• Please review our updated Terms and Rules here

CBM PET 3032 STRANGE BOOT

Check continuity between the socket of pin 39 of C5 to the actual leg of pin 19 of the character gen. I think its floating and picking up somehow.

Put the via back in C5, check from C5 let pin 29 to char gen pin 19.

Need to think about it.
 
I was just thinking tonight (yes, I know, it is always dangerous)...

We have 'gone for broke' and tried the entire video sub-system in one go; and it hasn't paid off. Possibly because we have more than one fault.

Have a look at what Gary has suggested first though - he has spotted exactly whatvI did before I went out...

Dave
 
Ok, so you might need to replace the socket of the VIA, but I would try a jumper from leg to leg first to make sure.

Do you have any IC test clips ? with those you could run a pin 39 to pin 19 bypass ?
 
Gary C said:
Ok, so you might need to replace the socket of the VIA, but I would try a jumper from leg to leg first to make sure.

Do you have any IC test clips ? with those you could run a pin 39 to pin 19 bypass ?
Click to expand...
Unfortunately i have not test clips Gary 😭
 
My guess here is that without the 6522 VIA in its socket, the floating length of track connected to the character ROM on its pin 19 is picking up stray signals. Obviously it doesn't do this in most PET boards of this model, or the fault would have been observed before when the VIA was out of its socket in other machines, so the possibilities are it is just a borderline condition. But the interesting thing is, that if it was just a susceptibility of that particular character ROM, the problem would have appeared to have gone away when Desperado fitted an eprom in the place of the character ROM, but the problem remained.

On that basis, it is suspicious that there is some sort of inter-track leakage on that length of track leading to pin 19, but it might not be physically possible to find it.

One thing that could be tried is to add a pull up resistor such as a 10k, from pin 19 of the character ROM to +5v, soldered on the back of the board, left there permanently. This won't interfere when the VIA is plugged in again later, but it would mean with the VIA out of its socket, probably the problem would go away. Or if you want the other " graphics case" connect the 10 k resistor to ground. But a pull-up resistor is probably preferable.
 
Last edited:
Its suspicious that he cant get continuity between the leg of pin 29 and pin 19 of the char gen. Sounds as if there is something wrong around that socket. Its possible its breaking down and almost touching an adjacent pin. Either way, I would replace the VIA's socket and see if I can at least get a connection between the output of the PIA and the graphic select of the char gen then even if there is some strange pickup, it should be driven by the VIA to the required state and held as the output of CB2 sources and sinks, ie it won't leave the line floating.


(NOTE 29, I made a mistake above in #729 when I said 39)
 
Last edited:
thanks a lot to all in the meantime, today I will start to change the socket of C5 and I will also check the card under it well.
 
Gary C said:
Its suspicious that he cant get continuity between the leg of pin 29 and pin 19 of the char gen. Sounds as if there is something wrong around that socket. Its possible its breaking down and almost touching an adjacent pin. Either way, I would replace the VIA's socket and see if I can at least get a connection between the output of the PIA and the graphic select of the char gen then even if there is some strange pickup, it should be driven by the VIA to the required state and held as the output of CB2 sources and sinks, ie it won't leave the line floating.


(NOTE 29, I made a mistake above in #729 when I said 39)
Click to expand...

It is an interesting scenario though. If you remove an IC from the computer, including a PIA or a VIA, that inevitably leads to some inputs on other IC's floating. Much of the time, if these inputs are to TTL devices, you "get away with that" because the nature of the input circuit architecture for TTL, is that they will assume a logic high, if left open.

But if the IC is a MOS IC of some kind, this is not definitely the case.

So, in all probability when it is suggested that PIA's and VIA's should be removed from the sockets, for fault finding applications, we need to think about this more carefully and probably apply tie resistors to the PIA and VIA output pins, or........we might be begging for trouble, just like this case.
 
Yes, removing devices can introduce further problems of course. But it helps to 'absolutely' know what the configuration of the machine is from test to test. Unfortunately, with the hundreds of posts, it is easy to get lost in the fine details.

Unfortunately, freezing E11 should not have had any effect if the problem was the floating address line on the character generator because of the missing VIA.

This begs the question as to whether E11 was really faulty in the first place - or did we have a problem with E11 that is now fixed, and now we have another concurrent problem?

The only sure way to know is to revert to our original test scenario (CPU removed, video RAM removed and video RAM outputs strapped to give us a screenfull of 'E' characters) and see if that is now stable. If it isn't, we can then try a pull-up resistor on the GRAPHICS signal.

Dave
 
If C5 is the VIA socket, then it seems logical (from what I have read) to replace it and THOROUGHLY check the PCB out.

Dave
 
As an interesting aside:

In early Neuroscience and Brain studies, to find out what part of the Human Brain did what, people who had lesions in parts of the Brain (from strokes, trauma, tumors etc) were studied. From that a rough map was created of Brain anatomy vs Brain function. Also, lesions were deliberately inflicted on experimental subjects, to find out what part of the Brain was performing specific functions.

This went on for many years until somebody realized the defective logic in it and remarked:

If an Alien came to Earth and started a study on the Internal Combustion Engine, knowing nothing about it in advance, they might wonder what a Piston was for. As an experiment, they would remove one from the Engine, creating a lesion in the Engine if you like. The Engine would then vibrate, and they would write in their report: "The purpose of the Piston is to stop the Engine vibrating".

It was realized that if a complex system is altered from its natural working state, all sorts of unexpected effects can occur, that can lead to false conclusions. So it is better, in the case of the computer, where possible, to leave the system in its original state, (not removing IC's for example) and fault find it with tools that create the most minimal electrical disturbance during testing (like x 10 probes).
 
On the expensive oscilloscopes at work I always use x10, because the oscilloscopes are 'smart enough' to know the probe is set to x10 and compensates on the display accordingly.

At home, I generally use x1 unless I am measuring clocks (or the waveform I am looking at doesn't 'look right') in which case I will switch to x10.

Horses for courses...

Continuing with your theme Hugo, I offer this observation:

If you train a spider to jump when you shout 'JUMP' and then pull the legs off one by one; you eventually get to the stage when the spider has no legs and doesn't jump when you shout 'JUMP'. Conclusion: a spider hears through it's legs...

Dave
 
Desperado said:
Need i change C5 socket or must i waiting?
Click to expand...
As you have reported that with the VIA in socket you have no continuity from the leg of pin 29 of the VIA to pin 19 of the char gen then that will leave the line floating.

Once the VIA output CB2 is connected to pin 19 of the char gen then it will be able to drive the line positively high or low.

So check pin 29 of the VIA and make sure its clean, maybe try to clean the socket and if you cant get continuity between the pin and the track then its time to change. You did say you could get continuity between the socket and the char gen so I'm assuming the tracks are ok and that maybe the socket has splayed and not gripping the pin.
 
You must log in or register to reply here.
Back
Top