[sdiy] Aaron's take on the Buchla 258 (rough draft)

[Magnus Danielson]

Aaron Lanterman wrote:
> Before trying to tackle the oscillators in the Music Easel, I thought it 
> would be wise to tackle something simpler to get some experience. I 
> thought the 258 would be a good thing to try, since many folks have 
> tackled it successfully:
> 
> http://users.ece.gatech.edu/~lanterma/b208_draft_schem.png
> 
> This is based primarily on:
> 
> http://rubidium.dyndns.org/~magnus/synths/companies/buchla/Buchla_2580_1_200.jpg 
> 
> 
> With some understanding gained by the notations of one of Mark V's 
> versions:
> 
> http://www.simple-answer.com/258mod.jpg
> 
> I made a few tweaks:
> 
> 1) The original 258 has some resistors that you change or leave in or 
> out depending on if you want a square wave or sawtooth wave. After 
> studying the schematic a bit, I figured you could switch between the 
> modes using a three-pole, double-throw switch (which are available), so 
> I went ahead and set it up to work that way.

If nothing else, a rotary switch can do that.

> 2) The buffer at the integration cap in the original is a FET/BJT combo. 
> I replaced that with a TL082 in a straight noninverting follower 
> configuration.

I recommended this trick to Jörgen when he made his 258 clone for the 
Bergfotron. Should not change the loop gain too much, but the noise may 
be different.

> 3) Cynthia's Buchla version of her Zeroscillator has separate "Don" and 
> "Bob" CV inputs, which I though was an insanely cool idea. My 
> implementation of it works by putting the trimpots on the input instead 
> of in the feedback loop of the summing amp. (I'm not sure if that's wise 
> or not, since it means that you're also sort of trimming the input 
> impedance.)

Not so great. Now your trimming depend on the output impedance of the 
source you have. Also, any paralleling into other oscillators detunes 
it. Baaaad idea. You want to be buffered before you hits the trim-pots.

> 4) I bring out the raw triangle output - seemed easy to do, just an 
> output protection resistor and another jack.
> 
> 5) I buffer the weird spiky waveform at emitter of the comparator 
> transistor pair, under the theory that it might sound interesting on its 
> own, and bring that out. Mark marks this as swinging from -2.7 to 1.3. I 
> buffer this with an op amp and also stuck in some other resistors to 
> level shift it. (Note the resistor values there around that op amp are 
> just place holders, to be changed once I figure out what values I need.)

Sync anyone?

> 6) The original 258 uses the unobtainium mu726. I adapted the expo 
> converter from the Bergfotron VCO, which I think itself is an 
> Electronotes/ASM/etc. NPN-pair-based variant. To this I added a 
> pot-changeable exponential CV input in addition to the linear CV and 
> trimmed inputs.

It is entertaining to look through old Buchla, Moog and ARP schematics 
and view the various compensation-tricks they did before they had low 
bulk resistance transistor pairs available.

> 7) There's a weird resistor on the lower right of the Buchla schematic 
> that says "230K to 470K" as needed, so I made this a resistor with a 
> trim pot to avoid having to actually swap resistors in and out.
> 
> 8) Boosted the output resistance of the main mix output to 1K.
> 
> 9) Specified J201 JFETs since I happen to have a bunch on hand. No idea 
> how well they will work, I guess that's what testing is for.

I though that deep thought and consideration could motivate a suitable 
replacement and possible adjustments in the design to handle it.

> Two questions:
> 
> A) Any other cool but easy hacks come to mind that I should put in while 
> I'm at it?

I don't recall the details and is to lazy to check, but a sync input 
would maybe be nice.

> B) Anything I'm doing seem blatantly stupid? Any obvious errors jump out 
> at anyone? ;)

Yezzzz. Your trim-pots on the input thing is BAAAAAAD. Are you trying to 
get into the mood of building an EMS Synthi-A or what? I just need to 
slap it in your face a bit it seems.

Cheers,
Magnus