Adding a tone control to a THAT corp compressor "One Knob Squeezer" design

[mordechai]

I've been looking at this design for a low noise alternative to an Orange Squeezer:

http://www.thatcorp.com/pedals/4320%20Battery-Powered%20One%20Knob%20Squeezer%20r00.1.pdf

I like that there's an option to include a volume control.  But I'd also like to include a tone control, and wondering if a tone control -- something similar to the active tone control on a TS-type circuit -- can be added to either the U2 or U3 spots on the schematic.  Any thoughts?

[EBK]

U2 and U3 are part of the side chain in that circuit.  Not good places for tone control.

[Groovenut]

I would think you could replace R4 with a 10k resistor, place a 4.7nF -6.8nF cap and 10k reverse audio pot in parallel with the R4 10k resistor for a treble boost control (this is the type of tone control most use with compressors, usually to restore the impression of pick attack). This will lower the overall output a bit. Not an issue if you implement the volume control.

The same idea would probably work using the existing R3 as well.

My two cents

[mordechai]

So, something like this, then?

[Groovenut]

So, something like this, then?

Yes exactly

[mordechai]

So -- just so I am 100% sure I understand...with the diagram I've concocted above, turning the 10KB pot clockwise will DECREASE the resistance, thus increasing the effect of the 4n7 capacitor and increasing the amount of treble?

[Groovenut]

So -- just so I am 100% sure I understand...with the diagram I've concocted above, turning the 10KB pot clockwise will DECREASE the resistance, thus increasing the effect of the 4n7 capacitor and increasing the amount of treble?

Correct

[EBK]

Unrelated to tone control, but if you're willing to make that design a little more complex, you might want to consider replacing C13 with a nonlinear capacitor circuit from one of the other sample designs....
More info: http://www.thatcorp.com/datashts/dn114.pdf

[mordechai]

I'm having trouble unpacking the info on that page you pointed me to. Can you help explain what benefit this will yield for the design?  Is this related at all to a feature in another one of their proposed designs (with a 4315 chip) for a compressor with variable attack and release controls?  That does appeal to me in making a squeezer-type compressor into something more flexible.

[Groovenut]

I'm having trouble unpacking the info on that page you pointed me to. Can you help explain what benefit this will yield for the design?  Is this related at all to a feature in another one of their proposed designs (with a 4315 chip) for a compressor with variable attack and release controls?  That does appeal to me in making a squeezer-type compressor into something more flexible.

The main point of the paper is using a variable cap circuit to have an "auto" type attack and release circuit. I think it would be a good addition to the one knob circuit. The variable cap circuit attaches to the Ct pin of the RMS detector.

[mordechai]

So in looking at this circuit addition, am I correct that it would cut down on potential distortion/noise at higher compression settings?

[EBK]

It's more of a distortion issue than a noise issue, but yes, it helps in this regard.  It minimizes the distortion caused by ripple at the output of the RMS detector.  You could do that with a bigger timing capacitor on the RMS detector, but it would cause the detector to be more sluggish.  This subcircuit essentially acts like a large capacitor for slowly-varying signal levels (where a sluggish RMS detector is acceptable), but acts like a smaller capacitor for transient signal envelopes (where you need a quicker detector). 

[mordechai]

Thanks for the explanation.  Between your comment and the design notes on the company's website, I think I now understand what's at work here in a general sense.  Looking at the sub circuit, it looks like only one half of the op-amp is represented on the schematic...so I assume the other half would still require the power feed and ground connection (pins 8 and 4 respectively).  How should the remaining pins be used/connected?  Just put pin 5 to ground and connect pins 6 and 7?

I guess the other question is how to plug in the values for the caps and resistors on this sub-circuit.  Since the cap on the original circuit (without the subcircuit) is 10uF, I'm guessing that should be the bigger cap on the subcircuit.  Would the other cap be much smaller, then -- like 100n?  The diodes, I assume, should be the Green LEDs that the company website suggests.  But I'm not clear on what the varying values of the other resistors on the sub-circuit would do to its performance.

[EBK]

Thanks for the explanation.  Between your comment and the design notes on the company's website, I think I now understand what's at work here in a general sense.  Looking at the sub circuit, it looks like only one half of the op-amp is represented on the schematic...so I assume the other half would still require the power feed and ground connection (pins 8 and 4 respectively).  How should the remaining pins be used/connected?  Just put pin 5 to ground and connect pins 6 and 7?

I feel it is much better to put an unused op amp to work, but if you must terminate it, connect pin 5 to some voltage in between the supply rails instead of grounding it.

I guess the other question is how to plug in the values for the caps and resistors on this sub-circuit.  Since the cap on the original circuit (without the subcircuit) is 10uF, I'm guessing that should be the bigger cap on the subcircuit.  Would the other cap be much smaller, then -- like 100n?  The diodes, I assume, should be the Green LEDs that the company website suggests.  But I'm not clear on what the varying values of the other resistors on the sub-circuit would do to its performance.

Keep in mind that 10uF was the compromise value, neither big nor small.  As far as specific component values, since "Pedal Design Five" is currently on my pedal board, I can personally vouch for those numbers.  It is set up for 24.3uF and 2.3uF being the effective big and small cap values, respectively.  I'd suggest copying straight from that design.  Note that there are a couple additional tricks in that design, including referencing that op amp to the regulated +5V, which let's you use a TL07x, if you like.

[Mark Hammer]

Here's a different approach.

U1D is set up as a buffer/preamp stage with a functionally "flat" response and gain of 2x.  Does it have to be?  No.  It should probably be in the ballpark, but there is nothing orthodox about it.

So let's adapt some tricks we've seen in a variety of simple pedals.

1) The 6k8/1uf leg sets the gain and provides a low-end rolloff around 23hz (which is why I say functionally flat).  If we were to run a parallel leg with 3k9 and 33nf, we would have this flat response, amended with a gain of 2.74x, starting just over 1200hz.  Using 3k3 and 39nf instead, we'd have the same rolloff but a gain for the mids and highs of roughly 3.1x.  Whether that addition produces the desired boost of upper mids and highs is up to you, but suffice to say it can be modified to come closer to the goal.

2) The 6k8/100pf feedback pair around U1D provide a rolloff above 200khz; helpful for chip stability but not anything you'd call "toneshaping".  We can use that feedback path for toneshaping, though.  Strapping a larger value cap across the 6k8 resistor will roll off highs.  If we stuck a 3300pf cap in there, it would begin to roll off highs around 7khz.  If we didn't want that much treble cutting, we could simply insert a resistance in series with that cap.  Once the resistance exceeds 6k8, that feedback path holds no more influence over tone than the existing 6k8.

3) So where are we going with this?  Imagine a pot, linear or whatever, whose wiper is tied to pin 17, with one outside lug in series with the suggested 3300pf feedback cap, and the other outside lug tied to the suggested secondary ground leg.  It would act as a reciprocal boost/cut circuit, whose characteristics would correspond to the desired treble-cut/treble-boost.  CAVEAT: This would alter the tone, but it would also alter what gets fed to the RMS detector, such that it would alter the quality of the compression.  That may or may not be a bad thing, depending on your needs/tastes.

4) The DOD Classic Fuzz actually used something like this as their tone control - SEE BELOW - albeit with a slightly different configuration than I suggested, and component-value choices clearly tailored to tone-shaping distortion, rather than compression.  But you get the idea.

5) When I saw "One knob squeezer" I was expecting something simpler, but that's a 28-pin chip!  Can you fit one of those in a 1590A? 

[EBK]

5) When I saw "One kob squeezer" I was expecting something simpler, but that's a 28-pin chip!  Can you fit one of those in a 1590A? 

It's 28 pins in a QSOP, so it is quite small.  Although I doubt you'll  get it into a 1590A, some people love a challenge.    My build of the 4316 version (16-pin QSOP) fits in a 1590BS.

Note, the 4320 version requires four fewer external op amps, so in theory, it can be made smaller, at least down to a 1590B (although I'm advocating adding one more op amp to the 4320 design....)

[mordechai]

So Mark...does this look right?

[Mark Hammer]

Aaaaallllllllmost.

The left-hand leg of that added 10k pot needs to add resistance on top of the 3k3, to reduce the impact of that 2nd ground leg.  You've drawn it such that the 3k3 is tied to the junction of R11 and R8, which means the pot will have no effect in the "treble" direction.  You need to lift that connection.

I well understand how attempting to graft a mod onto an existing drawing can corrupt our understanding of what we've drawn, because we had to put things where they could fit, visually, not where they made most sense to us. 

[mordechai]

Okay, I think I've got it...let me know if this is better:


Does 10KB strike you as a reasonable value for the pot?  I popped that value in since it seemed to be reasonably come close to 6.8K, and I thought a larger value might cause problems with everything else that sub-circuit network is trying to do...

[Mark Hammer]

Awkwardly drawn, but correct.

The pot value.  I suspect that if this were to operate as intended, with some reasonable dialability on either side of the mid-point of the pot, one may want to use a larger value pot, but tack on parallel resistors to each leg of the pot to each a reasonable range for each function.

So, let's say we had a 20k-25k linear pot.  We don't really "need" more than 4k or so to adjust the treble boost, and we don't really need more than about 7k to adjust the treble-cut function.  If you can find me an 11k pot with 4k on one side of the midpoint and 7k on the other, you're a better man than I.  The strategy would then be to use parallel fixed resistor values such that when the pot is rotated fully in one direction, there would be 0k on one leg (e.g., the one going to the 3k3 resistor), and something more than 6k8 placed in series with the added feedback cap.  Going in the other direction, the extra parallel fixed resistor should be able to add as much as 3k5 additional combined parallel resistance to the 3k3.

My math is probably a little off, but this is more or less what you want to do/have.  You could simply use a 10k pot, straight up.  But the question for me is what would yield optimal dialability.

Alternatively, if the goal is simply to add brightness,  just add the extra leg from pin 17 and put a 5k-10k pot in series with the 3k3 fixed resistor.  You lose the treble-cut function, but the math is easier and people complain about compressors being too dull-sounding anyway.