Cbm pet motherboard. 8096 80 coloumn.

[mjnurney]

Hi all,

another doomed pet motherboard needs help.

This board is out of an 8096 pet which used to run a farm in the 80s, anyway the former owner powered it on after 20 years and the power filter exploded and killed the machine..

it did nothing when i bought it,


so faults.

It used to display nothing but a dot on the screen which faded. but with a rom ram board fitted it displays have a screen of garbage and nothing else. i have changed UD13 and now it displays the TIM and half a screen of garbage.

if i substitute all the ram and rom for theonboard romboard it will display commodore basic 4.0 in half a screen if UB12 is removed.

i have run Daves pettest v2 in the kernal socket and it will boot without the romboard and displays half a screen of @abcdefghjklm etc and the flicks over to a half screen of bbbbbb.

sadly i cannot seem to get petteste4k running at all. - nothing happens in kernal or edit rom sockets.

i have socketed the roms / half the ram (upper) and the large ICs - 6502,6522 etc.


motherboard details: universal dynamic pet 80 column.

its currently connected to my 4032 as the 80 column pet is in the loft.

 

[mjnurney]

Pet test v2 results

 

[dave_m]

sadly i cannot seem to get petteste4k running at all. - nothing happens in kernal or edit rom sockets.

To get Daver2's pettestV4 to run, it must be in the EDIT socket (UD7) and a good kernal ROM must be in UD6.

When using the RAM/ROM board, did you remember to chose the option set for 80 columns even in the 4032?

 

[mjnurney]

Hi Dave, no option to select 40 or 80 column on the rom board as far as I know. Yep I’ve tried it in the edit and kernel socket without joy sadly

 

[daver2]

Isn't the 40/80 column selection mechanism the links just above the ROMs in your post #1?

Dave

 

[MikeS]

Isn't the 40/80 column selection mechanism the links just above the ROMs in your post #1?

Dave

Those and two or three other jumpers elsewhere; also an undocumented mod when going 80 to 40 to prevent the extra memory (normally not installed in a 40) from interfering IIRC.

 

[dave_m]

Hi Dave, no option to select 40 or 80 column on the rom board as far as I know. Yep I've tried it in the edit and kernel socket without joy sadly

I wasn't clear, I meant, on your RAM/ROM board did you make sure you selected the ROM set with the EDIT-4-80-b Edit ROM rather than the EDIT-4-40-b file? As your board is jumpered for 80 columns, you need that EDIT ROM even if you put the board into a 4032 PET chassis.

Read my message #3 again. You must put the PETTESTV4 chip in the EDIT socket (UD7), and you need a good kernal chip (901465-22) in UD6 for PETTESTV4 to work.

 

[Dwight Elvey]

It does look like you have working VRAM in the second half. It is just not being initializes. Maybe the firmware doesn't think it is a 80 column. That is what it looks like. I think Dave_m has the right idea. Some jumper is not right or your running the wrong firmware. The monitor display clearly shows that it thinks it is a 40 column.
Dwight

 

[mjnurney]

Hi all,

ive fitted the pet test in both kernel and edit sockets and nothing happens - maybe the kernel is bad? Also my rom/ram board is a very early version that I think is hard coded to 40 column - hence the split screen. I’ve spoken to Dave Curran about updating it if possible…

I’ll burn a new kernel rom on an eprom and try again. Something is wrong somewhere.

 

[mjnurney]

Some success, I’ve burnt a 4032 kernel to a 2542 and now it boots the pet test v04 and that’s a big result! I think I have 2x dead kernel roms.

anyway I now a quick screen of abcdefghi etc and then a lot of bbbbbbb some @@@@@ and then bbbb and aaaaaaa

bad main ram ?

 

[daver2]

I can't really see the image. Any chance of getting a closer / better resolution image at all?

It will be either hanging on the VIDEO RAM test or the Page 0/1 RAM test. We need to ascertain which. You still appear to have the 40/80 column weirdo - or a faulty VDU logic circuit...

EDIT: I read your post again - so a screen of @abcdef... is the VDU test. If that clears and PETTESTER moves on - then you must have a MAIN RAM (page 0/1) fault. You should (however) see 'g' and 'b' for the first 256 bytes (page 0) followed by 256 bytes of either '.' or the character in error. This should be repeated for the next 256 bytes (page 1).

If you video the screen - and slowly step the video - you can see what is going on. I wonder what you see when you look at the VDU test screen?

Cheers,

Dave

 

[mjnurney]

Blindly that is a rubbish photo, hang on

thats a bit better, I’m having to resize the photos to upload them.

So I suspect I have bad ram, as the VDU test clears. I can substitute the ram for the rom/ram board and that may hide that fault to allow rom testing ? Tomorrow’s task … although football is on.

 

[daver2]

Excellent - thank you...

Your RAM is always reading as $00 - irrespective of what is being written into it.

$00 is written into address $0000 - and that is 'g' good.

$01 through $FF is written into addresses $0001 through $00FF - and this is read as 'b' bad. The character that is returned is a '@' on the read - which is PETSCII for $00.

That is the assessment - now why is it happening?

Dave

 

[mjnurney]

They can’t all be bad can they? I can socket the ram to allow testing. I have ordered a ram tester for a few pounds rather than pulling apart zx spectrums…

 

[daver2]

I didn’t actually say that the RAMs were faulty. Only that they are reading as $00.

Now I have looked at your photograph more closely, I see that page 0 (addresses $0000 to $00FF) reads as $00, whereas page 1 (addresses $0100 to $01FF) reads as $01.

‘@‘ is PETSCII for $00 whilst ‘a’ is PETSCII for $01.

Interesting...

I will need to check my own PETTESTER documentation though to make sure I remember the test correctly. It is getting a little late in the UK!

EDIT: Out of curiosity (and for a good reason) can you identify exactly the part number of your PCB and post details of all of the link configuration on your board please. I have a hunch...

Dave

 

[Dwight Elvey]

I would say one of the first things to look at are the selects. You need to use a nop generator to do that.
Also, make sure the DRAM are getting the right rail voltages
Dwight

 

[mjnurney]

Some close ups if that helps. This board was working until the filter cap exploded on the previous owner…

 

[daver2]

Thanks,

So, from the PCB ID I see it is a Universal 2 main board. You will find all of the documentation here for it http://www.zimmers.net/anonftp/pub/c...iv2/index.html.

The images you have posted indicate it appears to be correctly configured for 80 columns - but I can't quite see some of the links on the left-hand side of the board. See http://www.zimmers.net/anonftp/pub/c.../8032090-7.gif details A, B and E.

I assume that you didn't see the PET running before it went bang? A mains filter going shouldn't have killed the PET. I suspect that the PET wasn't working before it went bang - so I am going to ignore that. All sellers lie!

After that, I am seeing 'confusing' results. So, let's go back to basics and find out what it isn't in order to rule out a load of issues.

I note that the higher bank of memory (second 16K) is mounted in sockets (even numbered devices UA4, UA6, UA8, ...). I assume this was done before you purchased the machine? In which case, either the machine has been upgraded from 16K to 32K - or the upper bank of memory devices have gone faulty at some time and have been replaced.

EDIT: Can I suggest you look at the soldering around the bottom of the RAM ICs for signs of poor soldering, bad joints and solder splashes or short circuits.

Let's concentrate on the voltages first and make sure everything is OK.

What I am going to try and find out is whether a common mode fault is present that would prevent the RAM from working correctly. With my PETTESTER stuck at the page 0/1 test, only the lower bank of DRAM will be active (odd numbered devices UA5, UA7, UA9, ...).

Use a multimeter set to read DC Volts to read the voltage on the following DRAM pins (say UA5) relative to UA5 pin 16 (0V):

UA5 pin 1 = -5 Volts.
UA5 pin 8 = +12 Volts.
UA5 pin 9 = +5 Volts.

All voltages should be within a tolerance of +/- 0.25 Volts.

What test equipment do you have? Do you have an oscilloscope or a logic probe?

The next thing to do is to check with an oscilloscope (best) or a logic probe (not as good) the following signals (all with my PETTESTER sat at the failed test screen):

UA5 pin 4 (/RAS0). This signal should be oscillating (i.e. not static logic '0' or '1').
UA5 pin 15 (/CAS0). This signal should be oscillating (i.e. not a static logic '0' or '1').
UA5 pin 3 (RAM R/W). This signal should be oscillating (i.e. not a static logic '0' or '1').

Any of the above signals being static will account for the RAM appearing to be malfunctioning.

The /RAS0 signal is not so clear cut. This signal is used for both the REFRESH and CPU access - so one type of access could be working with the other not. We have seen this type of fault before. We will come back to this later though. Let's just see if there are any 'stuck' signals first.

You say you have a RAM/ROM board. Can you perform another test with your RAM/ROM board. Configure the RAM/ROM board so that only the RAM is mapped into the CPU address space and not the ROM (i.e. we will use my PETTESTER and the kernel ROM for real). Does the PETTESTER get any further now?

Map the ROM into the RAM/ROM board and see what happens now.

I am just trying to get some consistent data points on the fault.

Dave

 

[mjnurney]

Hi Dave,

that’s a comprehensive response in the past I have measured the ram and all the required voltage were present but that was 2 years ago.

I’ll try the ram rom board in tonight. It will usually boot to commodore basic with everything on. But no cursor as I have to remove a 6520 or it boots to TIM.

I have a multimeter, logic probe and oscilloscope (not sure how to use it) some guidance would be needed on that.

the board has had repairs in the past, I’ll take pics of those later.

thanks for all the info !

 

[daver2]

>>> that’s a comprehensive response.

We try - once we get a handle on where we are at.

>>> but that was 2 years ago.

A lot can change in 2 years with vintage equipment! See some of the other threads related to PDP-8 machines !

>>> thanks for all the info !

Absolutely no problem. We have got worse repaired - but this fault is 'hiding' a bit. I suspect we may have more than one fault - but that is just a gut feeling though...

OK, Use the mutlimeter and logic probe for now - we will give you a 'crash course' with the oscilloscope presently ! What make and model is your oscilloscope (so I can look up the documents on the internet).

Dave