• Please review our updated Terms and Rules here

Diagnose issues: PET 4032 with PETTEST ROM

irataxy

irataxy

Member
Joined
Jun 10, 2022
Messages
46
Location
San Jose, CA
Hello,

I recently fixed the power problems with my PET 4032 but noticed that the display would not come on. The computer would also chime on startup. I bought a ROMulator with PETTEST ROM to help diagnose the issue. On startup, this is what I see now. In looking at the manual, I think this indicates errors on pages 0 and 1 of the RAM. Is this correct? If so, should I look at replacing some chips on the board? What is the next step for me here?

Thanks,
NIck

IMG-1009.jpg
 
Hi Nick,

As this is your first post - welcome to VCFED!

Don't forget to put your general location into your profile so we know where you are in the world.

Don't forget that you will also be moderated for the first ten (or so) posts - so it will take a while for them to appear. Just hang in there.

As the author of the (updated) PETTESTER, your assumption about page 0/1 RAM is 100% correct.

This could be a RAM fault. It could also be a data line or address line fault also (data line 0 or address line 0). Let me have a more closer look at the patterns returned. There is a consistent pattern in there - we just have to 'tease out' what it is telling us!

Do you have any test equipment - or access to any? Multimeter, logic probe, oscilloscope?

Dave
 
I assume you have found all of the schematics for your machine over at http://www.zimmers.net/anonftp/pub/cbm/schematics/computers/pet/2001N/index.html?

Are your RAM devices in sockets (by any luck of the draw)?

My gut feeling would be to look at replacing 4116 DRAM chip I9 - this is responsible for data bit 0 of the LOW 16K bank of RAM.

If that doesn't work - device E3 (74153) could also be a good candidate.

Do you have spares? If so, you could try the 'piggy back' test first.

Please ask if you are unsure of how I have arrived at this conclusion - or there is anything you are unsure of BEFORE you do it!

Dave
 
Dave - Am I right in thinking that V4 of your PETTester code is the version included in the bitfixer ROMulator's ROM images?

Nick - Can you post a photo' of the main board? That would make it easier to find the correct schematic for your PET. There were several different versions of the 4032 but your's is most likely to be covered in this document:


The RAM chips are likely to be soldered rather than socketed and they probably will be 4116s as Dave suggests.

Welcome to the forum!

Alan
 
>>> Dave - Am I right in thinking that V4 of your PETTester code is the version included in the bitfixer ROMulator's ROM images?

As far as I am concerned, that is correct.

Dabe
 
I have a multimeter, but no other test equipment. I am attaching a picture of the main board. Most every chip is soldered directly.

I don't have any spare chips. Are there recommended replacements? Also, I am at a very beginner level with soldering. Will this be above my skill level?


IMG-1010.jpg
 
That's what I was afraid of - most things (except the CPU, character generator and EDIT ROM) are soldered in.

You can purchase 4116 DRAM ICs from surplus electronic shops. You need to make sure they are faster than 300 nanoseconds access time. Yours were made by ITT in week 47 of 1981 and are 300 nanosecond. If you check ebay.com I am sure you can find something reasonable. However, ebay can be a bit of a cr*p shoot of course.

You can try and piggyback this replacement chip on top of I9. You need to be careful doing this of course - and there is no guarantee that this will work - but it is something simple that can be done to start with. We can talk you through this step if you want to try it.

If you are not competent at soldering and/or desoldering, I would suggest either practising before you attempt any sort of repair on this machine (otherwise you will wreck it further) or find some helpful electronics person in your area.

With the ROMulator you can also try various RAM/ROM replacement scenarios to see if other parts of the PET work OK (e.g. the keyboard etc.).

Dave
 
Really clear photo'. I'm afraid you'll probably need more than a meter to repair your PET. An oscilloscope would be a basic requirement unless a fault can be pinpointed using the ROMulator alone. If components need replacing you'll need reasonable soldering skills but these can be acquired with practice. Much depends on how keen you are to have a go yourself. The schematic you need is the first one in the document referred to in my earlier post.

Alan
 
Last edited:
I assume you have tried the romulator with switch index 3, (11000000) i.e. 1 and 2 on, all others off.
This would replace all the onboard ram. (except for the video ram, but yours seems to be working fine)

If it won't come up with that switch config, something is holding the bus either high or low on the data bit 0 apparently.
 
>>> something is holding the bus either high or low on the data bit 0 apparently.

This statement is not correct. Data bit 0 is being held LOW - but only for a certain address range. After that, it is fine. This implies it is NOT a permanently stuck data bit. Hence we can rule out certain things (like the data bus buffers). This could be as a result of an address fault - or a fault within the DRAM IC itself. These are prone to weird failure modes (depending upon the device construction - and internal arrangement of memory cells).

Dave
 
One thing I’ve noticed is that in addition to the PETTester V4 image the ROMulator has a ROM image described as ‘PET RAM/ROM/VRAM Test’. I’ve no idea what it does but it might be helpful if you’ve been provided with instructions. I couldn’t find anything about it on the web.

Alan
 
If I remember correctly, I think this is more targeted at the original PET with the static memory chips.

It provides more detailed diagnostic information on the specific chip to replace - because it is constrained to a limited subset of PETs. However, the testing is relatively limited (especially if you have dynamic RAM).

Dave
 
Maybe it's just me but it really bothers me that the board is so dirty and rusty. Wouldn't it be better to clean everything up? Or could it introduce more problems?
 
Cleaning it up, if done right, would probably be a good idea. When it's that much of a mess there's all sorts of possibilities for semi-conductive gunk between pins or traces to cause random havoc.
 
You can purchase 4116 DRAM ICs from surplus electronic shops. You need to make sure they are faster than 300 nanoseconds access time. Yours were made by ITT in week 47 of 1981 and are 300 nanosecond. If you check ebay.com I am sure you can find something reasonable. However, ebay can be a bit of a cr*p shoot of course.
Click to expand...

Just to clarify, I need to buy replacement 4116 DRAMs with greater than 300 ns access time? I haven't seen any of those in my (limited) searches.
 
On the topic of 4116 DRAM IC's there are a number of them out there. The ones you need have to be at least a 300nS access time (no longer), normally they have a suffix on the part number like 4116-3 or similar.

There are "faster ones" out there like 4116-15nl, these have a 150nS access time so they work fine in the PET, but the faster response is not required for the PET. I recently got a batch of the 15nl variants where all 20 were defective, I think they might be fakes.

The piggyback idea is imperfect.

The idea of it is; if the DRAM chip is defective and not responding, if you piggyback another good chip on top of it (all pins connected together of two IC's-the original one and the piggyback one) the good chip will "take over". But many a time the output of the defective DRAM will be stuck, so the process does not work and in my view has very limited utility as a "reliable diagnostic technique" unless you "get lucky", but good fault finding techniques should not rely on luck, if they did, then the average drunken punter at a Casino would have better luck fixing a broken computer than a trained technician.
 
You must log in or register to reply here.
Back
Top