• Please review our updated Terms and Rules here

Superpet Combo board not working

P

pttak

Member
Joined
Feb 21, 2024
Messages
18
The 8032 boots and works fine. When I plug in the combo board it will not boot in 6502 or 6809. This board used to work. The board is very clean like new. I have visually inspected very thoroughly and see nothing wrong.

I checked the ribbon cable thoroughly and 99.9% sure it is ok. Both 7805 regulators on the combo are at 5 volts. The 8032 has switches for RES, NMI and DIAG. I see activity on address and data buss when I RES, but it goes away in a few seconds.

I tried plugging a bitfixer romulator into the 6502 socket on the combo and it boots (green light) but no jingle. I tried NO-OP and looked at address and data lines on 8032 ROMs and do not get normal results.

Can anyone help with more things I should look for?

Thanks, Mike in Alaska
pttak
 
There's a lot going on with the Super PET. I'm assuming you have the two board system, just the 8032 main board and one SuperPet board?
Is the 8032 originally from a SuperPET? There's a mod that's required to send the system clock up the ribbon cable to the 6809.

Sounds like it's crashing shortly after startup.
Do you have any option ROMs installed in sockets UD11 or UD12?
Check the /NOROM signal and make sure that the 6502 and 6809 ROMs aren't interfering with each other. If both ROM sets are active at the same time, neither CPU will be able to read from their startup ROMs.

If you've validated all the system RAM on the PET board, that makes RAM less likely to be the problem.
There has to be something on the address or data bus causing a problem only when the SP board is connected.
The 6809 ROMs should be disconnected when in 6502 mode but the expansion RAM and the Serial UART are both still connected in both modes.
So if they are socketed, you might want to remove all the SP RAM and the UART to see if the PET starts without those plugged in.
 
Last edited:
Hello Hutch, Thank you very much for the quick response.

Yes, two boards. Not an original Superpet. The 8032 is version 8032080.

What is the mod you mention? That must be to get 16MHz on pin 6 of the ribbon via pin 35 of the 6502 socket? There is currently no jumper installed on the 8032 board.

I have no option ROMs installed.

I checked out the NOROM signal and I believe this is as it should be, but I need to confirm this. I'm learning a lot. The key statement you made is that the 16MHz clock needs to make its way up the ribbon.

There may still be a problem on the 6502 side. To be continued....

Thanks again.

Mike
pttak
 
Welcome to VCFED.

Whilst talking of the NO ROM signal - do you have the 'real' ROMs on the 6502 PET board or have some if them been replaced with EPROMs?

If you have EPROMs instead of ROMs they do not respond to the NO ROM signal and that will mess up your SuperPET...

A Universal board has the 16 MHz clock modification (I think) but an 8032 requires the modification.

>>> I tried NO-OP and looked at address and data lines on 8032 ROMs and do not get normal results.

It would help to share a video of this. This may prove crucial to identifying what is wrong.

Also, do you have the correct character generator insralled in the 8032 for the SuperPET? It won't stop it booting of course, but it will prevent you from running some of the languages (e.g. APL) if you don't.

Dave
 
Hi Dave,

Fortunately I have the original ROMs in this unit. I previously burned an EPROM to fix the edit ROM but that was for my other 8032 which is now working fine and set aside.

I need to recheck the NoROM status tomorrow, there could be a problem there. (it's going on midnight here so to be continued tomorrow).

I believe my character ROM is 8032. I presume the Superpet character ROM (characters.901640-01.bin) can be burned into a 2532 EPROM for plug and play?

Thanks very much for all the help.

Mike
pttak
 
If the /NOROM signal is LOW then the 8032 ROMs should be disabled.

If you check the schematics for the univ2 board you will see the 6502 CPU pin 35 link to 16 MHz for the SuperPET. You will need to do the same modification to the 8032 (as detailed above).

You can fit a standard 4K EPROM into the character generator socket. However, if you do, bend pin 21 of the EPROM out and directly wire it to pin 24 (+5V) of the EPROM after you have burnt it. The reason for this is that pin 21 on a ROM is a chip select logic input that is pulled up via a resistor to +5V. On an EPROM, pin 21 is specified as a power pin and can (in exceptional cases) draw too much current, reducing the voltage on the pin, and causing the PET video sub-system to malfunction. You can try it first without the modification and see if it works OK. It probably will. If the PET goes screwey, this is probably the reason why...

Dave
 
The Character ROM is not connected to the CPU Bus so no need to worry about the NOROM signal for the Character ROM. It remains in use no matter which mode you're in.
There's no way it could cause either mode to stop working.

My earlier mention of the SuperPET RAM and 6551 ACIA is because some failure modes can cause a chip to remain on the bus when it is not selected which will cause interference.
If the 6502 mode works fine with the SP board removed, but doesn't work when it's connected, this seems to be a likely cause since, in theory, those are the only parts that should remain available on the bus in 6502 mode.
 
Last edited:
Hutch said:
The Character ROM is not connected to the CPU Bus so no need to worry about the NOROM signal for the Character ROM. It remains in use no matter which mode you're in.
There's no way it could cause either mode to stop working.

My earlier mention of the SuperPET RAM and 6551 ACIA is because some failure modes can cause a chip to remain on the bus when it is not selected which will cause interference.
If the 6502 mode works fine with the SP board removed, but doesn't work when it's connected, this seems to be a likely cause since, in theory, those are the only parts that should remain available on the bus in 6502 mode.
Click to expand...

I modified the 8032 board for the 16MHz clock. I put everything back together and checked NoROM and it is high, enabling the 8032 ROMs. But still no boot.

I installed the bitfixer romulator (dip position #2) on the combo board and it boots into the mlm and is stuck. https://kl6m.com/c/MLM.JPG

I changed the dip to #13 and it fills the screen with constantly changing gibberish. I have yet to try #12 (noOp), which is my next attempt.

Thanks...Mike
 
Booting to the MLM means the CPU is running into a BRK instruction (00) when it shouldn't.
This is normally cause by a bad ROM.

NOROM might be OK, and the onboard ROMs fine, but IF the SuperPET ROMs are also on the bus and being selected, that will interfere with the PET ROMs in the same address space.
So it sounds like the ROMs on the SP board are not being disabled.
 
The NoROM line is high with 6502 switch selected. All three 6809 ROM select lines (pin20) U49, U48, U47 are all logic high. When I switch to 6809 only U49 select goes to logic low. Nothing on the display in 6809 position.

With romulator installed on the combo board and #13 selected, the gibberish screen I mentioned in the last message suddenly assumed a format that approached the proper screen for #13, but the ROM check sums were unstable and cycling. It actually attemped the RAM check. This all suddenly quit and I could not get it to repeat. This indicates to me, possible a bad connection someplace. I have rung out the ribbon cable twice, and it seems solid.

I removed the combo board and plugged the 8032 6502 back in to make sure something did not fail on the main board, and all was normal.

I guess next I will try romulator no-op #12 and o-scope it.
 
So, we are definitely sure that the problem is related to the SuperPET Combo board then...

I would start to check the SuperPET Combo board in a slow and methodical manner.

With the switch set to 6502 mode, check the following pins:

U17 pin 8 (VCC of the 6502 CPU).

U13 pins 39 and 40 (TSC and /HALT of the 6809 CPU).

If you changeover the switch to the 6809 CPU - then check the readings of the above pins again - they should all differ.

Dave
 
6502 mode:

U17 pin8 = +5v
U13 pin39 = high
U13 pin40 = low

6809 mode

U17 pin8 = low
U13 pin39 = low
U13 pin40 = high
 
That all looks normal.

I would next check all of the clock signals. We have the CPU pin clocks and the clocks divided down from the supplied 16 MHz clock.

I would also check both CPU /RESET pins. The /RESET should be activated when the CPU switch is changed over.

Dave
 
I removed the 6809 and tried to boot up in 6502 mode. Then I scoped all 40 pins. I have results here: https://kl6m.com/c/test1.gif

I have a reset button installed so I don't have to cycle the power switch. When I press it, I get action on data buss momentarily but quits after a few seconds.
 
With the 6809 removed I was able to boot in 6502 mode intermittently. Works sometimes and not others. I have yet to figure out the cause.

Unfortunately I believe there are multiple issues on the combo board. On the o-scope I see a lot of + and - transients on both edges of many of the pulses. I have carefully studied the schematic. I noticed that the design seems to be lacking in the recommended
decoupling capacitors for typical TTL circuitry. In many cases there is only a single 0.01uF glass decoupling capacitor shared between two devices.

I replaced C41 (the decoupling cap for VR1 and the 6502) with a 0.01uF and 0.1uF in parallel and obtain a lot more consistency of booting, but still sometimes intermittent. I think I will add a lot of 0.1uF around the board. Unusual since I had another identical combo board that worked perfectly as-is.

Also I believe there may be a 'mechanical' issue of some kind, like a bad connection, maybe even a faulty trace on the board. I tried to eliminate bad solder joints by resoldering every pin of both processors and also J1. I may have a bad 6809.

I have also rung out the ribbon twice now, and found nothing wrong with it. But I feel like I can't eliminate it quite yet. I also have carefully studied the the board itself under a microscope, looking for any visible problem but find nothing so far.

This is kind of a nightmare but I plan to stick with it until resolved.
 
Still no luck getting combo board to work. 8032 without combo works fine.

My 8032 is a version 8032080. Looking at 8032080 sheet 1 of 7, are any of these mods needed for combo board add-on? I did perform the 16 MHz mod (#12)
 
daver2 said:
So, we are definitely sure that the problem is related to the SuperPET Combo board then...

I would start to check the SuperPET Combo board in a slow and methodical manner.

With the switch set to 6502 mode, check the following pins:

U17 pin 8 (VCC of the 6502 CPU).

U13 pins 39 and 40 (TSC and /HALT of the 6809 CPU).

If you changeover the switch to the 6809 CPU - then check the readings of the above pins again - they should all differ.

Dave
Click to expand...
Hi Dave,

I've been taking another crack at fixing the combo board.

I had previously removed the 6809 but had never tried removing ROM U49. It happens to be socketed so I removed it and suddenly now the PET mode finally boots with the combo board installed.

I downloaded 970020-12 from Zimmer and burned a 68766 EPROM and plugged it in. And the PET mode still boots. I plugged the 6809 back in and now again it won't boot in PET mode.

I have another 6809 but it is a 6809E. With this installed the PET mode boots normally. So I think my 6809 might be bad. But with the 6809E installed when I flip the switch from 6502 to 6809 I get a screen full of bars. My question: is the 6809E usable in the Superpet?

Thanks, Mike
 
The 6809 and 6809E are not compatible with each other!

Please, please, check the data sheets and data books BEFORE exchanging a part with something else. In particular, ask us on VCFED first BEFORE doing the exchange.

For most parts, a suffix letter may be inconsequential. However it may also mean that the part looks physically different and will not fit. It can also mean that the component is in the same physical package (so it looks the same and will fit) but (in the case of the 6809 and 6809E) one cannot be used to replace the other.

To my knowledge only the 6809E is usable in the SuperPET (both the original and the combo). However, the schematic diagram for the combo indicates a 6809 whereas the parts list indicates the correct 6809E. The 'clincher' is the use of pin 39 as TSC. The TSC pin is ONLY available on the 6809E. On the 6809, this pin is XTAL.

So, was the original fitted a 6809 or a 6809E? If a 6809 was originally fitted when you got the machine - this was wrong and someone has attempted an abortive repair...

I don't think any damage has occurred as a result though.

Incidentally, some 6809 chips have been remarked as 6809E (and vice versa) by unscrupulous sellers...

Dave
 
Thanks very much. Good advice. I ordered a couple "E"

I have two 6809s here, plain and E. Unfortunately I don't recall which one came out of the combo board, but I would guess it was the E. I'm pretty sure this board worked when I first got it decades ago.

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