• Please review our updated Terms and Rules here

PET 4032 vertical collapse?

daver2 said:
OK, so that fault is back again. Let's put that on a list of things to investigate later.

I would now suggest removing the ROMulator and putting all of the correct ROMs in place.

I would go with this as the EDIT ROM (to start with): http://www.zimmers.net/anonftp/pub/cbm/firmware/computers/pet/edit-4-80-b-50Hz.901474-04-0283.bin.

See what we get on the initial start-up screen.

Dave
Click to expand...
Nothing with this configuration. No chirp, nothing on screen, similar to other tests without pettester
 
Well, if you are having a problem with a video RAM this fault is also a possible problem with video RAM?

I would have said the error screen in post #159 was main RAM and not video RAM though.

I think you have IC socket problems to be honest.

Dave
 
The video ram issue was due to me still having one old chip that had been socketed and I hadn't cleaned up one of the legs good enough on the chip so it was sitting a bit high.
I've gone through the rom sockets and tested all of those and they all seem fine, is it the main chip sockets you suspect?
 
That error screen is main RAM (low bank of 16K).

I will have a look at which bit/bits are at fault tomorrow morning. But you should also be able to do the same if you read my PETTESTER manual.

Which pin of which IC was cleaned?

Dave
 
I think it was the chip in UC5, I'm not sure what leg. The problem is when it seemed to be an issue with the video memory I then swapped the chips around to see if I got a different result with the character screen and I could get some uppercase characters showing.

I've just read your pettester manual and I had assumed I had gone back a screen with the memory error page but the manual suggests that it's after the VDU test. At the moment I am back at the screen with all lowercase letters.

The screen was probably still warming up so I didn't see that first screen but is that suggesting that it was happy with what it saw and it moved on to the memory test?

If to I think I should spend some time double check all of the VDU memory sockets and chips to see if I can get the VDU test satisfied?
 
Just rebooted a few times and where it mostly seems to sit on the character screen I did get it to the screen of Gs and Bs but on another reboot to see if it was consistent it then passed that screen and started the ROM CRC, unfortunately it crashed before the memory check. I'm now back to the lower case characters test on boots and I've got an issue with the brightness POT on the CRT where it's going full brightness, fixes if I press it in then goes full brightness again so I need to sort that out first.
Could it be some of the small caps that are out of spec giving these inconsistent results?
1000027068.jpg
The CRCs are not happy either on that screen
 
My gut feeling is that some bus line is marginal. I.E. some chip has a weak driver and/or some chip loads a line a bit too hard, and/or still oxide in some chip socket and/or on a chip leg.
And/or a power rail with too much ripple, but that would probably give less consistent results.

I think it's time to check more or less every signal with an oscilloscope and make sure that there seems to not be any case where the levels are in between the max allowed voltage for a logic 0 and the min allowed voltage for a logic 1 for TTL logic, except for the short transitions.

I don't know what modern oscilloscopes can do, but it would be nice if you can tell the oscilloscope to warn if you measure on a signal that wouldn't be correct TTL with a specified max rise/fall time, and either way tells if the problem is slow rise/fall time or an incorrect flat level.
 
Have I messed up here and been caught out with the romulator again?
I've swapped the ROMs out and gone from bad CRCs to good CRCs so I have presumed everything is fine and we've made progress. Now we are taking the romulator out and looking at the other faults. Pushing the pettester with the multiple reboots when it's in the editor slot and it managed to get to the CRC screen it doesn't seem happy with the roms.
I'm fairly certain that the rom burning is ok as I can read the rom back and do a verify. The other elements is the adapter.
The adapters that I have are 2532 to 2732 adapters.
The PET manual lists 2332 ROMs and I can see there are 2332 to 2732 adapters.
Looking online the pinouts do seem to be different between 2332 and 2532 so is it that I dont acutally have valid roms due to me using the wrong adapters?
 
Maybe use a burned ROM with the PETTESTER at the Kernal Fxxx location, and no other ROM's at all, and make sure the display and keyboard seems to work fine?
And then add your old ROM chips, one at a time, and see what happens?
 
MiaM said:
Maybe use a burned ROM with the PETTESTER at the Kernal Fxxx location, and no other ROM's at all, and make sure the display and keyboard seems to work fine?
And then add your old ROM chips, one at a time, and see what happens?
Click to expand...
The problem with the original roms is they dont seem to be standard roms. They are eproms again and dont seem to be standard images. I posted a copy of them earlier in this mammoth topic. The romulator seems that it might be a help and a hinderance especially without the option on it to disable RAM / ROMS. Just so weird how its gone from bad to good CRCs.

I dont have faith that the adapters that I have are right and I dont have faith that the romulator is reporting correctly. I think my plan is to order known good adapters for a PET.
https://www.pcbway.com/project/shar..._2332_ROM_to_2732_EPROM_adapter_50aae5c8.html
Whilst I wait for them, pay attention to the sockets testing for low ohms between the socket connector and the next point in the circuit.
Look at some of the smaller caps that I haven't looked at yet. I did notice a couple with high ESR and I have replacements for those now
 
Ignore the small capacitors. This is just 'noise' and irrelevant.

The only issue would be if a lot of the decoupling capacitors have failed. Highly unlikely.

Dave
 
A 2532 EPROM can be used as a direct replacement for a 2332 ROM (with one exception that I will explain shortly).

If you have a 2732 EPROM, then a 2732 (EPROM) to 2532 (socket) should be fine - with the above caveat.

In both cases, pin 20 is an active low chip select.

Pin 21 is the exception. On a 2332 ROM this is an active high chip select (i.e. a logic signal). On the 2532 EPROM this pin is a power pin.

The problem is with how this pin is driven and what current the EPROM(s) consume. If pin 21 is directly connected to the +5V rail, things will be happy. If pin 21 is connected via a pull-up resistor to +5V, you may have a problem (e.g. the Univ board). This creates a /NOROM signal (to disable the ROMs if the board is used with a second processor (e.g. the SuperPET add-on board)). However, if the EPROM consumes too much current, the voltage on this pin of the EPROMs will decrease (due to the presence of the 100R pull-up resistor) and the reliability of the EPROM will suffer - or they will not work.

Try measuring the voltage on pin 21 of the (EP)ROMs.

Dave
 
The original firmware installed looks like a 'hacked' version of BASIC designed for an embedded application.

There is text such as: "*** PPM 8032 CONTROLLER ***" and "DLOAD AUTOLOAD". This implies that there was probably a disk unit attached that booted an application called "AUTOBOOT" when the machine was turned on.

I suspect the above text "*** PPM 8032 CONTROLLER ***" should/may have been displayed when the machine was turned on?

PPM = Parts Per Million (unlikely), Pulse Position Modulation? Other...

Dave
 
Looks very similar to this thread:

C

Thread 'Unknown PPM PET roms'

All,

I acquired an 8096 SK a while back with some unknown ROMs - at least to me. When it boots it tries to load a file called "autoload" from device 8. The only clue I found to them was on a humanist website, which suggested that a company called PPM in Brookwood, Surrey, UK used to convert PET's and Apple II's for data acquisition and logging.

In the absence of any software or the hardware they interface to, I will probably just revert it back to a stock 8096 but the ROM's are posted here of they're of interest to anyone.

http://inchocks.co.uk/commodore/PET/ppm_roms.zip

Here's...

Colin.
 
daver2 said:
Pin 21 is the exception. On a 2332 ROM this is an active high chip select (i.e. a logic signal). On the 2532 EPROM this pin is a power pin.

The problem is with how this pin is driven and what current the EPROM(s) consume. If pin 21 is directly connected to the +5V rail, things will be happy. If pin 21 is connected via a pull-up resistor to +5V, you may have a problem (e.g. the Univ board). This creates a /NOROM signal (to disable the ROMs if the board is used with a second processor (e.g. the SuperPET add-on board)). However, if the EPROM consumes too much current, the voltage on this pin of the EPROMs will decrease (due to the presence of the 100R pull-up resistor) and the reliability of the EPROM will suffer - or they will not work.

Try measuring the voltage on pin 21 of the (EP)ROMs.
Click to expand...
daver2,
While I have never had a problem using a 2532 EPROM in a 8032 PET, you are technically correct as the Motorola Specification for the MCM2532 states that the Ipp max is 5mA, so the 100 Ohm pull-up will worst case drop Vcc by 0.5V at pin 21 to 4.5V. And Vpp minimum of the 2532 is specified at 4.75V.

I Suppose that I should add another 100 Ohm 1/4 W resistor in parallel to the R40 No_ROM pull-up? I will definitely call Commodore about this!

Also, as you mentioned the use of grounding the /NO_ROM signal in the SuperPet mode, the use of 2532 EPROMs are verboten in the 8032 mainboards of SuperPets.
-dave_m
 
I remember some problems with using a 25xx EPROM as a character generator and the dreaded /INIT signal that really messes things up, and extrapolated this to the /NOROM signal.

I found a fault in a DEC PDP-8 device driver once. I tried to report it to DEC, but it appears the warranty had expired!

Dave
 

daver2 said:
I remember some problems with using a 25xx EPROM as a character generator and the dreaded /INIT signal that really messes things up, and extrapolated this to the /NOROM signal.
Click to expand...
Good catch with the /INIT pull-up. It is a 1 K Ohm resistor which is fine for gate inputs, but would be definitely a problem for the Vpp input of a 2516 or 2716 EPROM. A possible solution there would be to use a 2816 EEPROM. Pin 21 on the 2816 is /Write Enable and /INIT will work to keep it at a logic high.
 
Or just stack sockets to create a simple adapter that routes VPP to VCC and leaves that socket pin not connected.
 
Quick update as to where I am so far.
I have been waiting on the new EPROM adapters. Whilst waiting I checked the sockets on the board for continuity to something down / up stream based on the circuit diagrams but didnt find any issue. I also fixed the brightness issue with the monitor.
The EPROM adapters I had previously gone for were these. I had assumed they were ok due to getting good CRCs with the pettester on the romulator when they were in and different results when they were out.
Since the previous post where I thought the romulator was actually doing me more harm than good. I ordered these as they were known to work with the PET.
They arrived yesterday and having used the new sockets and with a pettester in the editor slot for the first time it consistently would run the tests and report correct CRCs without the romulator.
After that it was a hunt for an editor rom that would work with the screen, 80 columns and the keyboard. After quite a few attempts I've landed with "sjg-edit-80-n-50 (2018-11-26).bin" which seems to work.
All I have done so far is type the program in below to prove the keyboard worked and it would run something. The i key was a bit iffy so there is definitely some keyboard cleaning to do and it doesn't boot every time so more troubleshooting to do but it's a great feeling to have something that is starting to be useable.

Thank you for all the help so far. I'll do some more testing over the next few days as I can get time. I have the other board to look at as well now I have know good roms and main chips and at some point I want to try the disk drive as well.

View attachment pet.mp4
 
Well done.

There is a little test program you can type in from a book to test the IEEE488 buffers without having anything attached. This is a good test before you start playing with the disk drives.

Dave
 
You must log in or register to reply here.
Back
Top