Compucorp 326 Scientist -- repair info needed

[jefo...@gmail.com]

Hello, 

Does someone have experience with repairing vintage compucorp 326 Scientist Calculator

All basic voltages are present and all waveforms are correct for the various "phi"signals on the so-called "converter Board"

It seems one (or more ?) of the 4 ics (ACL-4 to ACL-07 or equivalent TCL-04 to TCL-07) on the control board are broken. Possibly also a problem with the Data Board ics.

The ACL ics are American Microsystems custom Ics. The TCL ics are 

equivalent Texas Instruments vintage custom ics. 

Does anyone have an idea how to replace such ICs ? Substitutes by rebuilding these using modern components seems unlikely ? 

But perhaps there are still spare boards around ?  or ics from unused spare parts ? 

The device here uses the American Microsystems custom ics. Does anyone has a

description of the functionality of these ics ? Does anyone has the equivalent Texas Instruments ic numbers for this calculator ? 

Any information or hint on how to best proceed with this repair is most welcome 

[La Greenall]

I don't have direct knowledge with or experience of working on this machine, but firstly it is one of a range of very similar models and it might pay to include them in your searches, and secondly this particular model was also sold in rebadged form, as:

• Ataio 326
• Monroe 326
• Sumlock Anita 326

So again, also check out any data you can find on these.
See also http://www.calcuseum.com/SCRAPBOOK/BONUS/302360/1.htm (note that this is page 1 of 2).

One website lists ten models in the range of Compucorp models (not including rebadged ones), describing each one in detail:
http://calcmuseum.classiccmp.org//compucorp_portable.htm
In a subsequent section titled "Compucorp 300-Series Internal Construction" it begins by saying that:

"The internal construction of the various Compucorp models is virtually identical, except for the later Model 326 (which is similar). Compucorp designed a general-purpose set of integrated circuits and printed circuit boards PCBs) for the machines, and used different Read-Only Memory (ROM) and amounts of Random-Access Memory (RAM) depending on the functions of the machine."

It then goes on to list the ICs in each model, including several TI ICs.

So I think this might be a good starting point for research.

Secondly, it might also be worth contacting the Centre for Computing History (based at Cambridge University), because they have a Sumlock Anita 326 and have no doubt studied it very fully (even though their page on it is rather general),  so they might be able to answer specific questions.
https://www.computinghistory.org.uk/det/16678/Sumlock-Compucorp-326-Scientist/

The Computer History Museum in California also has a Compucorp 326:
https://www.computerhistory.org/collections/catalog/102692296
Another call you could make.

I also found this online:
https://usermanual.wiki/Manual/Compucorp326Service.1128033626

Rick Bensene's webpage on Compucorp has an account of the chipsets in these machines, including the fact that PCBs were designed to accommodate the different pinouts of chips produced by different suppliers:
https://www.oldcalculatormuseum.com/d-compucorp.html
Scroll about 60% down the page.

John Wolff's site has a detailed page on one of the other models in the series, including internal views:
http://www.johnwolff.id.au/calculators/Compucorp/Compucorp320.htm
It might be worth visually comparing this with your 326 to see if the similarities are enough to be informative.

I would also try contacting Joerg Werner whose website covers most calculators containing TI ICs; his site does not have a page on the 326 but does have a page on one of the other models, so again he probably has the detailed knowledge you are looking for:
http://www.datamath.org/Related/Compucorp/325.htm

Last but not least, a teardown video and related forum thread:
https://www.eevblog.com/forum/blog/eevblog-663-compucorp-322g-calculator-teardown/?nowap


HTH,
Lawrence

--
You received this message because you are subscribed to the Google Groups "OldCalculatorForum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to oldcalculatorfo...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/89fecf4e-3046-4c9d-878b-59b49f571163n%40googlegroups.com.

[J Ongena]

Dear Lawrence, 

Many sincere thanks for your quick and extensive reply. This is very much appreciated.The website of the Centre of Computing History triggered and that of Joerg Werner 

triggered me to send both of them a summary of my questions. I am looking forward to their response, and will keep you posted !

WIth my best regards, 

Jef Ongena

Belgium

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/fb500bf5-6bc2-7e8e-de26-f15acefdc561%40googlemail.com.

[Pieter La Grange]

Hi Jef

As I am typing this reply I realize who you are. I watched your Youtube video on the Riccar repair earlier this week.

I am very active on our vintage Facebook group but only joined this group here today.

I have a Compucorp 322G that I acquired a few years ago. I spent probably 20 hours on it with no success. I came to the conclusion that one or more of the IC's must be faulty. I recently acquired another model, and I have another that I will receive in the next week. Hopefully I can with the knowledge of the other two figure out why the first one never worked.

Pieter
South Africa

[J Ongena]

Hello Pieter!

Yes I am the person that repaired the Riccar calculator. You will find more info on the website of Mr Hilpert (madrona.ca), a fantastic source of info on old electronic calculators. 

It was a wonderful experience to see this old and nice nixie Calculator working again. And what a repair it was..!

Thanks to the precious help of Mr Hilpert it was succesful.

What you are undertaking is very interesting, and thank you for sharing!

It would be extremely useful to learn from your experience with the Compucorp calculator. Can you keep me informed? I have not been able to continue recently due to overload at work. Checked signal shapes and voltages as described in the manual and all looked fine. To be able to meaningfully continue I most of all need a detailed schematic to find by reasoning what could be the cause of the problem but that will need a bit more free time.

Looking forward to receiving more news ! 

Best regards

Jef Ongena

Belgium

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/06804a7b-95a7-4389-927c-a307075827ebn%40googlegroups.com.

[men...@gmail.com]

Hi,

As you have found, there is a service manual available with quite a lot of information. These machines are really boxes that hold the IC chip sets.  They are nice and modular and wellmade, but complex so there is a chance that it is the infrastructure rather than the ICs that are not working. If the ICs are broken, though, then you have a few options (my reply is about options  and 4):

1) You can hunt down ICs. New replacements are not available. Compucorp disappeared a long time ago. I doubt that new old stock ICs even exist.

2) You can get ICs form a donor machine. There's a good chance that ICs form a non functional machine would also not work, and taking ICs from a working machine is a bit self defeating.

3) You can recreate an old IC using new components. I have done this for some things. I replace an SP0256 IC with an RP2040, as the IC is not practically or economically available. This was fairly simple and work quite well., This is a 5V IC. I also replaced a RAM chip in a Casio calculator or two with Rp2040s. This was harder as the signals weren't always simple or 5V.

https://youtu.be/qNnYEQevR-Q
https://youtu.be/njsNJoiKlUM

https://youtu.be/XX5G-ph3CFc

This option requires a lot of reverse engineering,made easier with a service manual.

4) You can replace the electronics completely and run a simulation.

I have done this as well, for the psion organiser, Casio CQ-1 and Casio RPO-101. It requires manuals and, if possible, a working machine to check the simulator against. If you are lucky then you can dump ROMs and run the original code, this isn't always possible. Dumped ROMs and emulation has the advantage of the behaviour being very close to the original and also preserves the original code., It's probably less work that simulation.

https://youtu.be/qNnYEQevR-Q

https://youtu.be/XzfspKhKC18

https://studio.youtube.com/video/b8dYikCsfaM/edit

many other people have done this, there's the HP41Cl for instance.

[Michael Briley]

Hi Everyone,

Just wanted to send out thanks to everyone for some really useful and timely info. Yesterday, with the help of the service manual, I finished restoring a 324 - (Monroe badge). Mostly corrosion issues that made it onto the boards (sigh, I hope there’s a special circle of hell for people who leave NiCds in their devices).

Cheers.

Michael Briley

Department of Physics and Astronomy

Appalachian State University

On May 11, 2024, at 12:51 AM, men...@gmail.com <men...@gmail.com> wrote:

Hi,

--
You received this message because you are subscribed to the Google Groups "OldCalculatorForum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to oldcalculatorfo...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/995d31d8-a487-4c65-94ab-84b47955971cn%40googlegroups.com.

[osa...@gmail.com]

Isn't that a thing of beauty? Amongst my "calculators I desire the most list, but are rare and very pricey!"

[J Ongena]

Hi Michael, 

Wonderful that you were able to restore the Compucorp 324 ! I would be interested in the kind of corrosion issues you encountered. Also here, the specimen I have 

has had bad batteries....  Would it be possible to send some pictures illustrating what (and where on the pcb) you encountered (hidden?) corrosion ? Did you encounter

troubles with the "multiple" stack connector interconnecting the individual pcbs and if yes, how did you solve these ? Anything else  ?

If you find time, can you let me know ? 

Many greetings and thanks in advance, 

Jef Ongena

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/D24D02FB-423C-4EA3-AE2A-B13F672852D1%40appstate.edu.

[Michael Briley]

Hi Jef,

There was corrosion on the battery terminals (the batteries themselves were hopeless and are in the recycle bin), the power switch (inside - it had to be taken apart and cleaned), the AC adaptor connector (also taken apart and cleaned), the three-pin molex connector to the batteries/power (it had to be replaced), etc. The corrosion was mostly the blue stuff (copper sulfate?) and white (potassium carbonate?). No black corrosion (which is the worst!).

But by far the biggest issue was corrosion on the edge connectors of the PCB boards and the stack connectors. Of course, I didn’t have the forethought to take many pictures, but I do have one of a PCB connector before cleaning:

The stack connectors (copper) responded well to distilled white vinegar - each connector was scrubbed with a Q-tip, followed by a rinse in distilled water, then a bath in isopropyl alcohol.

The PCB edge connectors were a bit more tricky. Vinegar took a good bit of the corrosion off, but also led to oxidization of the metal, requiring a clean with alcohol immediately after the vinegar. I probably should have tried a baking soda mix… But that, and a reversion back to the old days when we would clean edge connectors with a pencil eraser, seemed to do the trick. A number 2 pencil eraser is your friend here :)

The whole thing took quite some time (over a week) - just be patient and go slow. I was also pretty lucky in that the corrosion hadn’t done much damage to the underlying metal. I have other calculators that have suffered far far worse.

Cheers.

Michael Briley

Department of Physics and Astronomy

Appalachian State University

On May 13, 2024, at 5:40 PM, J Ongena <jefo...@gmail.com> wrote:



[Thomas Baumgart]

Hello Jef,

I just subscribed to this forum since I'm a collector of old calculators as well -- I've got two well performing Scientist 324G's, a half-way operational "Statistician" (one segment output of the display driver TCL08 shot which was my own fault :-|  -- slipped with the scope probe...) and -- just like you -- a broken 326, tape drive included. The problem with my 326 is a faulty ROM chip; it seems just one of its data outputs is high-impedance. Since all the other ROMs on the same bus appear to be okay, my take is that the corresponding bond wire inside the IC casing may have got disconnected. These special P-MOS four-bit, latched mask roms are long time obsolete, but I'm pretty sure that I could replace all the roms of the calculator with a single (flash) EPROM and a little glue logic. With a little piggyback board that emulates a DIL28 footprint, some contemporary SMD components, and some black epoxy, it may be possible to make the "Data File Assembly" board even look half-way original.

That said, if I won't be able to source the microcode somehow or find a way to read out the original, faulty ROM one last time, all the effort will be in vain. It seems that with the two broken 326's that we've both got, we could easily make one working unit, but it would defiitely be preferable to get both calculator running again. I may even have a spare CPU PCB from a 324 that contains the same four control chips as the 326. I only don't remember right now if it's the TI or the AMI chipset, which unfortunately have different pinouts.

Would you be interested in a collaboration to get both our 326 running again (if possible)? If you haven't got the service manual for this calculator, I could email you a PDF.

Cheers and all the best,

Thomas

[J Ongena]

Dear Michael, many sincere thanks for the details of the corrosion and the tips on how to clean them !

This is very helpful for the rest of the repair ! I will give all contact the treatment you describe, to be able to 

exclude bad contact issues !

With my best regards

Jef Ongena

--
You received this message because you are subscribed to the Google Groups "OldCalculatorForum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to oldcalculatorfo...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/B0651D43-5E43-4B4C-B07F-6B3358A9720B%40appstate.edu.

[J Ongena]

Deat Thomas, your proposal is great. Yes, definitely, I would like to join the collaboration ! Perhaps we could more easily exchange emails

with our private email adresses, in order not to overload the mailbox of the colleagues of the forum ? My email is jefo...@gmail.com.

I will check here what you describe. For the pinouts of the ACL ICs one can make use of schematics of the compucorp 326 that are on the internet. That documentation contains for all custom made chips the pinouts, both for ACL 

and TCL, except for the printer output chip. I have backdecoded the TCL printer version, except for one pin. Could send you all this, assuming it is useful for you ? If you send me your email address we could start !

Best regards,

Jef 

--
You received this message because you are subscribed to the Google Groups "OldCalculatorForum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to oldcalculatorfo...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/b51b1b0e-bb98-443b-92cf-d6320e7276f8n%40googlegroups.com.

[men...@gmail.com]

Hi,

i think it's a bit more complicated than a simple piggy back board as these devices use voltages other than 5V, I'm pretty sure. So you would need some level shifters. the interface can be odd with some of these older chips as well.

it's possible though, i have used the RP2040 and level shifters (or sometimes not) to replace RAM chips, see here:

https://www.youtube.com/watch?v=QJGJ5JRpquA&t=92s

The RAM chip in that IC uses odd voltages and also has a strange interface, it's part parallel and part serial. If you do add a chip like the RP2040 then you get the bonus of being able to dump the RAM and alter the contents, so you can store the RAM on SD card with some work, for instances, and also examine the internal code structure and sometimes alter it.

To dump the ROMs you could do something like this:

https://www.youtube.com/watch?v=PYATa5GiB9w

Regards

Andrew

[Louis Angelucci]

Do you collect TI vintage calculators as well?

If so let me know

thank you

Lou Angelucci

--
You received this message because you are subscribed to the Google Groups "OldCalculatorForum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to oldcalculatorfo...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/5038cdb4-c282-41a3-9160-67caf9bbe5e6n%40googlegroups.com.

[men...@gmail.com]

I do, if you mean me. I have TI59s and 58s, and have just got an SR-60A, which is working quite well. I'm going to have a go at making some magnetic cards for it as it didn't come with any.

 I have attached a TI59 to an RP2040, but my emulator code doesn't agree with the calculator on one of the initial instructions. No idea why at the moment. 

regards

Andrew

[Dennis Tuckerman]

I have a number of spare TI Plastic Wallets with both TI and blank cards inside + standard Master Library Module 1 that you are welcome to have.

I do have a current issue if anyone could help me with please:-

My degree and academic training is in Pure Mathematics and Computer Science (graduated 1984, Ordinary Member of The London Mathematical Society since 2016) so I’m fine with algorithms and coding issues BUT absolutely hopeless with anything practical (not helped with getting Rheumatoid Arthritis in 1996!) - my wife has banned me from using or having Superglue in the house!! I have about 5 TI-59s in various states of disrepair that I’ve acquired over the years. I also have a TI PC-100A Printer and I can get these 5 TI-59s to work when attached to the printer. The printer came without a key but I’ve had one cut and taking the TI-59 that was on the cradle out revealed it contained the Leisure Library Module 7. However recent the PC-100A has started playing up. I can attach a calculator to the cradle, lock it and it’ll power the TI-59 BUT it no longer prints or responds to Trace - with or without anything attached the printer will light up and power the TI-59 but the PRINTER, TRACE and ADV buttons no longer work. Any ideas?

Thanks 

Dennis Tuckerman 

Sheffield UK

Leave the Planet (tlalticpac) in a better shape than when you first found it

I’m Undead, I’m not Unreasonable (Dracula)

On 18 Aug 2024, at 19:49, men...@gmail.com <men...@gmail.com> wrote:

I do, if you mean me. I have TI59s and 58s, and have just got an SR-60A, which is working quite well. I'm going to have a go at making some magnetic cards for it as it didn't come with any.

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/f171ac60-edb3-4ad8-86a9-1913c666a54bn%40googlegroups.com.

[osa...@gmail.com]

Hi Denis, I would love the Leisure module. I have a PC-100A cradle, the version with the internal power switch that allows multiple models to be powered. Mine also no longer prints, but it did. According to the experts on the Facebook groups it will require a complete teardown and repair. These printers were thermal; the thermal heads can give up, as can the paper. However, I don't know where you can get thermal paper to just try, what would be, an extremely simple repair. Oddly, I've just pulled mine out of storage with a view to making it accessible for undertaking repairs. Maybe we could swap finding and any repairs we make? I am in the UK as well, down and across in Suffolk.

[Dennis Tuckerman]

I’ve always wondered whether the ‘old’ in ‘oldcalculatorforum’ applied to my calculators or to me? As I’m 76 this year I’ve decided it must be both!

So I’ve lost the plot on who I offered both (1) a plastic TI wallet with Master Module 1 + cards and (2) the Leisure Module (7)

Both are free of charge - just send me a PM with delivery details

Dennis 

Leave the Planet (tlalticpac) in a better shape than when you first found it

I’m Undead, I’m not Unreasonable (Dracula)

On 18 Aug 2024, at 19:49, men...@gmail.com <men...@gmail.com> wrote:

I do, if you mean me. I have TI59s and 58s, and have just got an SR-60A, which is working quite well. I'm going to have a go at making some magnetic cards for it as it didn't come with any.

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/f171ac60-edb3-4ad8-86a9-1913c666a54bn%40googlegroups.com.