have another (working) machine of the same type that you can use as a reference machine?

If not, you will just have to "look and hope" and see if anything looks 'funny' to you... We are probably looking for short pulses that sometimes occur but mainly are not there.

This is another reason to work out what should be there from a working machine - i.e. you need to do some homework and document how the machine works and what you observe on the oscilloscope on a working machine (e.g. After repair). Sorry, no shortcut for hard work yourself. You can share your results with the VCFED community then!

I have never repaired one of these machines myself either - so I am in the dark regarding what you should observe...

Dave

 

[Nivag Swerdna]

I feel likely to be drawn into this thread soon... and to be honest I don't really understand all the previous 21 pages so...

My interpretation is that it now has multiple faults and is progressively getting worse? Am I wrong?

Please take a photo of the motherboard so we can see the condition of everything. Please make it in focus and of reasonable resolution.

Describe what is currently installed and what is not and what the behaviour is,... again with photos.

Finally, measure the power supply voltages on both DC and AC settings.

 

[dave_m]

I feel likely to be drawn into this thread soon...

Yes, I can feel the suction on you even from California!
Hopefully, isolating the ROM and BASIC RAM with your gadget will help separate the issues and indicate some direction to proceed. Daver2 has been heroic so far.

 

[Nivag Swerdna]

Yes, I can feel the suction on you even from California!
Hopefully, isolating the ROM and BASIC RAM with your gadget will help separate the issues and indicate some direction to proceed. Daver2 has been heroic so far.

I see my gadget is with customs at "MILANO ROSERIO CSI".

Given recent history of my shipments with the EU, particularly with Germany, it might never make it through customs. I have been seriously considering black listing any EU country from my tindie store as it is not worth the hassle.