S-2000 contour repeat issue ?

[papz]

I got a pair of new 34063 and replaced both. Now it' fixed for good

[buzzlegs]

I got a pair of new 34063 and replaced both. Now it' fixed for good

cool, what exactly did you do?... what is a 34063?... might try and get a tech to fix mine as I'm clueless about fixing things myself and soldering... thanks!

[papz]

The 34063 is a DIP8 IC used for voltage regulation.
There are 2 in the S-2000, socketed so no soldering is needed, you just need to remove them from the socket and place the new ones, paying attention that no leg is bent and the orientation is good.
It's a common and cheap part, easy to find for example on ebay.
Hope this will fix yours too.

[buzzlegs]

Brilliant, I've ordered a couple of new 34063's, hopefully that'll work, many thanks!

[metasonix]

Sorry you had a problem (I didn't see this last week because we were evacuated due to the fire). This is apparently due to a certain production run of 34063s made by ST. Their new ones don't do it and Motorola-branded ones don't do it either. The bad run of chips also run too hot. And it varied from unit to unit....

It's so damn frustrating for a major manufacturer to sell a whole production run of chips with weird little problems of this kind. We complained to Mouser about it and they said "sorry, no one else has complained" and so forth.

If you have an S-2000 that does this, adjust both the rise and fall controls until they pulse in a stable manner. Or just replace the 34063 chips.

[papz]

Thanks for confirming

Agreed on the frustration due to faulty parts, I've had problems with a batch of ceramic caps some of which fry and get shorted, which is not fun on power rails. Of course they didn't fry during the tests but after a few hours of use by the customer... or more, someone just reported a fried cap in a unit that I built for him 1 year ago.

[metasonix]

I just built another S-2000 for sale (they sell VERY slowly now, maybe one every 4-6 months). Even with the "better" Motorola 34063 chip it does the "jumpy oscillation" thing occasionally. No idea what to do except to hang more capacitors onto the board or somesuch. Arrgh.

Maybe we should just discontinue it. The first 40-or-so pieces didn't do this and most of the production units made after that didn't do it. What fucking misery.

[metasonix]

After much experimentation, this is the best way I found to stop the random contour triggering. It works on this particular sample of S-2000 anyway. Put a film capacitor, .047 uf or thereabouts (value not critical), between the "-4.5v" pad on the board to the nearest ground, that being the ground tab on the Gate Out jack. Keep leads as short as possible. The photo is of the back of the PC board.

This fixes it. On this one sample anyway. Now I am really not sure what is causing this, it's probably supply noise maybe. We put a scope on the -4.5v line and saw only 20 mV of noise spikes, but then hanging a scope on it might be enough added capacitance to fix it. I do not think it's a bad chip anywhere--different brands for the 34063, 4013 and TL084s didn't have any effect.

Try it and if it doesn't help, ¯\_(ツ)_/¯

[buzzlegs]

Thanks for following up on this Eric as I bought new 34063 chips but unfortunately they didn't work for me, still getting erratic contours, will look into your other solution although I've no soldering experience at all, so just bought some DIY kits to practise on before I try and tackle the S-2000!;)

[papz]

Sorry it didn't work for you. Getting the right 34063's looks like a bit of a lottery.
Just guessing, maybe this depends on how the pair is "matched". You could try various combinations of an old one and a new one. If you're lucky, one may work. Not risky and worth a try.

Otherwise, the capacitor as on Eric's pic may work better than my experimentations adding a cap, that worked as long as the panel was removed. Weird.