Redrawn Buchla 230A Envelope Follower

[Ricko]

Here is a redrawn version of the envelope follower section of the 1974 Buchla 230A (Triple Envelope Follower), based on some schematics floating around on the net. The original followers are paired with pulse extractors which operate when EF produces > 6V.

It is a convoluted circuit, but I hope the redraw makes it a little clearer. I have substituted modern components, and I ran it through TI-SPICE to figure out its behaviour.

It is definitely designed with percussive envelopes in mind! From what I understand, it would be used to drive the lowpass gate originally. So it is *very* snappy on attack or positive-going excursions of the input compared to other envelope followers, which means that it can be a little jagged or stepped rather than smooth with some kinds of input (this may colour the sound if driving a VCA but perhaps in the original systems it would be further smoothed by any LPG vactrol, I suppose.)

It is intended for line or mic in, rather than modular synth levels. Low inputs easily send the output up to the max 13.5, which makes it act like a gate extractor.

I added a switch with a second cap to possibly help track lower frequency envelopes (or even some kind of gates) a little better and perhaps make the response a little less nervy.

---

BUCHLA230A-EF.JPG
Description:
Filesize:	101.52 KB
Viewed:	1851 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[vtl5c3]

Thanks for your post and the redrawn schematic!

I guess there were three versions of the 230. I'd be interested in knowning what the differences are.

What you said about the signal levels that the 230 prefers makes sense to me. The audio path in the 200 series was line level.

[Ricko]

Anyone interested in seeing a contrasting approach, that can highlight what the Buchla EF is about, might care to look at the ARP 2600 EF. You can find it on page 11 of the ARP 2600 Service Manual that is available on the web.

It has a slightly similar dual diode arrangement on the input stage (no zener), however it has quite a few components dedicated to smoothing out the EF signal; in fact, it has four RC filters in series, followed by an LM301A with quite a high compensation capacitor (smooths out mid-upper audio range).

The Buchla has nothing like this. So perhaps we might say that the ARP is designed for legato envelope following, while the Buchla is designed for staccato.

[Ricko]

Just for the sake of obsessive research, I checked through the simple RC LPFs on a variety of other envelope followers. They all use simple LPF, nothing fancy.

PAIA 2720-11 EF has 1 stage 14Hz LPF.
Arp 2600 has four stage 59Hz LPF.
Roland 714 has three stages 30Hz, 30Hz and 60Hz LPF.
Roland 707 is 3x 28Hz LPF.
Emu modular EF has 50Hz LPF (from brochure.)
Korg MS-20 EF has just single 15.9Hz LPF after diode.

Electro harmonix Dr Q uses sim: 159Hz LPF.
Lovetone meatball lpf min 15.9Hz (attack) LPF.

The Roland 714 is their later more advanced model. It also has 3 RC filters on the input, fixed at 77, 40, and 18Hz HPF.

If you wanted to make a universal EF module, to kitchen-sink the various features, it would have

1) Mic pre-amp stage with senstitivy control
2) Input DC block cap and HPF RC
3) Active fullwave rectifier
4) LPF RC (multiple stages) including one variable pot for decay
5) Output buffer
6) Gate/trigger extractor
7) LED on input or output or gate

(The Buchla follows this general plan, but it has the zener and its fullwave rectifier is not the usual arrangement, with the decay control parallel to one of the rectifier diodes.)

[vtl5c3]

Don't forget the Polyfusion EF:

http://www.freeinfosociety.com/electronics/schematics/audio/pictures/polyfusionenvelopefollower.gif

[Ricko]

That link didn't work for me, but I found Polyfusion 2044 Envelope Follower
http://www.freeinfosociety.com/electronics/schemview.php?id=915

It has three settings if I read things correctly, approx:

Slow: 15 Hz
Medium: 60 Hz
Fast: 215 Hz

[Ricko]

Playing with the Buchla in Spice, it really has a lot of bleed of (a sawtooth version of) the original signal especially for bass: maybe 10% of the output!

What is interesting is that adding a series of RC LPF filters before the second op-amp (like the Roland configuration) does not result in a smoothed signal the way you would expect: instead the bleedthrough signal seems to lock on to higher harmonics: in effect it is like a really high-Q low pass filter, with a sawtooth at the resonant frequency.

So for an EF that does not color, this is not the design! For an EF that takes unpitched or semi-pitched precussive sounds, this will be very snappy.