Orange Squeeze - getting rid of the low end pumping

[Stompin Tom]

So, many many moons ago I built an OS. It turns out that I really love what this thing does, and it keeps finding its way back onto my board. In fact, I like it so much that I want to just leave it on all the time. My problem is that I find the pumping on the low end (or, to put it another way, the initial low end cut on my low notes) is too much. I've used it as is for years, just turning it off when I go down low. But, I've been thinking why not put a high pass filter on the part of the circuit that triggers the compression? A simple idea, but what do I know? (not much!). Looking at the GGG schematic (available here: http://www.generalguitargadgets.com/index.php?option=com_content&task=view&id=208&Itemid=26), could I just put a simple high pass filter like Jack's (available here: http://www.muzique.com/schem/filter.htm) in front of R11 (the 1.5k resistor before the diode)? I'd have to figure out what frequency I'd want the cut off at... I was thinking of using a pot with fixed cap to figure that out. Would this work, or am I off my rocker?

[Stompin Tom]

I'll add this useful thread I found in my many searches... includes RG's technology of the OS:

http://www.diystompboxes.com/smfforum/index.php?topic=54308.0

[Mark Hammer]

If I understood correctly, the 4u7 cap and 1k5 resistor form a sort of highpass rolling off below around 22hz.  Unfortunately, the 4u7 cap performs double duty, providing the DC blocking and rolloff for the envelope detector, as well as DC blocking for the audio output.

However, there is no reason why the op-amp can't drive two paths.  So, if you stuck a .47-1uf electrolytic cap, with the same orientation between the 1k5 resistor and the output of the op-amp (breaking the connection between the 4u7 and 1k5, while leaving the 4u7 connected to the rest of the output path), you would likely roll off some bass from the envelope detector's input.

However, what you probably want to do instead is simply reduce the amount of compression overall.  You can do that by simply reducing the gain of the opamp a bit.  Sticking a 5k pot between R10 and C5 would allow you to reduce the gain andreduce the amount of compression a bit.

[aziltz]

Honestly, I think cutting the bass going into the envelope detector works the best at reducing how much it compresses when you play low notes. HOWEVER, this only works if you are using both sides of the op-amp and the envelope detector is out of the signal path, otherwise you are cutting bass from the output as well.

Check out Mark's version called the Tangerine Peeler to see what I mean.  The envelope detector is separated from the audio path.  You can still get the same sound but it offers better control.

[Stompin Tom]

Thanks for the replies. Obviously Mark's suggestion for adjusting the compression would be the easiest, and I will definitely try it.

As for the envelope trigger being in the signal path... I'm a little slow and don't really get this... Maybe I don't understand why it wouldn't work in the original. Does the audio signal actually go through the envelope trigger (diode, back to the fets, etc.)? Thus, if the bass was not in that signal, it wouldn't be amplified like the highs... and would then be too quiet as it's added back to the dry signal? Sheesh. I thought I kind of understood this circuit. I don't!

Here's Mark's very cool looking Tangerine Peeler:

http://hammer.ampage.org/files/Tangerine_Peeler.gif

Ignoring the input buffer and treble boost on the output, it looks to me like he's added the second half of the opamp before the diode that, as far as I understand, determines how much signal is boosted/compressed? The envelope trigger can now be adjusted using the compression pot... So, I guess I don't really see how adding a high pass to that would be different then in the original circuit.

My minimal pedal knowledge is failing me! 

[aziltz]

no worries.

when I said "envelope detector"  I was referring to both the Op Amp and the resister/diode/FET feedback network.  In the OS, the Op Amp provides a boost for output, but also drives the envelope detection.

In Mark's Tangerine Design, the gain stage driving the diode/FET is not supplying the output signal, like the original OS does.  This is IC1c in the Tangerine.  The Audio Path runs from IC1a, to IC1b, and out through IC1d.  In this arrangement you can control the compression without affecting the frequency response of the audio signal.


I've found that cutting bass from the input of the pre-envelope detection gain stage (IC1c) removes some of the low end pumping while still retaining the compression on the highs.  I hope that helps.

I really like Mark's Design.  I've made my own and I'm still tweaking it, but my goal was to make an OS that uses both sides of a dual opamp and can get the same sounds as the original but have a lot more control.

[Stompin Tom]

Thanks for the explanation. It's still not super clear to me why cutting the bass that goes to R11 (and leaving the signal that goes to the volume pot intact) would cut bass in the overall signal... But I'll give it some more thought. Sometimes this stuff takes some time to sink in.

I went ahead and experimented a bit with Mark's compression mod suggested above (add the pot between r10 and c5). I found that a 10k pot works better then a 5k with my guitar. So far my ears are telling me that I like the added resistance of about 7-8k there. I'll be doing some demoing tonight with my real rig and I'll see how it goes. I'm thinking this will be a reasonable work around, although I think cutting the bass to the 'envelope detector' will give me the most satisfying results (I can still hear some pumping!). If you'd like to share your results azlitz (or your orange squeeze/tangerine peeler using both sides of the opamp), I'd be very interested. I'm crap with layouts, but will eventually give the peeler a try after I get done with a few 'easy' builds I'm in the middle of...

[Mark Hammer]

Think of it this way; there is what the envelope detector is responding to, and there is how much it responds.  The suggested gain adjustment of the op-amp varies how much the JFET responds by giving it larger or smaller signals to respond to.  Varying the low-end rolloff of the signal fed to the diode, however, alters what the JFET is responding by selectively attenuating the bass content via a smaller-value cap.  In a sense, it is like turning down the signal fed to the diode, but only doing so for the deeper bass, and leaving the rest of the signal at whatever its current amplitude it. 

Technically, you could do that via the ground leg of the op-amp where we inserted the gain adjustment, but then it would have the effect of not only altering the bass level of the control signal, but altering the bass level of the output signal as well.  The stock unit uses one cap to set the low-end rolloff of both audio and control signal.  The added cap I suggested in the earlier post simply imposes a different frequency response on the control signal than on the audio output.

Cutting the bass in only the control signal would NOT cut the bass in the "overall signal".  If anything, it would retain a little more bass because the JFET would not respond as vigorously to the bass, resulting in less attenuation via the JFET.

[aziltz]

Here's an example.



I took a hint from Mark and separated the Envelope Detector and Op-Amp from the Audio Signal.  The input signal comes in, passes the FET/Voltage Divider Attenuator.  Then its buffered, and split.  IC1B is and everything connected to it is basically the stock orange squeezer with a compression control in the 100k Pot in the Feedback loop.  Notice how this opamp isn't connected to the output.

IC1A is the same gain stage duplicated in parallel.  This one drives the output, so you have the same gain characteristics of the original orange squeezer, but with the envelope detection out of the audio path.  Remember the compression happens at Q1 and Q2.


Lowering C9, you can cut bass to the Envelope Detector, but still let the full frequency response through to the output via C15.  I found this is a nice place to adjust just how much the low notes thump the compressor, while retaining the nice snap that the high strings give. 

Let me know if you have any questions.

[Stompin Tom]

I think I actually understood things better then I thought. All that makes sense. But then I just tried Mark's suggestion of splitting the detector path off of the one pin on the opamp with a smaller electro cap. Strangely, this didn't seem to do anything (no matter what value I stuck in there). No compression, just sort of a dirty boost. hmmm.. Tried small and really large caps. The only value that did anything was a 470uf, which just cut the signal down to almost nothing. Maybe I'm doing something wrong, or perhaps this is why aziltz suggested using the other half of the opamp...

The good news is that the added compression pot worked pretty well last night while recording. I still don't want to leave the thing on all the time, but it can get a bit more transparent on the lows. Still, though, I miss the full compression on the highs. It'll have to do until I get around to putting together something like the peeler... thanks again.