Commodore Pet 2001-8 Stopped booting to basic.

[RetroGadgetMan]

Hi all I am back with my Commodore Pet 2001-8 that we got up and running a few years back.
PCB 320008
It has sadly developed a new fault.
I recently replaced all the horrible original basic eprom sockets with new ones as they were temperamental and all was working fine until one day I got a thin horizontal
Line across the screen.
Powered down and it came back to the boot screen but went back to a thin line again and then to a blank screen.
Now the tape motor runs constantly and I have gone from a blank screen to a full screen of the same symbols. But this can change upon power down to a different set of symbols. I have added some pics.
I suspect another logic IC has failed.
I have inspected the board for bad solder joints as this pet has had a-lot of the original logic replaced by myself and all appears to be ok.
I have tried running with all pias and via removed but still the same.
I have also tried my Tynemouth ram rom substitute and even with all of its features enabled the Pet will not boot to basic.
I have inserted a NOP adapter and i can see the pet is not resetting upon boot.
Pin 40 on the cpu is stuck on high no reset. I used a logic probe.
Sometimes the signature activity light flickers and sometimes it just stays on.
I have added a momentary switch to pin 2 of the 555 timer and to C6 capacitor.
When I press the switch the red reset light momentarily blinks in the NOP adapter but the screen does not clear just stays the same.

Hoping someone can advise on where to start. Thanks in advance.

Frank

 

[daver2]

1. When you press (and hold) your added reset pushbutton, does pin 40 of the CPU gow LOW?

2. When you release the pushbutton, does pin 40 of the CPU go HIGH after about 0.5 to 1.0 seconds.

3. When you put the NOP generator into the machine, does pin 7 of the CPU (SYNC) pulse (indicating that the CPU is fetching and executing instructions)?

4. I assume that you have checked ALL of the +5V DC rails?

5. Have you checked for a 1 MHz clock on the CPU (pins 37 and 39).

This looks like a video circuit fault to me (so it is possible that the CPU is running), but I am concerned that you don't think so.

Have you downloaded a copy of my PETTESTER program, and are you able to burn it into an EPROM? My PETTESTER replaces the EDIT ROM.

Dave

 

[RetroGadgetMan]

Well would you know it the video ram does seem To be an issue. I swapped it out for a wrangler substitute and now get the usual garbage before reset.
Also I didn’t notice that one of the ceramic caps had a broken leg C7. So replaced with a new one.
I am now getting a reset but not booting with the onboard roms or ram substitute.
In fact the only way I can get a boot screen now is with the ram and basic 4 enabled on the Tynemouth with the 6520 Pia #1 removed.
If it is left in lace I get the IRQ information.

None of the other basics boot from the Tynemouth.
Also I am now getting everything displayed twice on the screen.
Very odd behaviour.
I have added a pic of what I think is the diagnostics Rom that you mentioned.
Been a few years since I used it so can’t remember if this is it.

 

[daver2]

The 'double screen' issue is one of the video RAM address lines that is stuck.

Either one of the higher-address line video address multiplexor devices or one of the high address counters within the video system. I would guess IC D4 (74LS157) would to be the place to look, signal SA9. Check the two INPUTS to this gate that drives SA9 on the output - especially pin 14. The output (SA9 itself) may be 'stuck' or only stuck when the multiplexer is in one 'way'.

The 'machine code monitor' display can be entered by either the external DIAG pin being pulled LOW or a faulty keyboard PIA.

No idea what H7 is a picture of... It may be my PETTESTER, it may not be... H7 is the high half of the Kernal ROM, but this is also where a modified version of my PETTESTER was assembled for.

Dave

 

[RetroGadgetMan]

Well would you know it the video ram does seem To be an issue. I swapped it out for a wrangler substitute and now get the usual garbage before reset.
Also I didn’t notice that one of the ceramic caps had a broken leg C7. So replaced with a new one.
I am now getting a reset but not booting with the onboard roms or ram substitute.
In fact the only way I can get a boot screen now is with the ram and basic 4 enabled on the Tynemouth with the 6520 Pia #1 removed.
If it is left in lace I get the IRQ information.

None of the other basics boot from the Tynemouth.
Also I am now getting everything displayed twice on the screen.
Very odd behaviour.
I have added a pic of what I think is the diagnostics Rom that you mentioned.
Been a few years since I used it so can’t remember if this is it.

The 'double screen' issue is one of the video RAM address lines that is stuck.

Either one of the higher-address line video address multiplexor devices or one of the high address counters within the video system. I would guess IC D4 (74LS157) would to be the place to look, signal SA9. Check the two INPUTS to this gate that drives SA9 on the output - especially pin 14. The output (SA9 itself) may be 'stuck' or only stuck when the multiplexer is in one 'way'.

The 'machine code monitor' display can be entered by either the external DIAG pin being pulled LOW or a faulty keyboard PIA.

No idea what H7 is a picture of... It may be my PETTESTER, it may not be... H7 is the high half of the Kernal ROM, but this is also where a modified version of my PETTESTER was assembled for.

Dave

I just checked old messages it is a pet Tester. I bought it from member Nivag.

 

[daver2]

The ROMulan ID gives it away as being from Nivag...

Try that instead. Don't forget to read the PETTESTER manual!

Dave

 

[Nivag Swerdna]

That H7 thing you have might be configured as lots of things... does it have a label on the bottom?

 

[RetroGadgetMan]

That H7 thing you have might be configured as lots of things... does it have a label on the bottom?

It does lol. Can’t believe I didn’t think to read the bottom.
I tried it with all my roms and ram substitute removed using the Tynemouth ram rom enable and it failed the memory test.
The video glitch was just a bad connection on the video ram.

 

[Nivag Swerdna]

OK. That should be the H7 version of PETTESTER aka PETTESTE2KF04.bin which is basically @daver2 PETTESTER bodged to fit at $F800 rather than $E000 and a jump table $FFFA and a Dummy IRQ routine. It's based on the 2019 vintage of PETTESTER.
Should work in H7, make sure you put it in the correct way around.

One thing to note is that most devices have flat pins (as do my gadgets)... if you are using replacements with turned pins (round) then they should be put in turned pin sockets for best compatibly. Personally I hate round pins with the exception of PCB interconnects.

Anyway... keep up the good work

PS
Obviously when using a RAM/ROM board you don't want that board substituting F800-FFFF for it to have an effect.

 

[RetroGadgetMan]

Well would you know it the video ram does seem To be an issue. I swapped it out for a wrangler substitute and now get the usual garbage before reset.
Also I didn’t notice that one of the ceramic caps had a broken leg C7. So replaced with a new one.
I am now getting a reset but not booting with the onboard roms or ram substitute.
In fact the only way I can get a boot screen now is with the ram and basic 4 enabled on the Tynemouth with the 6520 Pia #1 removed.
If it is left in lace I get the IRQ information.

None of the other basics boot from the Tynemouth.
Also I am now getting everything displayed twice on the screen.
Very odd behaviour.
I have added a pic of what I think is the diagnostics Rom that you mentioned.
Been a few years since I used it so can’t remember if this is it.

 

[RetroGadgetMan]

I have checked the four regulators and confirmed 5v.
Using the pet tester with this configuration it doesn’t work. I get a garbled readout from the pet tester pictured here.

 

[RetroGadgetMan]

With the following configuration the pet tester runs and I get these results.

 

[RetroGadgetMan]

And at the end of my testing the screen shrunk to a thin white line. But a power down and back on again to a full screen.
This thin line of information is what I had three years ago in the early stages of repair.

 

[daver2]

Post #11. If you read my PETTESTER documentation, this is NOT a garbled screen. It is telling you that you have RAM faults in page 0 (address range $0000 to $00FF)! The characters are displayed can be decoded (using the PETSCII tables) to identify which bit(s) or IC locations are potentially faulty.

Not sure what the configuration is in post #12. But I can see some issues whatever it is (ROM checksums and keyboard PIA)...

Post #13 could be a bad solder joint somewhere on the VDRIVE signal from the logic board to the monitor, or an intermittent fault in the vertical deflection circuitry of the monitor.

Dave

 

[MiaM]

One thing to note is that most devices have flat pins (as do my gadgets)... if you are using replacements with turned pins (round) then they should be put in turned pin sockets for best compatibly.

Also: In particular once you have used round pins in a socket that doesn't have round holes, that socket tends to only work reliable with round pins. I.E. if you insert round pins in a non-round-hole socket, you then have to have an additional round-hole socket in the non-round hole sockets to reliably use regular chips (with flat pins).

 

[Nivag Swerdna]

I'm not going to interfere too much here but... I would remove H1..H6 and the PIAs and VIAs and then just run PETTESTER from H7 in a divide and conquer style... add them in later.

I also notice that your buffers are socketed... that's a good thing!

 

[RetroGadgetMan]

It’s strange that I am getting ram issues with a modern ram substitute. The computer won’t do anything with the ram substitute only with the Tynemouth ram enabled.
Sometimes I get a boot screen using the rom ram substitute basic 4 but not basic 1 or 2 when pia#1 is removed but obviously no curser.
I was suspecting a bad connection somewhere but can’t see anything under a magnifier.
Something is pulling it down somewhere.

 

[daver2]

The basic problem is that you are trying to fault find and get the whole system going in one go, rather than taking it in steps to find the faulty sub-system(s) and fault find from there.

Use my PETTESTER...

If my PETTESTER won't even work, we have to strip the machine down as far as we can (remove ROMs, PIA, VIA etc.) and then look to use a NOP generator.

Once we can get my PETTESTER going, that will test the video sub-system, followed by page 0 and 1 RAM, followed by the ROM checksum and keyboard, followed by an exhaustive DRAM test.

If the Tynemouth ROM/RAM board only works with one version of BASIC, it indicates that something 'strange' is going on - as my PETTESTER does with just page 0 RAM...

As I have stated elsewhere, my PETTESTER was trying to indicate what was potentially wrong back in post #11.

I am away on a business trip for a few days, so only have limited access to my information until I get back home.

Dave

 

[Nivag Swerdna]

Recently I wrote some words...

https://forum.vcfed.org/index.php?t...lp-diagnosing-ram-issues.1254810/post-1470523

So it's not strange that PCB RAM and ROM/RAM board RAM behave differently... The PCB RAM needs the address decode and all the buffers to work correctly amongst other things. Similarly the video RAM requires the buffers and decoding to work correctly. A ROM/RAM board will hide a multitude of evils but can be a useful tool as you can turn things on and off selectively.

 

[RetroGadgetMan]

Well I found the culprit for the loss of the vertical. I followed it back on the circuit diagram to C6 then B6. I was naughty and cheated by piggy backing B6 74LS107 and got a full screen. I have now soldered in a socket and replaced.