Expander, ssm2216 or fencepost?

[blackieNYC]

The 2216 is an expander (along with a compressor) that appears to have a variable threshold but a fixed expansion ratio of 3:1.  Or was that 1:3. I don't know what this will sound like - is it noticeably more gentle than full noise gating? Am I correct- it is internally fixed? A
And the fencepost, courtesy armdnrdy, and Mark Hammer's expander - I see the 47k fixed resistor there. Is this where one might put a pot for variable expansion? Expansion ratio is dB in to dB out, meaning that, under threshold, a 1dB drop in input level becomes a 3 dB drop in output.
I'm wondering if the 2216's ratio is just great for guitar fx (front end? Back end?)
And, was 47k chosen in the fencepost because it did the job best?  Any idea what that particular ratio turned out to be?
I don't expect to find myself getting too picky about it, but the chips cost a little.

Extra points - can the engineers thumb method get applied to an expander/gate? All audio does not pass thru the VCA because it is in a neg dB loop. Basically.

[Mark Hammer]

Well keep in mind, Alan, that it is a downward expander, and not a full-range expander across all possible signal levels.  So when one is dealing with very small signals, I'm not so sure that variations in expansion ratio would be all that noticeable.  On the other hand, I don't operate in a studio environment with good monitors, so it might be far more audible than my meagre setup would suggest.

[blackieNYC]

I got it.  Of the two mentioned, how do you prefer to use them. Work well both first and last?

[Mark Hammer]

What I appreciate most about the 2166 is that the expander accompanying the compressor.  No breathing, and as a result, no hiss to amplify by subsequent gain pedals in the rest of the chain.  And like any compressor/limiter, or gate, sticking it at the end of the signal chain imposes too much burden on that puny little expander.

[Transmogrifox]

Best signal/noise ratio when considering any form of gate or expander is by stopping noise at the source.

For example, if you go into a high gain distortion pedal or compressor, the noise level signals get amplified while the program material gets clipped or compressed.  In this situation the noise level approaches the signal level and there comes a point where you can no longer find a threshold to discriminate between signal and noise (unless you want to gate at a high threshold for the gating effect).

If using an expander to reduce hiss then always put it right in front of a high gain effect so you can gate noise coming from the lower gain effects before it gets amplified.  That way you can gate the hiss when it is still a small fraction of the signal level.  You could keep hiss pretty low if you had one noise gate/expander right before all overdrive, fuzz and distortion pedals, then another one at the input of your amp (or mixer input).

With one before and one after you are gating noise going into the high gain circuit so it doesn't get amplified, then afterward gating noise generated by the high gain effect (which hopefully is a small ratio of the amplified signal).

I tend to agree with Mark about the ratio.  It makes a real audible difference when your threshold is relatively high and you're using it more for the task of dynamic enhancement than as a hiss-reducer.  Not so much when gating a noise-level thresholds.

[blackieNYC]

I've experienced the front and back thing. Single coils and fuzzboxes - how can we expect complete silence?   I would think an expander at both ends would be fine.  I'll have to try myself. 
The gator- any idea how one could make it an adjustable expander?  A pot in place of R 16?

http://gaussmarkov.net/layouts/gator/gator-schem.png

[Transmogrifox]

Threshold control is sort of thresh and ratio in one.

If you leave this and re-label it "ratio" then you can put threshold control off the output of IC5b.  IC5b output would have to get above a certain threshold to start building charge on C8.

Probably something like a PNP transistor with collector going to C8.

Pot between a couple resistors from +9V to ground connected to BJT base (the couple resistors could be used to set the max/min limits on the pot so you don't have a bunch of unusable range on the pot).

BJT emitter connected to IC5B output through the attack pot.  Diode still bypasses the whole thing (C8 to IC5B output).

Then the existing "threshold" pot sets ratio and the new Threshold control actually sets threshold.

One way to skin the cat anyway.  You could add resistor to the existing "threshold" pot to set a more useful minimum.