Hello

I was wondering if someone could answer a question?

Would lowering or raising the values in the voltage divider on the 2nd fet that connects to the envelope sensor in the squeezer make the sensor more sensative?

The two 470k resistor that meet at the gate.

Thanks in advance
Jeff

My guess is no.

The 2n2 cap and 470k resistor are in exactly the same spot they are in many varieties of FET-based phasers to increase the headroom of the FET and reduce distortion.  I take it they are used similarly here, especially given the almost universal comments about the sonic transparency of the OS.

If you want greater compression (which I assume is the reason for your query) you can do a few things.  One is to increase the gain of the op-amp.  You can do that either by increasing the value of the 220k feedback resistor, decreasing the value of the 10k resistor to ground, or both.  JD Sleep posted the mods I did to mine over at www.generalguitargadgets.com (you'll need to scroll down when you get to the OS page).

A second mod to increase the apparent compression might be to increase the value of the 100k resistor in parallel with the 4.7uf cap.  That 100k resistor sets the decay or discharge time for the cap by providing a path to ground.  At present, it is set for a relatively fast decay time, which means that the unit recovers from transients fairly quickly.  Lengthening the decay time by means of the discharge resistor or by increasing the size/value of the cap (e.g., to 10uf or even 22uf), will lengthen the recovery period.  This makes the unit less responsive to quickly-occurring notes, because it takes longer to be ready for them.  If a transient reduces the volume level, it will stay there for a little bit, even if you pick really hard on the next one.  One of the many reasons why some manufacturers have erroneously labelled things like this an "attack" control, because you tend to notice it in the attack rather than in terms of what it actually does electronically.

The other 470k resistor going to the FET gate makes sure there isn't too much current going there, which means that goosing the gain of the op-amp shouldn't be problematic.

The thing to remember about that whole series of Armstrong pedals/boxes is that even though there is a ton of parameters you could tweak if you know how they work, they were intended to be non-adjustable, on-off units.  At the time, we fools thought that was "convenient".

So what you're saying is by increasing the output of the opamp (the peak to peak voltage) the follower will compress more?

I've already replaced the 100k resistor with a pot as well as the 1.5k resistor with a pot.

Is the opamp simply a booster or is it doing more?

Thanks again
Jeff

The op-amp is acting as both booster, recovery stage, and envelope follower.  I'll try and lay out the chain of contingencies that produces this state of affairs.  You will note a certain resemblance between my explanation and "The house that Jack built", "There's a hole in the bottom of the sea", or if you're Jewish "Chad Gadya".

Everything about the OS is intended to do one basic thing: reduced the D-S resistance of the FET connected to the 82k resistor.  The FET and trimpot to ground, in conjunction with that 82k resistor, act like a pot   The 82k resistor is one leg, the FET+trimpot is the other leg, and the FET/82k junction is the wiper.  With any voltage divider/pot, changing the value of one leg of the pot changes the resulting voltage available at the wiper.  In this case, the leg that is changed is the leg to ground, formed by the FET+trimpot.  In the true spirit of compressors, the OS is simply a glorified volume control. and all the other parts are just there to help the volume control work.

To produce the desired change in resistance of the FET, a voltage within a specific range needs to be applied to its gate.  The FET can't help it; it's made that way.

In order for that change to correspond to signal level, you need to use only one half-cycle of the signal so that the FET receives something that ranges from ground to positive.  The diode assures that only positive-going (or is it negative?  I can never tell) signal passes on to the FET gate.

Because the half-cycle is derived by chopping off half the signal by means of a diode, right away the signal amplitude is cut in half.  As well, the voltage drop of the diode, while small because it is germanium, is still a voltage drop, reducing the voltage from the op-amp that gets applied to the FET gate.

Sooooooo.........

The op-amp boosts the signal up to a level where presumably after passing through the diode and being chopped in half, and having a fixed amount of voltage subtracted, it is still in the right zone for providing a suitable voltage to the gate.

This logic applies to just about every envelope-controlled device out there.  In some instances the amount of gain that needs to be applied in the booster stage just ahead of the diode is quite large.  Many folks here have found a need to stick feedback resistors as large as 3.3meg in the corresponding gain stage in the envelope follower of the Dr.Q/Quack and Nurse Quacky.  To put it in perspective, that provides an amplification factor of around 70 for the DQ, in contrast to 23 in the OS.

The OS uses this booster stage part of the envelope follower as its output stage, sticking a plain vanilla volume control at its output.  As an economical solution, this is brilliant.  One of the weaknesses of it, though, is that its default boost IS x23, which means that any noise coming its way is boosted 23 times.  What would work far better, for noise minimization purposes, would be something where the boosting work is divided up, such that, say, the existing stage boosted by maybe x5, and after the signal is split and sent off to the output control and the rectifier/diode circuit, a second stage would boost the signal to be rectified by something like x4 to x10 to bring the level up hot enough and also provide degrees of compression.  That way, the noise might be boosted a whole lot for whatever gets to the envelope follower, but only boosted x5 for the path that you actually get to hear.  With compressors, the secret to noise minimization is to keep the boost out of the audio path as much as is feasible.

If you look at the mods I did over at GGG, you'll see what I did to permit a greater range of boosting, as well as tame the hiss a bit.  I chickened out and used caps for hiss-taming, rather than introduce another boost stage.  Since I only planned on using it for guitar, losing all that stuff over 9-10khz wasn't a great loss.  Variations in the gain of the basic booster stage will yield different degrees of squash, though, which is what you're after.