Commodore PET CBM 4016 ASSY 8032080 RIFA burn and more

[DrAlis]

Using the momentum swiftly to return to the black screen 4016 with pettester i get the attached 3 video signals. hsync vsync video

 

[DrAlis]

But your other 16K RAM is also faulty though!

If you have BASIC up and running, you could try POKEing and PEEKing memory above the 16K limit ((16*1024)+1) to see what is wrong...

EDIT: If your keyboard is faulty, ignore what I just said...

Now you know how to fault find the RAM, you could swap the banks back to check the faulty bank...

Dave

with banks unswitched, we had a different screen that did not allow you to find any particular ic if i remember correctly

 

[DrAlis]

so the first b (pre bank switch) was second position so 01 in hex. thereafter, i fail to follow your explanation with the @ sign etc how to arrive at the broken ram...

4032:

View attachment 1278676

 

[daver2]

D0 and D1 are both stuck at '0'.

PETTESTER fills the memory with patterns from $00 to $FF.

The first 256 characters on the screen indicate 'g' or 'b' indicating whether the correct pattern was found when the memory was read or not.

The next 256 bytes on the screen indicate a '.' if a 'g' was displayed (in the first 256 bytes), otherwise (if 'b' was displayed) the PETSCII character that was read.

gbbbgbbbgbbb indicates that:

00 is good.
01 is bad.
02 is bad.
03 is bad.
04 is good.
etc.

Looking at the PETSCII characters indicates:

.@@@.ddd.hhh.

The '.' is displayed because the memory cell was good when read.

The PETSCII for '@' is $00.
The PETSCII for 'd' is $04.
The PETSCII for 'h' is $08.

Hence, I have deduced that D0 and D1 are both stuck at '0' (to start with).

Dave.

 

[DrAlis]

thanks that made it much clearer. still likely too tough for me to diagnose since i am not familiar with petscii and the memtest algorithm. well, great I got you here thanks.

so it seems at least two drams are broken / stuck 0 (is that low voltage or high voltage btw). will these correspond to to ua0 and ua1 on the universal pet pcb or was that pure luck? i do have a pet microsd to load some diagnostic programs. and i could switch the keyboard around also if that helps.

do the video signals look ok to you? if just realized some are upside down. i flipped them now. hsync and vsync have something. video is blank as far as my litte understanding goes. thanks

 

[daver2]

All of the information is in my PETTESTER manual...

Stuck '0' = 0 Volts.

Where do you get UA0 and UA1 from?

I see a good HDRIVE signal.

I see the correct VDRIVE frequency, but the voltage levels do not look correct.

The VIDEO signal is either fully black or white (depending upon which monitor is connected.

When measuring TTL signals, please set your oscilloscope to 1 V/div and ensure we know where the 0V level is!

Otherwise, we get ourselves confused as to what us a 'healthy' TTL signal and what is 'noise' scaled to fit the screen...

Dave

 

[DrAlis]

4032: I was looking at what is printed on the PCB re UA0 and UA1. But it turns out the lowest number is UA4 for a DRAM. Even and Odd are the banks as I can tell so only odd x in UAx are defect candidates.

4016: screen all black. signals hsync vsync again rescaled

 

[daver2]

HDRIVE and VDRIVE look OK.

You need to work your way back from the VIDEO signal (using the schematic) to see where the video signal appears and disappears...

D0 is either UA18 or UA19.
D1 is either UA16 or UA17.

Depending upon which bank of RAM is the lower bank.

Incidentally, it is unlikely that the piggybacking technique will work for a stuck '0' bit.

Dave

 

[DrAlis]

As @daver2 and @Hugo Holden and some others invested their valuable time into helping me, I owe all another update. So I found someone who can actually solder properly and got all DRAMs removed and sockets put in. Yeah! Ok so the outcome is that the

4032: is completely healed and will live happily on in my collection until a random new defect will occure some day. Thanks!

4016: still stays dark despite the new RAM. I might take another debug round if Dave feels like it. Overwise I will try to find a new home for it or best equip it with a MicroPET.

Thanks all for now it has been a lot of 100% fun, 100% of learning and also 50% success!

 

[daver2]

I am game for another crack at the 4016. Perhaps we can start again at the beginning to make sure we are not being derailed by false assumptions from last time.

PETTESTER in place of the EDIT ROM.

Power supply rails (noise an mains hum) first, followed by checking that the power-on-reset is present and correct on pin 40 of the CPU, followed by checking for pulses on CPU pin 7 (SYNC) to make sure that the CPU is executing instructions.

Report back.

Can you also remind me what the part number of the PCB is, so I can locate the correct schematic.

Dave

 

[DrAlis]

Great Dave! I will deliver the information requested in a couple of days. Alex

 

[DrAlis]

Alright. Pettester in place. Photos of version (img_6495), -5v (img_6500), 5v (img_6502), 12v (img_6501) [all 3 supply voltages measured at a dram], reset pin (img_6509) [5v after about one sec - before it is 0v], sync pin (img_6506) [it isnt regular when looked at wider timeframe]. Thank you.

 

[DrAlis]

Bonus photos in descending order from J7 (video output) pin 5 (img_6510) to pin 1 (img_6514)

thank you

 

[DrAlis]

Extra bonus to add some novelty, here is a thermal image of the PET Assy 8032080 16k that does not work:



And since I have another one (similar but different Assy 8032089 32k) that does work:


Both quickly show the CPU as the hottest spot. Machines have only been powered on for a short time. You can see the different RAM amounts clearly. I find warmth pattern fairly conistent, except for the rightmost column, but this is also where the Assy's differ. So unfortunately, no clues to malfunction found from this. And finally I have another image of a PET, my cat (Assy unknown):

 

[Gary C]

Ah, that looks like the M0gg-y32 assy.

 

[daver2]

OK,

So, I see from your oscilloscope traces that we have SYNC activity on the CPU (indicating that the CPU is actually executing something), that you have VDRIVE and HDRIVE (at the correct frequencies) on J7 - but no VIDEO signal. I think this was where we were at last time.

One thing I did notice is that you stated you have measured the +5V rail on the DRAM. However, this PET (803208 - Universal) has two (2) +5V rails. By measuring the +5V rail on the DRAM, you are reading one of the +5V rails - but missing the other. Since we have no idea which +5V regulator feeds which ICs, it is usually better to measure both +5V rails from each regulator. However, since we are seeing signs of life, I don't think the other regulator is causing us problems...

The next thing is to use your oscilloscope and look at all of the pins of IC UE12 (74154). This decodes the CPU address signals into blocks of 4K, with each block generally decoding for a specific sub-system within the PET.

Check each pin and post a list of whether the pin is HIGH, LOW or PULSING. I don't need a load of photographs - unless some of the displays do not look like nice TTL signals.

We know the CPU is executing instructions (from the activity on the SYNC pin). This test will give us a rough idea of what the CPU is accessing (e.g. which ROM(s), the video screen etc.).

Incidentally, the CPU SYNC signal will not be a regular 'clock-like' signal. This signal pulses HIGH each time the CPU fetches/executes an instruction. As each instruction takes a different amount of time to execute, the pulses will not be regular. If they were, I would 'smell a rat'! In this case, your PET Moggy would have already pointed it out !

Dave

 

[DrAlis]

UE12:

1-12: Puls High High High High High High High High High High Low

13-24: High High High Puls High Low Low Puls Puls Puls Puls High

I had a hard time measuring. Sometimes stuff changed suddenly. Some also showed superslow puls like activity one a second but maybe i just did bad contact. I did it a couple of times above is best consensus.

thanks

 

[daver2]

I am surprised that you do not have pulses on pin 9 (/SEL8). This pin is used by the CPU (under control of the PETTESTER ROM that is being correctly selected by pin 16 - /SELE) to select the screen.

Can you recheck pin 9 please.

Dave

 

[DrAlis]

Pin 9 has pulses (that did unfortunately look only high on high temp resolution to me). Image below

 

[daver2]

The key thing is there are low-going pulses there though.

Dave