Orange Squeezer Output Capacitor Question

[vanessa]

I still have been trying to get to the bottom of a friends issue with OS noise. I read in a post that changing the output cap to .5uf or even 1uf can help a lot. (I've tried opamps). So I was looking at the schematic (yes I'm a newbie) and see two that look like they could be the output cap. Using JD Sleep's schematic (http://www.generalguitargadgets.com/diagrams/orangesqueezer_sc.gif ) I'm new at this but C6 looks like it's blocking not outputting. Am I right on that? Is C7 the output cap?


:roll:

[Mark Hammer]

Close but no cigar.  C7 is an averaging or "time constant" cap that smooths out the rectified envelope.  The actual output cap would be one that blocks DC and sets the low-end frequency response of the unit.  That particular beast would be C6.

I have not done such a change myself, but there are a few things that might give pause for thought, in principle:

1) The drop in value is not only a quantitative one, but a qualitative one as well.   Move from >1uf to <1uf and chances are you're making the move from electrolytic to non-polarized plastic, which might affect noise.

2) The envelope signal is derived *after* C6, hence is limited by it.  Trimming back low-end in the to-be-rectified envelope signal will likely reduce the amplitude of the envelope signal driving the FET, conceivably reducing the maximum amount of squish possible.  This would be over and above any possible bass-reducing effect it might have on the output signal.  On the other hand, my read is that low end rolloff occurs at just over 3hz with a 4.7uf cap, so dropping it to 20% of original value (i.e., 4.7uf -> 1uf) still wouldn't take away much bottom, either from the output or from the envelope signal...assuming my calculations are correct.

3) What people call "noise" or "distortion" in compressors can sometimes be misleading.  Certainly, it is an undesirable component of the audio output, but because it is a vague term the cure may not present itself quite so easily.  In just about any device using a half-wave recitified sidechain (D1/C7/R12 in this instance), there is a risk of "ripple" coming from the imperfect smoothing of the cap. Since the envelope signal is NOT perfectly smooth, it will produce subtle audio-frequency (i.e., at an "LFO rate" >20hz) changes to whatever the effect is doing, be it, reducing gain (compressors), sweeping a filter (ECF/autowah pedals) or gating.  Those changes can sound like distortion, or more likely a kind of burr or sputter during note decay.
One of the sources of this can be old strings that introduce ultra-low frequency "beats", or dissonant notes, or some other source of something at a low enough frequency that you wouldn't hear it on its own, but you'd hear what it does to other things.  This would be more apparent with single strings than chords, since the beats from 6 strings sort of blend in together and can even cancel each other to some limited degree in the odd instance.
I mention all of this because it may well be that the "noise" that sparked the original comment was not hiss, but rather ripple-related, and decreasing the value of C6 reduced the amplitude of the low-frequency ripple in the envelope signal.  
Incidentally, if cranking C7 up to 22uf or even 33uf gets rid of the "noise", then clearly envelope smoothness was the source of the problem to begin with.  If so, should you change C7 instead of C6?  No, because a larger value of C7 will make the unit respond more sluggishly to transients.  It is the rapidity of the OS's response that gives it its reputation as "transparent".  On the other hand, a slight reduction in C6 (say, 2.2uf)and a slight increase in R12 (say 150k) could conceivably improve ripple rejection while leaving bass and attack time intact.

[mojotron]

Mark, ever thought of doing a "Technology of the Squeeze" write-up?

Which, at first, is bound to raise the eyebrows  of the un-informed :wink: , but would be greatly appreciated by a number of people I'm sure. I would help in any way I could to get your wisdom on the matter in one place.

[vanessa]

That did the trick! Sounds great!

Thank you Mark!!!

[gaussmarkov]

Mark, ever thought of doing a "Technology of the Squeeze" write-up?

Which, at first, is bound to raise the eyebrows  of the un-informed :wink: , but would be greatly appreciated by a number of people I'm sure. I would help in any way I could to get your wisdom on the matter in one place.

i second this motion!    i would learn a ton.

i've thought that it would be cool to gather all of mark's posts together and edit them into a book.   :shock: but i don't have the know-how for that kind of work.  narrowing things down to compressors would help a lot. ... then phasers.  .... then ...  

[Mark Hammer]

That did the trick! Sounds great!

Thank you Mark!!!

You're most welcome.  Just so I know, what exactly did "that" consist of? :lol: Seems to me there were a couple of possibilities mentioned.

And thanks to Mojo and Gauss for the flattery.  First let me get my dining room table and workshop cleaned up.  Then let me finally finish my Hyperflange from 13 years ago.  Then let me fix my MXR Digital Delay.  THEN we'll talk.

[vanessa]

You're most welcome.  Just so I know, what exactly did "that" consist of? :lol: Seems to me there were a couple of possibilities mentioned.

I up'd R12 to 160k (that's all I had in my resistor box). It got rid of a lot of the "dirt" aka ripple. What a cool mod. I think I'm going to do it to mine too.