• Please review our updated Terms and Rules here

SuperPET no video issue

snuci

Veteran Member
Joined
Nov 22, 2012
Messages
1,545
Location
Richmond Hill, Ontario, Canada
I acquired a SuperPET sometime ago and I've been trying to fix it but this one is a little weird. I've removed the BMB board and the Computhink boards and am focused on the 8032 board at the moment.

I know the monitor works fine as well as the power supply because if I put in another 8032 motherboard, it works fine.
I have a ROMulator to help me troubleshoot (it's really awesome)
- If I set up the ROMulator to emulate the ROMs and memory for an 8032, I get a start up beep but no video.
- If I set up the ROMulator to emulate the ROMs and memory but with the PETTESTER ROM, I get nothing on screen
- I've checked to make sure that the CRTC chip works in another machine and it's fine.
- I've replaced some of the memory that was bad in that I checked the first 8 chips and 6 were bad but I suppose I should check all of it. I figured the PETTESTER ROM would help me identify the other bad chips so I wouldn't have to desolder all of it to check it.
- I've checked J7 and have 20khz on pin 5 and 60Hz on pin 3. I get nothing on pin 1 but it's not in the schematics that I can see easily.

I guess my first question is, does the start up beep indicate that the PET is running properly? Without any video, I cannot tell.

Any help is much appreciated.
Santo
 
The beep indicates that the CPU is running and has correctly made it into the EDIT ROM where it is initialising the hardware. That’s it. The CPU may go into the weeds after that. Faulty memory could be the cause of that. The Kernal ROM calls the EDIT ROM and the subroutine return from the EDIT ROM could go into the weeds.

My PETTESTER doesn’t use the stack (initially) for this very reason.

Because you can see the correct H and V frequencies, the CRTC must have been initialised correctly.

I would work backwards from the missing video signal on J7/1 by checking the following:

UA2 (74166) pin 13. This is the ‘raw’ video signal from the parallel to serial shift register. If you get nothing here we need to work further backwards.

UC2 (7486) pins 8, 9 and 10.

UD4 (74LS10) pins 12, 1, 2 and 13.

Dave
 
Thank you Daver2. Once I knew where J7/pin 1 was in the schematics, I could work backwards. The culprit was UC1.in the end.

Going through PETTESTER now and moving ahead.

Thanks again,
Santo
 
Yes, I can see something finally :) Works great with the ROMulator but now I peel back the emulation and have video RAM issues. I alternate between these two screen with PETTESTER.

IMG_2404.jpg IMG_2403.jpg

Time to do some digging. i thought i had already checked the video RAM.
 
When you say it works with the ROMulator - what do you mean?

My guess would be something wrong associated with SA9 or TA9 (in the video circuitry) or BA10 in the CPU. Check around UC10 on pin 13 (BA10), pin 14 (TA9) and pin 12 (SA9).

It looks like half of the video memory is not being accessed.

Depending which ROM emulator you have - you may have the NOP generator in-built so you should see BA10 ticking away at half the frequency of BA9.

Dave
 
With Bitfixer's ROMulator, I can emulate onboard RAM memory and ROMs so when I run it fully emulated like this, the PET works fine. When I remove the memory and ROM emulation and just add PETTESTER via the ROMulator, it didn't work and you see what I see above. What I did learn was something interesting though.

I've since removed the 16 RAM memory chips to test it and of the 16 RAM chips, 14 chips were bad. I also had a physical PETTESTER EPROM I had made and put it in the kernal socket (UD6). Little did I know, the PETTESTER ROM on the ROMulator replaced the Edit ROM ($E000) as that was how it was configured out of the box unless I did something wrong by mistake. So I technically had two PETTESTER ROMs including an emulated one at $E000 (via ROMulator) and a physical EPROM I had created in $F000 at UD6. Once I sorted this out by putting the kernal ROM back in UD6, I was able to test properly and along with the bad memory that was fixed, all is now 100% working without any ROMulator at all.

Now I have to move to the two SuperPET boards.

Just to confirm, PETTESTER should replace the kernal ROM (901465-22) in UD6 at $F000, right?

Thanks again for your help and your wonderful PETTESTER ROM. I've used it many times in the past and will continue to do so.

Santo
 
>>> Thanks again for your help and your wonderful PETTESTER ROM.

Cheers.

>>> I've used it many times in the past and will continue to do so.

That's what it was designed for. Please, carry on!

My original development was designed to be used in place of the EDIT ROM socket at $E000. This (EP)ROM should be in a socket and should, therefore, be easy to swap out. This configuration requires the reset vector and half a dozen instructions to 'work' from the KERNAL ROM in order to enter my PETTESTER code.

However, others have taken my code and reassembled it to run in place of the KERNAL ROM socket at $F000. The KERNAL ROM is not always in a socket - but soldered into the board. This (obviously) makes it more difficult to swap out.

If you downloaded the code from my Google drive, it is assembled to run in the EDIT ROM socket at $E000.

I can't vouch for any other location you obtained it from.

My naming convention for the PETTESTER file itself indicates the ROM size required (2K) and the base address (E) along with the version number (latest V4).

Dave
 
Thanks for the explanation Dave. I do have the Edit ROM version.

This SuperPET is driving me a little crazy. I'm not sure if you or anyone else can help because I know it's difficult to troubleshoot remotely but maybe someone has an idea.

As noted, I got the 8032 board working now and passes all tests on the PETTESTER ROM.. For the SuperPET boards, this is a three board version where the BMB board (6809 + 6502) and the CompuThink board (the memory board) are known to work. I've swapped them in another SuperPET and they are fine. However, they don't fully work with the original 8032 board. I can run 6502 mode just fine but when I switch to 6509 mode, this shows a blank screen in most cases. Occasionally a screen full of garbage but also a single white dot other times.

For the working SuperPET, I swapped the 8032 boards only (keeping the cables and switches original for each machine) and the fixed 8032 board does the same thing in the working SuperPET. Putting the working SuperPET 8032 board in the non-working SuperPET makes it work. It is definitely the 8032 motherboard.

I am assuming that since 6502 mode works, there may be issues related to the memory expansion port that the 6509 board attaches to. I've checked all of the ICs between the 6502 and the expansion port and everything looks fine. If you have any ideas, I'd love to hear them. 6502 mode works fine but not 6509 mode. It is a pretty unique issue so I would completely understand if you aren't sure. I have a chip tester and have removed some of the TTL logic to test it and it all seems fine.

Thanks again.
Santo
 
Before we go off at a tangent can we just check one stupidity first please.

Does the 8032 board have 'original' ROMs or have some of them been replaced with EPROMs?

If they are all original you shouldn't have a problem. If some have been replaced by EPROMs, it is possible that the replacements will not respond correctly to the /NOROM signal that is asserted by the 6809 second processor to disable all of the 6502 ROMs. If they do not disable properly, we will potentially have bus contention when the 6809 CPU is selected.

I would just like to rule this scenario out from the outset. This would fit your scenario that it is something on the 8032 board itself.

Previous post: 6509 == 6809...

Dave
 
Does the 8032 board have 'original' ROMs or have some of them been replaced with EPROMs?

Interesting. 901465-22 @ UD6 and 901447-10 @ UB3 are not original on the bad board but they are all original on the working board. Sadly, I don't have originals for either. I swapped the character ROMs at UB3 and the programmed EPROM works on the good SuperPET. Putting the real one in the bad SuperPET still has the same bad results.

UD6, on the other hand, is soldered in and I would prefer to keep it there but if I have to I can desolder it and try it.. For reference, this is what I get infrequently when I start it up after not starting it for a bit. Any subsequent off/on switches are a blank screen. If I switch from 6502 to 6509 while powered on, the screen slightly corrupts and then hangs.


IMG_2426.jpg
 
The character generator isn’t the problem - it is not connected to the CPU data bus.

The question is (with regard to UD6 and UB3) what devices have they been replaced by? Are they EPROMs? And, if so, what type?

Dave
 
I desoldered an original 901465-22 @ UD6 from a 4032 known working. I put in this chip and the original UB3 and we still have the same issue. I think we can eliminate that as a possible issue. Thanks for noting the possibility though. I didn't know there might be an issue. I use TMS2532A-45JL EPROMs normally and that was what was in there originally.

Santo
 
For a TMS2532 EPROM having a VPP (pin 21) at any voltage other than +5V and the programming voltage is not specified by the manufacturer.

Are there any other EPROMs in the language ROM sockets or are they now all original ROMs?

It might be worth checking the voltage on pin 21 of the language ROMs. This pin should be HIGH for 6502 mode and LOW for 6809 mode.

The /NOROM signal is derived on the SuperPET board, but is routed via an ‘unused’ pin on the 6502 CPU (and hence the IC socket) onto the logic board.

Dave
 
Hi Dave,

The option ROMs are original MOS BASIC ROMs. I'm not sure why but these worked fine with the working 8032 board so those aren't an issue.

Pin 21 on the language ROMs are high in 6502 and low in 6509 as shown here.

6509:

Pin21-6509.jpg

6502:
Pin21-6502.jpg

At least that seems to be working okay.
 
Yes, this signal looks OK. I forgot it should oscillate in 6502 mode - dependent upon the state of the read/write signal but (in 6809 mode) it should be permanently LOW disabling the onboard ROMs.

So, to recap, another 8032 board works with the ‘duff machine’ SuperPET boards OK and the ‘duff machine’ 8032 doesn’t work in another (working) SuperPET. But, the 8032 works OK as a bog standard 6502 PET. Is this statement correct?

If it is, then we are looking for the differences between the 8032 working as a bog standard machine and a SuperPET.

Out of interest, which cables did you use when you moved the 8032 into the other (working) SuperPET?

Dave
 
Another thought. Can you post the link settings for your 8032?

Not expecting an issue, but just want to rule out another variation.

Is it also a standard 8032, a universal or a universal 2? Just want to get the correct schematic...

Dave
 
So, to recap, another 8032 board works with the ‘duff machine’ SuperPET boards OK and the ‘duff machine’ 8032 doesn’t work in another (working) SuperPET. But, the 8032 works OK as a bog standard 6502 PET. Is this statement correct?

This statement is correct.

If it is, then we are looking for the differences between the 8032 working as a bog standard machine and a SuperPET.

Exactly. This is a three board SuperPET so since two cables are used, one being one connected to the 8032 6502 socket and the other being the memory expansion connector, I figured the issue was around the expansion connector since it's not really used in 8032 mode.

Out of interest, which cables did you use when you moved the 8032 into the other (working) SuperPET?/QUOTE]

I made sure I used everything from the working or non-working machine in their respective machines including the cables, switches and option ROM so the only thing different is the 8032 board. At one point I suspected the cables and then the switches. I had also done swaps with these with no change in working ability. I believe I've been fairly thorough. I have been at this for days so I've tried every combination. In fact, I have a spare 8032 board and this also does NOT work. It, however, has all ROM chips replaced but has the same issue (6502 working, 6509 not working).
 
Another thought. Can you post the link settings for your 8032?

Not expecting an issue, but just want to rule out another variation.

Is it also a standard 8032, a universal or a universal 2? Just want to get the correct schematic...

Picture of the board here. All jumpers that i could see are the same.


Click image for larger version  Name:	IMG_2427.jpg Views:	0 Size:	149.0 KB ID:	1223785

The working SuperPET is 8032089. The non-working one is 8032090.

EDIT: Posted a better res temporary pic here: https://vintagecomputer.ca/files/temp/SuperPET-MB.jpg
 
Last edited:
This is fairly weird...

Question - Are your ROMS on the SuperPET board in IC sockets or not?

The one of interest (on the schematic I have seen - but this may not necessarily be your particular SuperPET board of course) is U22 (Address range $Fxxx).

The reason I am asking is that I was just wondering whether I could convert my PETTESTER ROM from 6502 to 6809 so that it could be run on the SuperPET. This may take a while to do fully, but we could do it in stages. Get the CPU initialisation, CRTC initialisation and video memory test done first and then work on other aspects.

Just a thought.

I think the memory expansion port of the 8032 may be 'clear' as the data bus buffers (UB9/10) would be checked by the video circuitry (this is on the far side of the buffers). Similarly with the address buffers, there is a lot of standard 8032 PET logic on the far side of these buffers also.

Dave
 
Back
Top