• Please review our updated Terms and Rules here

Cbm 2001 Pet strange boot

a good friend of mine who worked years ago for Commodore support always told me that the original white pet sockets are a major cause of errors ...
Do you agree Dave?
 
This is another Pettester ram fault:
 

Attachments

  • Schermata 2022-05-15 alle 13.02.37.jpg
    Schermata 2022-05-15 alle 13.02.37.jpg
    63.9 KB · Views: 8
Dave_C78 said:
a good friend of mine who worked years ago for Commodore support always told me that the original white pet sockets are a major cause of errors ...
Do you agree Dave?
Click to expand...

I'm not sure what Dave will say, but I have found the white sockets on both my dynamic PET boards to be very good, but I have looked after them with lubrication.

One thing it pays to keep in mind; it is not a good idea to take IC's in and out of IC sockets over and over again without lubricating the sockets. If they are dry there can be significant wear. It doesn't take much lubricant, I do it by adding the lubrication (Inox's mx-3) to the IC's pins before they are plugged in and the lubricant transfers to the socket claws. This way you can take IC's in and out of the socket many times without damage to the socket. (mx-3 leaves a long lasting high purity oil behind and is better than any other product for this task)

If you want to check the socket claws, you can do it one by one on a test. It requires that a pin is removed from a dud IC and soldered to a small wire handle. You can also dip the pin in the lubricant. Then you pass it into the socket hole on the correct axis and you can feel the tension in the socket claws. Occasionally, a socket claw has been damaged by rough insertion of an IC, or somebody forcing pins in there that were thicker than those on an actual IC and you will be able to feel if the tension in the socket claw is lost.

Don't put any other kind of pin than an actual IC pin, or pin of identical geometry, into an IC socket, this includes the round pins on machine pin IC sockets. They stretch the socket claws apart and later they won't connect properly to an actual IC pin. If you are going to make any adapters for IC's, out of IC sockets, they must be made from the dual wipe type of socket, which has flat thin pins of the same geometry as an actual IC. Never make an adapter out of a machine pin socket, or the pcb socket you push that into, may not be suitable to take an IC again directly later.

It sounds like you need to stock up on some spare RAM IC's.
 
Last edited:
Hugo Holden said:
I'm not sure what Dave will say, but I have found the white sockets on both my dynamic PET boards to be very good, but I have looked after them with lubrication.

One thing it pays to keep in mind; it is not a good idea to take IC's in and out of IC sockets over and over again without lubricating the sockets. If they are dry there can be significant wear. It doesn't take much lubricant, I do it by adding the lubrication (Inox's mx-3) to the IC's pins before they are plugged in and the lubricant transfers to the socket claws. This way you can take IC's in and out of the socket many times without damage to the socket. (mx-3 leaves a long lasting high purity oil behind and is better than any other product for this task)

If you want to check the socket claws, you can do it one by one on a test. It requires that a pin is removed from a dud IC and soldered to a small wire handle. You can also dip the pin in the lubricant. Then you pass it into the socket hole on the correct axis and you can feel the tension in the socket claws. Occasionally, a socket claw has been damaged by rough insertion of an IC, or somebody forcing pins in there that were thicker than those on an actual IC and you will be able to feel if the tension in the socket claw is lost.

Don't put any other kind of pin than an actual IC pin, or pin of identical geometry, into an IC socket, this includes the round pins on machine pin IC sockets. They stretch the socket claws apart and later they won't connect properly to an actual IC pin. If you are going to make any adapters for IC's, out of IC sockets, they must be made from the dual wipe type of socket, which has flat thin pins of the same geometry as an actual IC. Never make an adapter out of a machine pin socket, or the pcb socket you push that into, may not be suitable to take an IC again directly later.

It sounds like you need to stock up on some spare RAM IC's.
Click to expand...
Thanks Hugo, i used some contact cleaner spray on the sockets....
 
I have a ram tester....maybe i can remove step by step the ram ics to test....but what are the ram ics of the lower bank?
 
If you look at the schematics you will observe that the low 16K bank of RAM is driven by the signal /CAS0 whereas the high 16K bank of RAM is driven by the signal /CAS1.

With this fact, you should now be able to identify which ICs are associated with the low 16K bank.

This (and other useful facts) I would write down in the back of my notebook so I can refer to them quickly in the future.

That last error indicates the lower bank of 16K RAM and data bit 1. Is this different to other memory error reports?

Of course, the problem may not be the RAM itself, but the RAM address multiplexers or data buffers or...

Dave
 
Dave_C78 said:
i haven't all these ram ics :(
Click to expand...

Remember that you only need eight working RAM chips for the dynamic to PET run. The jumpers do have to be set appropriately of course.

Alan

Edit: Added attachment
 

Attachments

  • 320351.pdf
    82.6 KB · Views: 5
Last edited:
daver2 said:
If you look at the schematics you will observe that the low 16K bank of RAM is driven by the signal /CAS0 whereas the high 16K bank of RAM is driven by the signal /CAS1.

With this fact, you should now be able to identify which ICs are associated with the low 16K bank.

This (and other useful facts) I would write down in the back of my notebook so I can refer to them quickly in the future.

That last error indicates the lower bank of 16K RAM and data bit 1. Is this different to other memory error reports?

Of course, the problem may not be the RAM itself, but the RAM address multiplexers or data buffers or...

Dave
Click to expand...
So what do you recommend to do? I remove one by one the low bank ram and test them with my ram tester or is there a way to understand if they are the buffers?

Thanks
 
With this type of fault I would use Nivag's ROMulan RAMulator first, with the PETs ROM and RAM mapped out. In this configuration, all we are using of the PET is the power supply, the video sub-system and the I/O devices.

If this is rock solid, the next thing would be to map the PETs ROM back in and see if that is rock solid. What you are left with then is the RAM to diagnose.

If Nivag's emulator doesn't run properly, it is not the ROM or RAM...

There are a wide range of tools out there, and you need to choose the right tool for the job. PETTESTER is good at detecting fairly 'hard' faults, although some 'soft' faults can be detected.

Dave
 
Dave, i tryied to run Pet with original ud9 rom and i have this fault :(
Strange strange..... :( I am desperate!
it would seem that there is something that heats up and crashes...

 
So, what you have working now is infinitely better than it was when you started this thread.

The behaviour is sort of expected given what I have observed with the PETTESTER. It was never likely to 'fix' the problem...

The usual way of identifying a thermal issue is with a hot air gun and a can of freezer spray.

That behaviour does (however) look PIA-related; but it could also be attributed to the I/O space in general - which is a block of memory space.

what we need to do is to collate all the data we have together and look for one orctwo components that are common to the sub-systems that are failing.

Dave
 
Note that the PET got to the machine language monitor by way of a 'Break' instruction. That means while at the ready prompt or in the interrupt handler, the CPU fetched an all zeros instruction.
That's telling us something, but what? Bad ROM? There are no transceivers between the CPU and ROMs.
 
dave_m said:
Note that the PET got to the machine language monitor by way of a 'Break' instruction. That means while at the ready prompt or in the interrupt handler, the CPU fetched an all zeros instruction.
That's telling us something, but what? Bad ROM? There are no transceivers between the CPU and ROMs.
Click to expand...
Hi Dave_m how are you?? Do you think that i have some bad rom?
 
daver2 said:
So, what you have working now is infinitely better than it was when you started this thread.

The behaviour is sort of expected given what I have observed with the PETTESTER. It was never likely to 'fix' the problem...

The usual way of identifying a thermal issue is with a hot air gun and a can of freezer spray.

That behaviour does (however) look PIA-related; but it could also be attributed to the I/O space in general - which is a block of memory space.

what we need to do is to collate all the data we have together and look for one orctwo components that are common to the sub-systems that are failing.

Dave
Click to expand...
If i leave Pet running, can i feel if some ic gets hot?
 
You must log in or register to reply here.
Back
Top