CBM Pet 4032 DOA

[mjnurney]

I spent a good few hours on the PET last night.

i checked every address lines from the roms and the are all good.

removed all the ram and replaced it.

ive spent some time examining the board , cleaning the new sockets and 9 times out 10 im greeted with the screen below and a short chirp

it probably means something to someone ?

Or I get a nice chirp and no screen at all...


update:

heres a funnily little quirk.

I have changed LS244 UB5 , because it went crazy when I piggy backed it and it constantly had hi readings on the logic probe.

Now... with pettester rom fitted , it initialises the screen quicker

oh and I tested the removed LS244 in my logic tester - it failed on 4 gates.

UB4 & UB5 are linked to the video ram / display generation. I changed both anyway

 

[dave_m]

ive spent some time examining the board , cleaning the new sockets and 9 times out 10 im greeted with the screen below and a short chirp

it probably means something to someone ?

Or I get a nice chirp and no screen at all...

Good work on the PET. The CRT display looks steady not not wavy.

What you have now is the classic Garbage Screen of uninitialized Video RAM at power up condition. It just means the processor ran in program just long enough to run the BEEP routine, but quickly got lost before it could clear the screen. Check the CPU SYNC signal (pin 7 on the 6502). Does it get stuck after a very brief period?

It seems we can not READ good instructions or READ/WRITE good zero page RAM for some reason.

I'm getting worried we are cycling the system too much. Let's attach a RESET switch to the /Reset line such that when you depress the pushbutton or toggle the switch, we ground the pin at expansion J4-22 or at UD15-pin 6. This will give us a good reset without cycling the power supply.

 

[mjnurney]

Good idea on the reset , I was thinking about do that last night but bed time called.

I'll check the sync in a min. Is this worse or better than being stuck at the ready prompt ? Sometimes the screen stays blank too but that could be that it freezes after clearing the screen?

But it's progress I guess

 

[mjnurney]

I hate to say it but I'm fascinated why this thing wont work

 

[mjnurney]

hi Dave

sync is low , the pet chirps then it flashes hi (garbage screen appears) and then goes low & high and pulses

 

[dave_m]

hi Dave

sync is low , the pet chirps then it flashes hi (garbage screen appears) and then goes low & high and pulses

If Sync is pulsing, at least it means the CPU is running somewhere (stuck in a tight loop?), just not running in program.

 

[dave_m]

I hate to say it but I'm fascinated why this thing wont work

This PET is fighting you, but when you finish, you will be an expert on repairing this computer.

 

[mjnurney]

hi Dave yes the sync is pulsing

 

[dave_m]

Off the cuff things to try in no special order:

Get into Monitor mode by holding down diagnostic input on power up.

Check the Interrupt Request line (IRQ) (CPU pin 4) for pulsing at 50 Hz. If not pulsing, it should be HIGH and not LOW.

Are any of the Data lines not pulsing.

Check for pulses on the ROM Select lines: SEL D, SEL E and SEL F on UE14 (74154 decoder)

Check for pulses on RAS0 (UA5 pin 4) and CAS0 (UA5 pin 15)

 

[mjnurney]

ok ill do it now , thanks

I cant get it in ti diagnostic input - is there a trick to it?

Blank screen.

irq=high
A11=high
a13=high
a15=high

UE12 encoder (if I read the schematic correctly? pin 14,15,16,17) ?
sel a=high
seb b=high
sel c =high
sel d=high
sel e=high/low and pulse
sel f=low
sel 9=high




ua5 pin 15 low pulse
ua5 pin 4 low pulse

I hope you can make sense of this ?

garbage screen

irq=high
A0 to A15=high

ue12
pins 14,15,16,17
14 sel c=high
15 sel d =high
16 sel e=high
17 sel f=low

the garbage screen is back on every reboot and its different every time - it wasn't before.

How does the Boot sequence work ?

How does the ROM selection work ? the SEL line is used to select the ROM i suppose? if that is the case i usually have the kernel and editor running and the last basic ROM but not the two other ROMs so it can never boot.
i may look in to the rom selection circuit and remove/test a few of the chips in that circuit. Unless you can suggest any other?

How it works now is that - if its a black screen i usually have no editor running or sometimes even kernel isn't running (if high on SEL is not what we want )
or i have a garbage screen with SEL pulsing on Kernel and editor and usually the last basic rom as well but I've never had SEL pulsing on the other basic ROMs except when its reset. then all SEL lines appear to pulse for one second and go high.

Sorry to be a pain

A bit of info I've just been told.


The select lines should be high most of the time unless that particular ROM is being accessed. If any of them get overly hot, it can be caused by more that one being active at a time. You could use a nop generator to test the functioning of the address buffers and the 74154 (http://blog.tynemouthsoftware.co.uk/2014/08/commodore-pet-repair-part-5-some-roms.html).
Garbage on the screen is normally the video RAM, but can also be caused by the data latch that feeds the video RAM, either not being activated or being faulty, or the latches coming out of the video RAM.

Only one ROM is accessed at a time, starting with the kernal ROM, then jumping to various points in the others as required, and ending up in the editor ROM. Only one device which is writing data to the data bus should be active at a time otherwise they conflict.

mike.

 

[dave_m]

Garbage on the screen is normally the video RAM...

Mike,
This is true only if the rest of the PET is running properly. When the PET is not running very long in program, the list of suspect parts or circuits gets longer.

With IRQ high and not pulsing at 50 Hz, we are not getting to the point of enabling Interrupts so going to the Monitor program will not help as the keyboard will never be scanned.

Perhaps you are right, it is time to build the NOP Generator to dynamically check the addressing circuitry with incrementing address lines. Believe it or not, the address circuits are the most complicated part of the PET. The 16 address lines are buffered and sent in a lot of directions. In the video memory area, the screen refresh address counters are multiplexed and latched so there are a lot of places to cause problems.

 

[mjnurney]

thanks Dave

i'll make one tonight probably.

i have done some more work on the board and i have IRQ low and SYNC low now. still a garbage screen but occasionally i get IRQ and something PC on screen.

I've ordered a ROM/RAM replacement board so that should help when it arrives.

mike.

 

[dave_m]

i have done some more work on the board and i have IRQ low and SYNC low now. still a garbage screen but occasionally i get IRQ and something PC on screen.

That's strange for the Interrupt to be stuck low (asserted). Remove the source of the 50Hz Interrupt Request UB12 (PIA) and see if we get the Commodore message. Of course without interrupts the keyboard scan and blinking cursor will not work.

 

[KC9UDX]

I hope you don't end up with what mind had: A short from /IRQ to ground that I couldn't find. I mean, it isn't to be found anywhere. So, I put in a smaller pullup resistor hoping the short will burn out someday. The system works in the meantime (long time now).

 

[mjnurney]

i don't know why it did it but it hasn't done it since , IRQ is pulsing as is Sync so thats good.

Ive had the CRT board out today and measured the caps and resistors and everything looks good. however the 25v 470uf cap measures fine but has a slight smell of what i could smell the other day - i'll change all the caps i think shortly.

No Commodore message yet but I'm sure there will be soon.

mike

Video RAM tested in a Sinclair ZX81 (timex to you guys)
main ram tested in an Apple II

so all ram is good and checked now.

 

[dave_m]

i don't know why it did it but it hasn't done it since , IRQ is pulsing as is Sync so thats good.


so all ram is good and checked now.

Yes almost running. Must be in a tight loop. Maybe check address lines for a clue where the problem is. Or ground the diagnostic line at J2- pin 5, or J9-pin 20 or UD5-pin2 and power up to get into the machine language monitor program.

 

[mjnurney]

Ok I'll try that. It's booting to a blank screen but it's clearing the instant garbage screen. Just not making it to the commodore bytes free

 

[KC9UDX]

Soon you may see this:

 

[mjnurney]

Apparently my devibe is not present , pins grounded for diagnostic



Sorry about the pic , mobile upload

 

[mjnurney]

One day I will have a Komondors!?@ 4.0 ladic