Desperado
Veteran Member
- Joined
- Nov 25, 2017
- Messages
- 6,827
Yes...recently I tried to change the original white socket with this new black one... I don't have a spare rom but I can try to write an eprom!
- VCF South West - June 14 - 16, Davidson-Gundy Alumni Center at University of Texas at Dallas
- VCF West - Aug 2 - 3, Computer History Museum, Mountain View, CA
- VCF Midwest - Sept 7 - 8 2024, Schaumburg, IL
- VCF SoCal - Mid February 2025, Location TBD, Southern CA
- VCF East - April 2025, Infoage Museum, Wall NJ
-
Please review our updated Terms and Rules here
CBM PET 3032 STRANGE BOOT
- Thread starter Desperado
- Start date
Hugo Holden
Veteran Member
Isn't the character generator ROM on this board IC F10 ?
Don't forget if you replace it with an eprom to disconnect its pin 21 and tie it to +5V (pin 24) or in some cases it can drag down the /INIT line to an indeterminate logic value.
One way to cool off individual IC's is just to apply contact cleaner via a saturated cue tip to their top surface, and keep applying it as it evaporates quickly. The latent heat of evaporation significantly cools the IC. This is better in some ways than freeze spray because it is more targeted. It also works on other pcb parts, like transistors, diodes etc to check for temperature related problems. Dentists use a similar trick to find cold sensitive teeth.
Don't forget if you replace it with an eprom to disconnect its pin 21 and tie it to +5V (pin 24) or in some cases it can drag down the /INIT line to an indeterminate logic value.
One way to cool off individual IC's is just to apply contact cleaner via a saturated cue tip to their top surface, and keep applying it as it evaporates quickly. The latent heat of evaporation significantly cools the IC. This is better in some ways than freeze spray because it is more targeted. It also works on other pcb parts, like transistors, diodes etc to check for temperature related problems. Dentists use a similar trick to find cold sensitive teeth.
daver2
10k Member
This is why I suggested to @Desperado to make a small containment box out of cardboard to ensure that the freezer spray is affecting only where we are targeting.
Yes, F10 is the character generator ROM.
Dave
Yes, F10 is the character generator ROM.
Dave
Nivag Swerdna
Veteran Member
daver2
10k Member
Desperado
Veteran Member
- Joined
- Nov 25, 2017
- Messages
- 6,827
There's news...
I wrote a 2716 eprom and inserted it in place of the original rom characters. I tried for some time the full screen of "E" and nothing stuck. I then removed the pieces of cable and reinserted the original video ram, I also inserted the cpu on top of the Nivag card. I launched the pettester and it seems to be running fine..however, if I try to touch the character eprom with my finger, the problem reappears!! I am desperate!!
I wrote a 2716 eprom and inserted it in place of the original rom characters. I tried for some time the full screen of "E" and nothing stuck. I then removed the pieces of cable and reinserted the original video ram, I also inserted the cpu on top of the Nivag card. I launched the pettester and it seems to be running fine..however, if I try to touch the character eprom with my finger, the problem reappears!! I am desperate!!
daver2
10k Member
Desperado
Veteran Member
- Joined
- Nov 25, 2017
- Messages
- 6,827
no...i inserted eprom only....
daver2
10k Member
Go back and read post #602 then!
If that doesn't fix it - you have a faulty solder joint or PCB track somewhere between UF9, UF10 and UE11.
Bringing your finger close by causes an increase in the 50 Hz (or 60 Hz) mains noise and this is being coupled to an IC.
Dave
If that doesn't fix it - you have a faulty solder joint or PCB track somewhere between UF9, UF10 and UE11.
Bringing your finger close by causes an increase in the 50 Hz (or 60 Hz) mains noise and this is being coupled to an IC.
Dave
Desperado
Veteran Member
- Joined
- Nov 25, 2017
- Messages
- 6,827
Ok, tomorrow I'll double check all soldering and connections...Thanks so much
Nivag Swerdna
Veteran Member
Of course the purists would say that waving your finger at a circuit was a really mad idea. In fact the lack of an earthing strap would give some people cold sweats. 
I'm wondering what @Hugo Holden is thinking about the observed behaviour? The finger test is more suited to finding broken output stages but here we have a 'new' EPROM...
Maybe time to bite the bullet and just shotgun the shift register?
I'm wondering what @Hugo Holden is thinking about the observed behaviour? The finger test is more suited to finding broken output stages but here we have a 'new' EPROM...
Maybe time to bite the bullet and just shotgun the shift register?
daver2
10k Member
Let's just take stock of what we are doing for a minute or two...
Is what you are saying now that we largely have a screen full of 'E's, but when you bring your finger near we get our interference?
Replacing the ROM with an EPROM has changed the dynamics of the circuit slightly - and this probably accounts for the observed slightly different behaviour.
I am going back to the idea of putting the oscilloscope probe on the VIDEO output of G11 pin 6 (J7 pin 1) and using the HDRIVE as the timebase trigger signal (with the oscilloscope in CHOP mode). This should give us a stable image on the screen (assuming there is a load of 'E' characters displayed.
If you then bring your finger close to the area of the board - and you get a displayed image that is now different; you MUST be able to observe something different on the oscilloscope screen. It may be subtle - but you should observe a change.
I still don't like what happened when you measured E11 pin 2 (CLK8) with your oscilloscope. Short of replacing G5 (the source of CLK8) I don't know what we can do about that.
It would be worth (tomorrow) inspecting the top and bottom of the PCB around all of the ICs immediately adjacent to the character generator ROM (F10). All of the solder joins and all of the visible PCB tracking for bad solder joints, potential 'breaks' in the track and contamination on the surface of the board.
I would also use a multimeter (with the power off) set to a low resistance range to check for continuity between every pin of the ICs we are looking at and where they are supposed to be connected to. Go directly on the pin of the IC to the pin of the IC. Do NOT press too hard - otherwise you may make a connection where (originally) there wasn't a connection.
Look very carefully at the pins of each and every IC around F10 looking for signs of corrosion. On an Apple IIe disk controller, the pins of some of the ICs were completely eaten through. They looked like they were intact, but the pin of the IC itself was not making contact with the pin you could see (and hence the PCB). Fortunately, in this case, the ICs were in sockets. When I removed the IC, a couple of pins got left behind! I had already deduced that the IC pin was not making contact - but this confirmed it!
Bright light (preferably daylight), magnifying glass and printout the schematic. Tick each end of each pin as your test it (to make sure you don't miss any). Don't forget that the schematics do not always show the +5V (VCC) and 0V (GND) pins. Look on the data sheet to ensure you have captured ALL of the pins. Don't miss any. The pin you miss will be the culprit...
Dave
Is what you are saying now that we largely have a screen full of 'E's, but when you bring your finger near we get our interference?
Replacing the ROM with an EPROM has changed the dynamics of the circuit slightly - and this probably accounts for the observed slightly different behaviour.
I am going back to the idea of putting the oscilloscope probe on the VIDEO output of G11 pin 6 (J7 pin 1) and using the HDRIVE as the timebase trigger signal (with the oscilloscope in CHOP mode). This should give us a stable image on the screen (assuming there is a load of 'E' characters displayed.
If you then bring your finger close to the area of the board - and you get a displayed image that is now different; you MUST be able to observe something different on the oscilloscope screen. It may be subtle - but you should observe a change.
I still don't like what happened when you measured E11 pin 2 (CLK8) with your oscilloscope. Short of replacing G5 (the source of CLK8) I don't know what we can do about that.
It would be worth (tomorrow) inspecting the top and bottom of the PCB around all of the ICs immediately adjacent to the character generator ROM (F10). All of the solder joins and all of the visible PCB tracking for bad solder joints, potential 'breaks' in the track and contamination on the surface of the board.
I would also use a multimeter (with the power off) set to a low resistance range to check for continuity between every pin of the ICs we are looking at and where they are supposed to be connected to. Go directly on the pin of the IC to the pin of the IC. Do NOT press too hard - otherwise you may make a connection where (originally) there wasn't a connection.
Look very carefully at the pins of each and every IC around F10 looking for signs of corrosion. On an Apple IIe disk controller, the pins of some of the ICs were completely eaten through. They looked like they were intact, but the pin of the IC itself was not making contact with the pin you could see (and hence the PCB). Fortunately, in this case, the ICs were in sockets. When I removed the IC, a couple of pins got left behind! I had already deduced that the IC pin was not making contact - but this confirmed it!
Bright light (preferably daylight), magnifying glass and printout the schematic. Tick each end of each pin as your test it (to make sure you don't miss any). Don't forget that the schematics do not always show the +5V (VCC) and 0V (GND) pins. Look on the data sheet to ensure you have captured ALL of the pins. Don't miss any. The pin you miss will be the culprit...
Dave
Hugo Holden
Veteran Member
If we go with the theory that what we are seeing represents an open circuit gate input, and signals from hand capacitance are being coupled to it then there are at least 4 ways this could come about and we then would have to guess perhaps which one would be more likely:
1) there is a broken track or bad connection between IC's
2) there is a failing input connection inside an IC body
3) The output stage of an IC is going open circuit.
4) a complex failure inside the IC die.
In cases where an input is left floating, at least for TTL, it simply assumes a logic high, though it is less noise immune than if tied by a resistor to +5v. But, if there was a long length of track connected to it, it could pick up enough signal by stray coupling to activate the gate. Especially if the output stage of an IC driving an input was defective to the extent it kept it just near the gate's threshold voltage.
If a gate's input connection, to the pin on the IC package goes open circuit (I have seen this before a few times) the gate inside the IC acts like it is tied high. It would certainly have a very "short antenna" and not likely to be able to activate the gate, unless there was a more complex failure and some sort of cross coupling in the IC die.
So overall I think there must be an IC with a failing output stage, leaving a length of track connected to an IC's input pin partially open circuit condition and behaving like an antenna.
All TTL output pins should be able to drive the 30pF capacitance of a x10 scope probe, so I would look for any change in the fault, when the scope connects to an output pin on the IC body, and if it does, suspect that IC, being careful not to get confused where the presence of the hand and probe in the area causes a change in the fault. So I would go back and see what effect the scope has on output pin G5 pin 3, and if it is affecting that, suspect that IC and test all the other output pins on the other ICs with the scope.
1) there is a broken track or bad connection between IC's
2) there is a failing input connection inside an IC body
3) The output stage of an IC is going open circuit.
4) a complex failure inside the IC die.
In cases where an input is left floating, at least for TTL, it simply assumes a logic high, though it is less noise immune than if tied by a resistor to +5v. But, if there was a long length of track connected to it, it could pick up enough signal by stray coupling to activate the gate. Especially if the output stage of an IC driving an input was defective to the extent it kept it just near the gate's threshold voltage.
If a gate's input connection, to the pin on the IC package goes open circuit (I have seen this before a few times) the gate inside the IC acts like it is tied high. It would certainly have a very "short antenna" and not likely to be able to activate the gate, unless there was a more complex failure and some sort of cross coupling in the IC die.
So overall I think there must be an IC with a failing output stage, leaving a length of track connected to an IC's input pin partially open circuit condition and behaving like an antenna.
All TTL output pins should be able to drive the 30pF capacitance of a x10 scope probe, so I would look for any change in the fault, when the scope connects to an output pin on the IC body, and if it does, suspect that IC, being careful not to get confused where the presence of the hand and probe in the area causes a change in the fault. So I would go back and see what effect the scope has on output pin G5 pin 3, and if it is affecting that, suspect that IC and test all the other output pins on the other ICs with the scope.
daver2
10k Member
Let's get rid of the desperation then...
1. Put the original character generator ROM back in.
2. Do we get a screenful of stable 'E' characters again or not (with your fingers out of the way)?
3. If you let the machine warm up, does the screen start to change from stable 'E' characters into a mess?
4. Make a little box out of cardboard just larger than an IC.
5. Go around the various IC's as I have stated in post #591 and spray each device in turn - shielding the others with the little cardboard box.
6. Once you have frozen the IC, remove the can of freezer (and your hands)...
7. Has it made a difference to the screen display or not?
8. Spray the next device in the list and so on...
Dave
1. Put the original character generator ROM back in.
2. Do we get a screenful of stable 'E' characters again or not (with your fingers out of the way)?
3. If you let the machine warm up, does the screen start to change from stable 'E' characters into a mess?
4. Make a little box out of cardboard just larger than an IC.
5. Go around the various IC's as I have stated in post #591 and spray each device in turn - shielding the others with the little cardboard box.
6. Once you have frozen the IC, remove the can of freezer (and your hands)...
7. Has it made a difference to the screen display or not?
8. Spray the next device in the list and so on...
Dave
Desperado
Veteran Member
- Joined
- Nov 25, 2017
- Messages
- 6,827
right now I no longer have the cables in the video ram sockets... yesterday I inserted the video ram and the cpu on the diagnostic card... I have to go back to the old "E" configuration or I can also carry out the test with the pettester ?