[sdiy] Random bits on the Buchla LPG

[Grant Richter]

OK, I do know some thing about this.

The LPG was call a "Frequency and Amplitude Domain Modifier". It has  
three modes (4 on the easel) "Frequency", "Both" , "Amplitude"

>
> I've been going over the different versions of the LPG - the 292B  
> and 292C, and the Music Easel, and pondering the various differences.
>
> I see the basic Sallen-Key structure, and the influence of  
> switching in the VCA mode in addition to the VCF mode. I'm  
> wondering about the purpose of random odds and ends, capacitors and  
> resistors here and there, that fall outside that structure.
>
> Let's start with the 292C.
>
> http://rubidium.dyndns.org/~magnus/synths/companies/buchla/ 
> Buchla_2920C_200.jpg
>
> For instance, the feedback cap is a - looks like 4.7 nF cap.  
> (Interesting, that's what he uses in his oscillator cores. Must  
> have had a bunch handy?)

When this is switched in, in "Frequency" mode, the Sallen-Key  
topology "swamps" the Butterworth topology.

>
> Anyway, between the feedback cap and where it ties between the  
> vactrol resistors, there's this 220... uh, I can't really read the  
> units! cap to ground, and I can't figure out its purpose.

It is a 220 pF ceramic disk. When the 4.7 nF is switched out in the  
"Both" mode, the circuit forms a classic passive Butterworth 12 dB /  
Oct two pole. The reason it is smaller is because of the passive  
loading of the 910 pF cap. If the sections were buffered, the 220 cap  
value would be 910 pF.

>
> Then, the second Sallen-Key cap is 910 pF. If the VCA mode is  
> switched in, a 10K resistor to ground is dropped in in parallel to  
> it. But, regardless of what mode, there's also a massive 4.7 meg  
> resistor there. I can't figure out it's purpose either.

In "Amplitude" mode, when the resistor is switched in parallel with  
the 910 pF cap and the 4.7 nF is disconnected, it becomes a simple  
resistive voltage divider. That is, the voltage divider mode now  
"swamps" the Butterworth mode.

>
> In the 292B version, the 4.7 meg resistor is 2.2 meg instead,

The Vactrol will go to 100 Megohms when off. There needs to be  
something there to bias the FET buffer, that is purpose of the 2.2 or  
4.7 meg to ground, to maintain input bias to the FET when the  
Vactrols are completely off.


> the resistor that gets dropped in is 15K instead of 15K,

In "Frequency" mode, the 4.7 nF cap is switch in regeneratively. This  
provides more gain in that mode than the other two.
To keep the output gain constant between the three modes, he used a  
couple different techniques. Either cut the gain for one mode, or  
boost the gain for two modes.