PDA

View Full Version : Victory! Sol-20 repair



ef1j91
November 20th, 2017, 03:35 AM
I had to share a hard-won victory getting this Sol-20 up and running. I'll post repair notes to my site soon. It involved a few iterations of power supply, PC board, and keyboard troubleshooting.

Next up is refurbishing the keyboard, a topic of much discussion on the board. I'm planning to punch out my own foam pads and was wondering what recommendations were out there for particular foam products? What thickness and foam density work best? I was going to try a lighter weather-seal type foam. Also: the keyboard is the type without the pad clips and the foam inserts are held in with an adhesive membrane.

42032

Corey986
November 20th, 2017, 05:32 AM
Use low density foam or it won’t feel right. I found only one package/thickness of this foam at Home Depot and not all carry it. I use space blanket material from Walmart. One blanket is enough for many keyboards. Then use permanent double sided tape or contact cement to attach it to the plastic sheets to make the circles, which you can buy uncut sheets from a hobby store. Then you punch it all out together. I don’t remember the exact thickness of the plastic, but if you look at a current disk and then go to the store it should be obvious.

DaveL
November 20th, 2017, 05:35 AM
Congrats! I'll be very interested in what the power supply needed (and where you found replacement caps, if any).

ef1j91
November 20th, 2017, 07:23 AM
Congrats! I'll be very interested in what the power supply needed (and where you found replacement caps, if any).

I ended up replacing Q1 and Q2, the Darlington pair on the 5V line and the MC1458P op amp that is part of the regulator circuit. Caps seem to be fine and powered it all up initially with a variac. Oh, and I needed to replace the fuse holder, but Littel still manufactures the same fuse block.

snuci
November 20th, 2017, 08:01 AM
Use low density foam or it won’t feel right. I found only one package/thickness of this foam at Home Depot and not all carry it. I use space blanket material from Walmart. One blanket is enough for many keyboards. Then use permanent double sided tape or contact cement to attach it to the plastic sheets to make the circles, which you can buy uncut sheets from a hobby store. Then you punch it all out together. I don’t remember the exact thickness of the plastic, but if you look at a current disk and then go to the store it should be obvious.

Just to give some more ideas...

I used weather stripping that was of good consistency. This already had one side sticky. I stuck that side to shiny wrapping paper from the dollar store. I then used two sided tape on the foam side and used overhead transparency sheets for the plastic disks to make the the rigid side that fit into the key switches. I punched the plastic disks separately but did everything else including the two sided tape with one shot. That worked out for me.

NF6X
November 20th, 2017, 04:16 PM
Congratulations! I've gotten my SOL-20 working-ish, but my first attempt at rebuilding the keyboard didn't work. Maybe I'll get around to trying again over the holidays?

Corey986
November 21st, 2017, 05:04 AM
I ended up replacing Q1 and Q2, the Darlington pair on the 5V line and the MC1458P op amp that is part of the regulator circuit. Caps seem to be fine and powered it all up initially with a variac. Oh, and I needed to replace the fuse holder, but Littel still manufactures the same fuse block.
Post the part number for the fuse holder. I’d like to order some spares.

Thanks,
Corey

ef1j91
November 21st, 2017, 08:25 AM
Post the part number for the fuse holder. I’d like to order some spares.

Thanks,
Corey

Jameco part no.: 1711701
Manufacturer: Littelfuse
Manufacturer no.: 03420038H

http://www.littelfuse.com/~/media/electronics/datasheets/fuse_holders/littelfuse_fuse_holder_342_datasheet.pdf.pdf

ef1j91
November 22nd, 2017, 05:42 PM
Almost there...

42066

falter
November 25th, 2017, 12:36 PM
Question.. I am looking around for 'space blankets' here in town to get my keyboard repaired.. none are made from mylar... they all appear to be aluminized something or other now.. will that work? Or does it have to be mylar?

Chuck(G)
November 25th, 2017, 12:54 PM
"Mylar" is a trademark by DuPont for Polyethylene Terephthalate (PET) film. Most non-infringers generally call it "polyester" or sometimes just "polyethylene".

For example: this (https://www.mec.ca/en/product/0199-034/Emergency-Blanket).

snuci
November 25th, 2017, 01:02 PM
Question.. I am looking around for 'space blankets' here in town to get my keyboard repaired.. none are made from mylar... they all appear to be aluminized something or other now.. will that work? Or does it have to be mylar?

Try Dollar Store metallic looking wrapping paper.

ef1j91
November 29th, 2017, 03:04 AM
I wrote up details of the Sol and a summary of my repairs in case this helps anyone else:

https://sites.google.com/site/retroborkenwerk/sol-20

https://sites.google.com/site/retroborkenwerk/sol-20/basic-function

I haven't had a chance to test the foam inserts for my keyboard yet, but if they work, I'm happy to share my now seemingly lifetime supply of metalized Bo-PET film.

ef1j91
December 3rd, 2017, 05:11 PM
Keyboard repaired.

However, had my first tantalum 'splode testing a RAM board in the machine. A good pop and small flame. I had tested for continuity / resistance, but that wasn't enough.

One culprit may be the voltage going to the cards. The unregulated power on the Sol expansion card seems to be running high. I measure 11V on the 8V line and 19.7V on the 16V line (-19.5 on -16V line). The voltages on the regulated outputs seem to be fine.

Any tips / thoughts?

DaveL
December 5th, 2017, 04:12 PM
Can you check your transformer for a "4 3991" marking? If it has that you need Change Notice #9 from here (http://deramp.com/downloads/processor_technology/sol-20/hardware/manuals/SOL-20%20ECNs.pdf). Otherwise, something bad is going on in the power supply.

Since it's unlikely you'll find the parts needed by CN9 (if you need it) you might begin thinking about some switchers.

ef1j91
December 6th, 2017, 02:55 AM
It's the T2 transformer. I'll go in at some point and double check the transformer connections. Otherwise, not sure what could have gone bad. The circuit is just the transformer, rectifier, smoothing cap and bleed resistor.

The Sol and DRAM board are working fine, but 11V on the 8V line is towards the upper operating limit of the 7805 regulators. It could come down as I add more cards.

Does anyone else have measurements of their Sol PSU outputs, especially on the unregulated side?

ron shenk
February 5th, 2019, 11:51 AM
I wrote up details of the Sol and a summary of my repairs in case this helps anyone else:

https://sites.google.com/site/retroborkenwerk/sol-20

https://sites.google.com/site/retroborkenwerk/sol-20/basic-function

I haven't had a chance to test the foam inserts for my keyboard yet, but if they work, I'm happy to share my now seemingly lifetime supply of metalized Bo-PET film.

I am having trouble with my Sol20 keyboard repair. I have tried several materials for the "silvered mylar" component
with no luck, this includes space blanket material, dollar store metalized paper, harddrive static bags, aluminum
model airplane aluminized covering. (Note small 1/4 inch aluminum foil dots elicit a key, but the wrong one, e.g.
pressing D gives C and so on.)

My finger works as does the material from a Sun type 4 keyboard.

If you have a "lifetime supply of metalized Bo-PET" I could stand some.

ron shenk

Hugo Holden
February 5th, 2019, 07:03 PM
Just some tips from my Sol-20 keyboard experience;

The capacitance increase required to reliably trigger the key detect circuitry is around 10pF to 15pF. The original film was around 0.06 to 0.07 mm in thickness, but the conductive material appears to be in the center of the layer, roughly about 0.035 mm from the surface. However the replacement foam discs I bought from ebay had about the same thickness plastic film, but the conductive layer was on the foam surface, placing it about 0.07mm from the keyboard pcb surface with the key pressed, which did not raise the capacitance high enough on some keys to reliably detect the key press.

And there is more.... The Sun 4 pads are perfect for the job and don't have the above problem. However, although my keyboard looked in most respects identical to others, the method which retained the foam discs was different. In most it would appear that small plastic tongues on the plunger base retain the foam disc and its stiff plastic base in position and the discs "click into place". My keyboard must be an early model without these. The originals were attached with adhesive. So to get the Sun 4 discs to work I had to punch out discs of ultra- thin double sided tape fit these to the Sun 4 discs and fit those to the base of the key plungers as the discs would not self retain. I was not prepared to put any glues or liquid adhesive near it.

Experiments also showed that the keyboard could be reliably triggered with a DC link between the pcb pad areas, anything less than about 68K triggers it. The output pulse width from the detector circuit becomes wider compared to the capacitive coupling case (increases from about 800nS to 5uS), but the keyboard function remains normal. I bought some conductive rubber discs (calculator or TV remote control spare parts) which have a low Ohmic resistance, this would also have worked, but in the end I went with the Sun 4 discs, but the rubber ones are an option if you cannot get the Sun 4 discs or others that work.

That was not the end of my keyboard woes. The keyboard would intermittently stop from time to time, it took quite a while to get to the bottom of it. One of the counter IC's had an intermittent output and randomly fall to zero while all the inputs to the flip flop doing this were normal, as was the power supply to the 7493 IC. 74 series IC's seldom go intermittent like this and it was not temperature related. Finally I tracked that down with the scope and replaced the IC. I have had no keyboard troubles since.

ef1j91
February 6th, 2019, 03:19 AM
I've had intermittent keyboard issues as well. Usually LOCAL, UPPER CASE, ALPHA LOCK become unresponsive. Other keys that have issues are W, Q, :, and sometimes E. They will work for a while, then over time (days) stop working.

A sure fire way I've found to restore the keyboard function is to take it out and lay it flat on my bench. This makes me think there is a mechanical component involved; a cold solder joint or marginal trace somewhere on the PCB. After I do this and reinstall the board, it works fine... for a few days.

I also had the earlier keyboard that used adhesive membranes instead of the clip-in pad.

Hugo Holden
February 6th, 2019, 02:31 PM
I've had intermittent keyboard issues as well. Usually LOCAL, UPPER CASE, ALPHA LOCK become unresponsive. Other keys that have issues are W, Q, :, and sometimes E. They will work for a while, then over time (days) stop working.

A sure fire way I've found to restore the keyboard function is to take it out and lay it flat on my bench. This makes me think there is a mechanical component involved; a cold solder joint or marginal trace somewhere on the PCB. After I do this and reinstall the board, it works fine... for a few days.

I also had the earlier keyboard that used adhesive membranes instead of the clip-in pad.

Sounds somewhat similar to mine.Though mine was not affected by movement but it was a very random fault, disappearing for days, sometimes. It is possible that it could be a similar fault. Given the somewhat rare nature of the type of failure that the 7493 IC had (in my keyboard) it might possibly be a manufacturing issue. The IC was a Fairchild brand (usually incredibly reliable) We probably have the same vintage keyboard, both with the adhesive membranes. When the fault occurred on mine, I noticed that some of the keyscan pulses vanished intermittently and that the clock oscillator was still running and I was lucky enough to trace it back to one of the 7493's while the output pulses from it (from the first flip flop within it) were coming and going. It was very random.

I could suggest hanging the scope on the keyscan line associated with the keys that deactivate with the fault and seeing if the keyscan pulses have dropped out at the time of failure (they probably will have) then trace that back, or you could simply as a punt, replace the two 7493's in the keyscan generator circuit and you might get lucky if it was the same fault as mine.