UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Specific Vintage Equipment > Vintage Computers

Notices

Vintage Computers Any vintage computer systems, calculators, video games etc., but with an emphasis on 1980s and earlier equipment.

Closed Thread
 
Thread Tools
Old 17th Mar 2021, 2:02 am   #1081
ortek_service
Octode
 
ortek_service's Avatar
 
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,392
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by SiriusHardware View Post
On the circuit diagram, inside the chip outline, there are arrows helpfully indicating the mode / direction in which I/O pins are used. Some are INs, some are OUTs and some are IN/OUTs. DIAG is one of those.

I wouldn't ask anyone to do anything to an IC pin without a better idea of what it does.
You could always try pulling it high / low via a (4.7k etc) resistor, just in case it's not configured as an input. I've previously done this on systems where I didn't have much info / schematic, in order to work out what are I/P's etc.
ortek_service is offline  
Old 17th Mar 2021, 2:33 am   #1082
Refugee
Dekatron
 
Refugee's Avatar
 
Join Date: Apr 2012
Location: Worksop, Nottinghamshire, UK.
Posts: 5,549
Default Re: Non-working Commodore PET 3016

Can you run off a list of RAM chips that you could use so that I can have a look in the chest of drawers I rescued from a computer repair shop that shut down a good few years ago.
4116 might be close.
256K ones will be far too high but I might have 64K ones but they are still a bit high.
Refugee is offline  
Old 17th Mar 2021, 3:53 am   #1083
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
Default Re: Non-working Commodore PET 3016

There was a list of compatible memory in the pdf in post #1035.

I think the links could be changed to use 8 of 4116 type instead of 16 of the 8k.
Mark1960 is offline  
Old 17th Mar 2021, 8:57 am   #1084
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Non-working Commodore PET 3016

I think it's 4108s ideally as that would directly replace any faulty devices found, or possibly 4116 as Mark says. Slightly premature though, any apparent system RAM fault could still be down to a dud 74LS244 buffer.
SiriusHardware is online now  
Old 17th Mar 2021, 9:27 am   #1085
ajgriff
Nonode
 
ajgriff's Avatar
 
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,583
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by SiriusHardware View Post
I think it's 4108s ideally as that would directly replace any faulty devices found ....
Much googling suggests that they are unobtainable as NOS at any price and there doesn't seem to be a drop in equivalent either. I may have missed something of course.

Alan
ajgriff is online now  
Old 17th Mar 2021, 9:28 am   #1086
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
Default Re: Non-working Commodore PET 3016

I think the 4108s might be 4116s that have a defect in one half of the memory. So a 4116 could probably be used as a replacement for a defective 4108.

I think its the same as using 4164 in place of 4132.

I can check the datasheets in more detail if we think we need to find replacements for some of the 4108s.
Mark1960 is offline  
Old 17th Mar 2021, 10:00 am   #1087
ajgriff
Nonode
 
ajgriff's Avatar
 
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,583
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by ajgriff View Post
One thing worth noting is that Colin's PET is fitted with sixteen TI 4108 dynamic RAM chips (see post #8) whereas the circuit diagram we've been working from has sixteen 4116 chips ie, as in the PET 3032. This is significant because I can't find anywhere currently selling legacy 4108 chips. On the other hand NOS 4116s are still available, possibly because some versions of the Spectrum used them. In fact there's one vendor offering eight 4116s for £10 plus £1.60 p&p. In theory and with changes to jumper settings it would appear that eight 4116s could be used as direct replacements for the sixteen 4108s in Colin's machine.
If the buffers check out and with Colin's ability to socket multi-pin ICs it's well worth considering these:

https://www.ebay.co.uk/itm/STC-4116-D-memory-ic-ram-QTY-8/393015482007?hash=item5b818c5697:g7kAAOSwMFhfrqUA

The only problem is I can't find a datasheet for these STC chips but the seller seems to have dozens of them so upgrading to 32k could even be an option.

Alan
ajgriff is online now  
Old 17th Mar 2021, 10:03 am   #1088
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Non-working Commodore PET 3016

Quote:
...(4108) unobtainable as NOS at any price
Welcome back.

I did remember you said that, but Refugee was saying he had an assortment of old ICs that he could look through, so who knows?
SiriusHardware is online now  
Old 17th Mar 2021, 10:20 am   #1089
ajgriff
Nonode
 
ajgriff's Avatar
 
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,583
Default Re: Non-working Commodore PET 3016

Definitely worth seeing what Refugee can unearth.

*For Refugee's benefit the chips originally used by Commodore are listed in the PDF document attached to post #1035.

Alan

Last edited by ajgriff; 17th Mar 2021 at 10:30 am. Reason: Additional comment
ajgriff is online now  
Old 17th Mar 2021, 12:54 pm   #1090
ortek_service
Octode
 
ortek_service's Avatar
 
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,392
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by ajgriff View Post
Quote:
Originally Posted by ajgriff View Post
One thing worth noting is that Colin's PET is fitted with sixteen TI 4108 dynamic RAM chips (see post #8) whereas the circuit diagram we've been working from has sixteen 4116 chips ie, as in the PET 3032. This is significant because I can't find anywhere currently selling legacy 4108 chips. On the other hand NOS 4116s are still available, possibly because some versions of the Spectrum used them. In fact there's one vendor offering eight 4116s for £10 plus £1.60 p&p. In theory and with changes to jumper settings it would appear that eight 4116s could be used as direct replacements for the sixteen 4108s in Colin's machine.
If the buffers check out and with Colin's ability to socket multi-pin ICs it's well worth considering these:

https://www.ebay.co.uk/itm/STC-4116-D-memory-ic-ram-QTY-8/393015482007?hash=item5b818c5697:g7kAAOSwMFhfrqUA

The only problem is I can't find a datasheet for these STC chips but the seller seems to have dozens of them so upgrading to 32k could even be an option.

Alan
STC 4116 are just standard generic 4116 (which are 3 supply, whereas the 4816 used in BBC's are 5V only)
https://www.petervis.com/Sinclair/Si...tatic_RAM.html
- The ZX Spectrum had many of these ones (2N = 200ns?) originally fitted.
Although it's not clear what speed 4D marked ones are. It doesn't seem that DRAM were ever 400ns, and that a higher single number on some were actually faster (and same coding may be used by other manufacturers): https://www.electronicspoint.com/for...m-parts.65261/


TMS4532 and OKI M3732 were the one-half reject faulty 4164's (= M3764's) (single rail) DRAM's

DRAM IC's are inherently a 'square array', due to Row/Column addressing.
So standard all-working sizes will be 1, 4, 16, 64, 256 etc.
Most were 1bit only, so needed 8off.
But later on 4bit ones (usually with 1 changed to 4 / an extra 4 inserted into part no.) were released, to minimise chip-count. And revised versions of many computers (and PC memory modules) used these.

Last edited by ortek_service; 17th Mar 2021 at 1:06 pm.
ortek_service is offline  
Old 17th Mar 2021, 1:37 pm   #1091
ajgriff
Nonode
 
ajgriff's Avatar
 
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,583
Default Re: Non-working Commodore PET 3016

Looks promising if the buffers turn out to be ok and Refugee doesn't manage to find anything suitable.

Alan
ajgriff is online now  
Old 17th Mar 2021, 2:10 pm   #1092
Refugee
Dekatron
 
Refugee's Avatar
 
Join Date: Apr 2012
Location: Worksop, Nottinghamshire, UK.
Posts: 5,549
Default Re: Non-working Commodore PET 3016

I only had a quick look last night.
I have definitely got 4164 and 4116 and several other bigger sizes like 256.
Has a chip select line rearrangement board been looked at?
If anyone has reverse engineered the PROM data to see 64K with an adapter board the land freed up by removing the faulty chips would easily accommodate it. It could then be repopulated with 4108s if any are found later.
Refugee is offline  
Old 17th Mar 2021, 2:19 pm   #1093
ajgriff
Nonode
 
ajgriff's Avatar
 
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,583
Default Re: Non-working Commodore PET 3016

That's great news! Out of interest how many 4116s do you have going spare?

Alan
ajgriff is online now  
Old 17th Mar 2021, 2:31 pm   #1094
Refugee
Dekatron
 
Refugee's Avatar
 
Join Date: Apr 2012
Location: Worksop, Nottinghamshire, UK.
Posts: 5,549
Default Re: Non-working Commodore PET 3016

I have not counted them or established if the drawers have all the same type in them.
Most of them are 1980s date codes.
The chests of drawers were simply rescued and incorporated into my component store with any empty drawers populated with modern components as I needed the space. I am not sure exactly what I have got.
There are some chips with very faded part numbers so a look with a digital microscope might reveal something. I will have a better look later today.
Refugee is offline  
Old 17th Mar 2021, 2:33 pm   #1095
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Non-working Commodore PET 3016

I think plan 'A' is to keep the machine as close to its current configuration as possible, so assume for the moment that there are no plans to add non-standard expansions. It may certainly be necessary to fit a set of 4116s or possibly just half-enabled 4116s as 4108 substitutes, if that is possible.

We should really wait and see what comes back from Colin, but I really appreciate Refugee having taken the initial step of looking to see what he has.
SiriusHardware is online now  
Old 17th Mar 2021, 6:44 pm   #1096
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Non-working Commodore PET 3016

I've just been doing a bit of looking about. The pinout for the 4108 seemed remarkably hard to find but I got it eventually, so attached, the pinouts for the 4108 and 4116 respectively. From a connection point of view they seem identical so presumably the 4108 has nothing in the 'unused' half. I think Mark suggested as much earlier.

Am I right in thinking, in that case, that if we keep the machine linked for 16 * 4108 as it is now, we can replace individual suspect 4108s with 4116s without changing anything else, at least until the machine is up and running?
Attached Thumbnails
Click image for larger version

Name:	4108_4116_Pinouts.jpg
Views:	84
Size:	41.2 KB
ID:	229429  
SiriusHardware is online now  
Old 17th Mar 2021, 7:01 pm   #1097
ajgriff
Nonode
 
ajgriff's Avatar
 
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,583
Default Re: Non-working Commodore PET 3016

Seems plausible but I'm no expert as you know. Do you have a straightforward strategy for identifying faulty 4108s?

Alan
ajgriff is online now  
Old 17th Mar 2021, 7:18 pm   #1098
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Non-working Commodore PET 3016

In the first instance, just to see if there are any suspicious differences in the signal levels on the data connections from the RAMs. As each RAM IC has one bit input and one bit output and shares its single-bit connection with only one other RAM, any visible anomaly on one of those nodes will hopefully narrow it down to 2 RAM ICs and the associated 74LS244 buffer.

Then there is also the possibility of either swapping the high and low banks in the hope that one is working, and / or switching off one half to see what difference that makes.

We also have Slothie's test code, and I've also found another diagnostic ROM image which looks quite comprehensive and may also attempt to identify the actual faulty area. Obviously both of these will require EEPROMs or EPROMs to be programmed and an adaptor from the programmed device to the pin outline of the original PROM to be made, so that won't be the first line of attack, although it may not be far behind.
SiriusHardware is online now  
Old 17th Mar 2021, 7:26 pm   #1099
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
Default Re: Non-working Commodore PET 3016

I think the hope is that by looking at each pair of inputs and outputs on the buffers we can see a problem on one of the data lines. Then by comparing that data line with the buffer control signals to determine if its the buffer failing to drive correctly or one of the two ram chips. Then from the ram control lines which of the two ram chips on that data line.

A ram test rom would probably help a lot. Its going to be difficult to capture ram access on the scope as the pet software is not going to be making regular access to ram.
Mark1960 is offline  
Old 17th Mar 2021, 8:05 pm   #1100
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Non-working Commodore PET 3016

Indeed, we should probably also have Colin check the RAS and CAS refresh signals going to each DRAM as well. Given the seemingly random pattern of IC failures we have found so far, it would be quite possible to have an as yet undiscovered fault in the refresh circuitry.

There was a recent discussion elsewhere on the forum about the relative merits of plastic vs. ceramic IC packages and there was some mention of plastic packages being prone to damage by moisture ingress. From the condition of the PROMs I have no doubt that the machine spent some time in a damp environment somewhere.

I've only belatedly realised there is no CS or CE pin on these RAMs so that's why the UI10 / UI11 buffers are there, to gate the connection between the RAM inputs/outputs and the buffered address bus.
SiriusHardware is online now  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 11:40 am.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.