Commodore 8026 printer / typewriter

[TonyDuell]

The processor was marked '3874'. According to the Mostek databook (on bitsavers) this was renumbered as the 38P70/02.

It's the external ROM version of the 3870 family, to be used, normally, for code development. Once your firmware works correctly you'd have a mask-programmed verson made for production.

I suspect the Olympia typewriter used a 3875 of some type as that's the one that supports battery backup of the RAM.

[SiriusHardware]

The 'zimmers' website resource (which we have made so much use of in the past when attempting Commodore related repairs) might appreciate a copy of that firmware - but maybe only when we have reason to believe its contents are intact and have not suffered data fade.

[ScottishColin]

And Tony's drawings too, if he is OK to share them.

Colin.

Originally Posted by SiriusHardware

The 'zimmers' website resource (which we have made so much use of in the past when attempting Commodore related repairs) might appreciate a copy of that firmware - but maybe only when we have reason to believe its contents are intact and have not suffered data fade.

[TonyDuell]

I have no problems with either the drawings or the firmware dump being shared elsewhere.

[ScottishColin]

I've just found an interesting set of Jerry Walker on Youtube videos fixing an Olympia ESW printer which looks very similar to mine (although it has a serial interface not a IEEE).

I need to find some time to settle down and watch them now.

https://www.youtube.com/@JerryWalker...?query=olympia

Colin.

[SiriusHardware]

I believe Mr. Walker is actually a forum member here, after you've seen the videos he may be able to share some insight with you if you can get in touch with him.

[TonyDuell]

The ribbon cable from the GPIB (IEEE-488) board in the 8026 is soldered to a set of holes on the main board that could have taken a 24 pin DIL package. Just checked and it could have been a 6850 ACIA originally. There are several other empty spaces round it that could have been RS232 level shifters, baud clock generator, etc.

Of course the microcontroller firmware would be different.

But I think a lot of my diagrams would apply to the serial version as well.

[ScottishColin]

I've watched video number 1 (of 5) and it's a very similar set of problems so far. Leaky batteries, destroyed capacitors, missing transistor.

Onto the next one now.

COlin.

[ScottishColin]

Watched them all. Very interesting. Electrical faults where the batteries had leaked which affected the reset line to the processor.

While I was digging around, I found this on GitHub which I can't swear is the same (obviously no IEEE interface) but it looks similar.

https://github.com/RoSchmi/ES100/tree/master

Colin.

[TonyDuell]

I've taken a look at those diagrams. They are of a related machine (maybe just an earlier/later version) but are different enough that you will have problems if you follow them blind. Use them for clues as to the basic machine operation, not details of what connects to what!

Interestingly there _is_ a GPIB (IEEE-488) interface in the schematics there, and it closely matches the one in the 8026. It's called an 'IEC' interface, IEC-625 being IEEE-488 on a DB25 connector, not a 24 pin Microribbon one.

[TonyDuell]

Any news on this? I wonder how Colin is getting on.

[TonyDuell]

A little more information.

With Colin taking lots of photos and measuring the resistance between points I speciifed, I've worked out the wiring for the mechanical bits both on the main chassis and the carriage. The attached .pdf gives the details.

Enjoy!

[TonyDuell]

Based on more of Colin's pictures I've produced this schematic of the power supply unit and mains wiring.

It's a ferroresonant transformer providing the motor supplies (+9V, +36V) with adequate regulation and a 3 terminal linear regulator for the +5V logic supply.

Also note there is no mains earth. The input connector takes a normal IEC cable but has only 2 pins. Colin has confirmed this is correct. I guess it's designed to be double-insulated.

[SiriusHardware]

All great work from the two of you - Colin is very methodical about taking large numbers of measurements and recording them. The wiring of the carriage coils must have taken some visualising, especially as they don't set off from the J1 connector in the easiest imaginable order.

What is the main thing impeding progress now? The cracked cog, I would guess. Does that have 'normal' teeth with the edges of the cog teeth aligned at a right angle to the flat side faces of the cog, or is it one of the awkward ones with 'skew-cut' teeth?

While I know Colin is quite expert at 3D printing I'm wondering if the old fashioned approach of pressing the original cog into a mould, removing it and casting a new one in resin / 'plasticraft' could also work?

[TonyDuell]

We're currently completing the schematic by working out the keyboard matrix.

I am not sure what the issues with the printer are. I believe there's a cracked plastic gear in the carriage drive [strictly 'cog' is a tooth, the gear a 'cogged wheel'. These terms all grate on me a bit!].

Colin can decide what to do. I'm still old-fashioned when it comes to repairs like that and would probably mill a replacement.

[SiriusHardware]

Careful, you might get the job.

I'll try to remember to use the terms 'cog' and 'cogged wheel' where applicable in future, but you should brace yourself for the possibility of there being other people who, like me, always thought a 'cogged wheel' was a 'cog'.

[ScottishColin]

I believe I can sort the gear with my resin 3D printer. We still have yet to put any current through the PCB (and I'm fine with that while we're (well, Tony is} drawing the diagrams.

I predict something won't work as I've been at the board replacing the components that have 'rotted' away from the leaking batteries.

Colin.

[SiriusHardware]

I think the main thing you replaced was what appeared to have been obvious supply rail decoupling capacitors near the power input connector - although we couldn't be sure of the values we knew what the typical value range of such capacitors is, and we knew the highest voltage they were going to need to handle, so that bit should be OK. I think you also lifted and either cleaned up or replaced the power supply input connector?

Tony mentioned (and showed through his hand drawn diagram) that the rotted transistor (now replaced?) was in the circuit which holds the uP in reset until the power is OK (and you mentioned that Jerry Walker had made a similar observation in his video which featured a unit similarly afflicted by battery rot) so I have some hope that the situation, electronically speaking, has improved quite a bit already.

[ScottishColin]

I've also replaced a couple of resistors on the print head PCB and one more resistor and a transistor on the main PCB that were also affected by the battery leakage.

You have more confidence in my soldering on what is an old PCB than I do...

Colin.

[SiriusHardware]

I really do, remember those far-off days when we were advising you to always chop the legs off ICs and remove them individually in order to minimise the possibility of damage to your one and only PCB - compare that to when you were working on the last PET and you were removing 40-pin ICs intact with great insouciance - and as far as I know you still don't have a 'proper' (electrical pump assisted) desoldering iron, so you do very well with the tools you have.

I would never have any qualms nowadays about suggesting you remove multi-pin parts from double sided PCBs. I know you'll do it all right.