Q & D compressor SSM2166 chip

[theman]

Thanks for the info and glad to hear that it worked out for you. I plan on building it soon ...

After performing the "surgery" on my Q&D last night, and playing through it for a couple of hours I must say that it's effortless to get a good tone out of this....infact, you have to work at it to make it sound BAD!
That's kind of a nice change of pace ya know?
Anyway, build it! It's worth it.
I plan on messing around with some of the comp values and rotation a bit but even as it sits right now on the bench, it's fine.
As a matter of fact, any other "mods" I do will probably take place later on as I am quite content with it as is now.

Hi Markm,

Thanks for the review -- yes, I plan on building it very soon. Thanks for your review and feedback! I have a few ideas regarding the mods that Mark Hammer spoke of. They relate to the noise pot and "Rotate" pot. I like to call the latter a "Limit" pot since it sets the point where limiting starts. In any case, if you look at the plots of resistance to +V for pin 9 (Noise Gate) in the datasheet (plot TPC2), you'll notice that the noise threshold changes drastically from 10mV to 5 mV in only a span of 50K. It looks more log-like in the decrease. If you use a regular linear pot (say 250K), and you want to try and hit something like 8 mV, it might be hard. It'll be near impossible with a 1M lin pot -- the rotation range is just too small.  As you go out farther past 250K, the noise threshold drops below 1mV and doesn't decrease as fast as you increase the resistance. My idea is like Hammer's (trimpot+pot) except instead of a trimpot, use a 220K fixed resistor in parallel with a SPST switch and a 1K resistor. And the Noise pot I'll use is a 250K C (reverse audio).  You'll see why in a minute.

I want the noise gate knob to work such that as I turn it CW, it sets the noise threshold HIGHER, not lower.  This is the way most noise gates work.  I also need it to be gradual as it approaches the CW position since this is where the noise threshold is most sensitive to resistance changes. Because of this, you need a reverse audio pot, and take the resistance between the center lug and CW lug.

Now put this pot is in series with a parallel combination of a (1K resistor + SPST switch) and a 220K resistor. When the switch is open, I call this the LO range position. The resistance is 220K. With the pot at fully CCW, the resistance is 250K. Add the 220K and you get 470K. This corresponds to a noise gate threshold of 0.6mV. This sets the minimum. As you turn the pot CW, it decreases to zero, so the total resistance is 220K which is a noise gate threshold of about 1.1mV.  So the LO range is 0.6 mV to 1.1 mV.

When the switch is closed, this is the HI range. The resistance is now 1K in parallel with 220K which is very close to 1K. With the pot fully CCW, the resistance is 250K, so total resistance is 251K. Noise gate threshold is then about 1 mV.  As you turn the pot CW, it goes to zero. The total resistance is then about 0.995K, and the noise gate threshold is > 10mV.

Thus, you can cover a good range while at the same time use the reverse audio pot to fine tune the selected range.

I chose the 250K and 220K since for my purposes, as I've seen my noise gates act on high distortion devices in the -40 to -30 dBV range. -30 dBV is 20 mV (can somebody confirm?), so most of the time, I'll probably be operating in this range. But other times when playing clean, I could be down in the -60 dBV range which is 1 mV. Choosing the values above seem to be a good compromise in this area. Of course maybe you only operate in one range or the other, in which case you can get rid of the switch alltogether and choose the pot and fixed resistor sizes to suit your needs.

Regarding the Limit pot ... you'll see that the plot of Limit point vs resistance also looks like the noise gate plot in shape. And it's intuitive (at least to me) that as I turn it CW, it raises the limit point. This is not the effect in Orman's schematic. I would use a 50 K C (reverse audio) again, with the resistance between the middle lug and the CW lug.

Let me know what you think. I'll be building it sometime soon!

[Mark Hammer]

Elegant solution.  Smart.

I suggested the trimpot+pot arrangement not so much because it was a "best" solution, but rather because the one shown in the schematic was NOT a "best" solution for guitarists.  The downward expansion facility is for controlling noise, and setting the threshold so that one loses as much noise but keeps as much signal as possible is tricky.  When you consider the difference between playing a single-coil guitar in a room with fluorescent fixtures and a computer (like I use) vs playing a humbucker-only instrument in a space devoid of many sources of hum, it seems fairly clear that no single control arrangement will afford a guitarist the degree of dialability in the threshold control that they would like or need.  Their threshold will certainly be in there, just hard to nail reliably.

Arrangements such as the one described/analysed in pleasing detail by "theman", or the one I suggested, may help the user to tailor the very pleasing noise-reduction capabilities of the unit/chip to their context.

As always, keep in mind that this chip (SSM2166) was not designed for guitarists but for microphone processing, where noise floors, noise sources, gain factors and such can be very different, not to mention operating contexts.

[markm]

I have to agree with Mark Hammer on this (I don't have much choice as this is really over my head!  )
but it does seem to make some sense. I have to say that dialing in the rotate pot hasn't been much of an issue for me however, that said, between the rotate adjusment and the gate, you do have to be caeful because it will cut off the end of a sustained note or chord if set to more extreme settings.
The interesting thing is the Rotate pot, to a less knowledgeable DIY guy like myself, acts like a second Gain control......why is this?

[theman]

Elegant solution.  Smart.

I suggested the trimpot+pot arrangement not so much because it was a "best" solution, but rather because the one shown in the schematic was NOT a "best" solution for guitarists.  The downward expansion facility is for controlling noise, and setting the threshold so that one loses as much noise but keeps as much signal as possible is tricky.  When you consider the difference between playing a single-coil guitar in a room with fluorescent fixtures and a computer (like I use) vs playing a humbucker-only instrument in a space devoid of many sources of hum, it seems fairly clear that no single control arrangement will afford a guitarist the degree of dialability in the threshold control that they would like or need.  Their threshold will certainly be in there, just hard to nail reliably.

Arrangements such as the one described/analysed in pleasing detail by "theman", or the one I suggested, may help the user to tailor the very pleasing noise-reduction capabilities of the unit/chip to their context.

As always, keep in mind that this chip (SSM2166) was not designed for guitarists but for microphone processing, where noise floors, noise sources, gain factors and such can be very different, not to mention operating contexts.

Hi Mark,

Thanks for the very kind words. The idea came from your idea of getting more control of the gating, and I'm sure there are more possibilities as you pointed out depending on the context of the device's use.

[markm]

Do you suppose there is a better "generic" setup for the gate than what is illustrated in the original schem?
If so, what would it be??

[theman]

Do you suppose there is a better "generic" setup for the gate than what is illustrated in the original schem?
If so, what would it be??

It's hard to pinpoint what is "generic" because it depends on the noise range you normally deal with. But if I had to make it simple, I'd use the same setup shown in the schematic except choose a reverse audio taper pot. Maybe start out with 250K C pot, except in the schematic, have the CW lug connected to +V and the other two connected together to the 1K. In this configuration, as you turn CW, the noise threshold set is HIGHER. If you need to cover lower noise gate ranges, increase the pot to 500K instead, but you'll lose some fine tuning ability in the higher threshold region.

[markm]

Okay.
I think the one in the schematic does work in reverse rotation.
CCW is maxxed.

[theman]

Well, I finally got around to making this. I put a MOSFET-based buffer in front.

First off, the GATE works beautifully. The HI and LO switch work very well over a wide range. You don't have control over the decay, but it isn't a big issue since it clamps down pretty well.

I do find that rotating the COMP pot CW gets higher compression ratios as well as higher perceived volume. I'll then use the LIMIT (rotate) pot to set the extreme LIMIT range.

The GAIN knob isn't needed since there's more than enough gain to go around. But it's cool as a boost.

So it's a compressor, limiter, noise gate, and boost -- all in an MXR-sized box!

Tonewise, it can squash like a Boss, but retain the highs.  You can also have it set more transparent. In this mode, there's not a heavy squash to the signal. The highs are still there, and the overall tone sounds like your guitar in bypass, only softer notes standout.

Overall, though it has some characteristics which make it different than most guitar compressors, this thing is a winner!

[Stephen]

Yeah this thing is quite amazing in a single chip.....

I have one more chip and havent decided how to set that one up yet!

Smallbear sold them to me ...they are amazing

[markm]

This is what you can do to calm down the gain in this baby and make that gain control functional;

If that on-board op-amp is a *real* op-amp, then replace the 10k resistor (R2) and 1uf cap to ground (C2) with....nothing.  You could replace the 10k feedback resistor (R1) with a wire link but the key thing is to make the resistance to ground as large as possible so as to turn it into a unity-gain stage.

Although the gain control CAN be used for level setting, what seems to have escaped most of us is the option to preset the gain via trimpot and tack a 100k log output level pot after C10.  Setting the gain level will be a pain in the arse, but do-able.  The trouble is that, unlike a number of more familiar compressor designs, on this one output level goes up as compression amount is increased.  So, you'd need to set the compression to max, the added output pot to max, and the gain level to what you think you can tolerate.  That will be the max output level attainable.  A/B it with bypass and adjust to taste.  The added output level pot can now attenuate below that max level.

There is still more than enough gain left to use the unit as a booster, just not a ridiculous amount as it is in stock form.
This will make the gain control useful to you, not just a knob to take up space on the control panel! 

[theman]

This is what you can do to calm down the gain in this baby and make that gain control functional;

If that on-board op-amp is a *real* op-amp, then replace the 10k resistor (R2) and 1uf cap to ground (C2) with....nothing.  You could replace the 10k feedback resistor (R1) with a wire link but the key thing is to make the resistance to ground as large as possible so as to turn it into a unity-gain stage.

Although the gain control CAN be used for level setting, what seems to have escaped most of us is the option to preset the gain via trimpot and tack a 100k log output level pot after C10.  Setting the gain level will be a pain in the arse, but do-able.  The trouble is that, unlike a number of more familiar compressor designs, on this one output level goes up as compression amount is increased.  So, you'd need to set the compression to max, the added output pot to max, and the gain level to what you think you can tolerate.  That will be the max output level attainable.  A/B it with bypass and adjust to taste.  The added output level pot can now attenuate below that max level.

There is still more than enough gain left to use the unit as a booster, just not a ridiculous amount as it is in stock form.
This will make the gain control useful to you, not just a knob to take up space on the control panel! 

I connected pin 5 and 6 together to set the VCA buffer gain to unity. However, when I increase the comp ratio, I get a large perceived increase in volume that can only be tamed by introducing limiting. This is not a big deal to me, as I can set the master volume of my pedalboard using an EQ which is post compression, or in many other ways.

This increase in volume is standard behaviour, however, as most compressors sound louder as you increase their "sustain." It's just that most comps also have a level knob to compensate, while this design doesn't. You could certainly add one, however. I currently like it the way it is as a 4-knobber ...

[markm]

I don't think you want to connect 5 and 6 together.
They need to be grounded.
This would eliminate the opamp in the chip.

[theman]

I don't think you want to connect 5 and 6 together.
They need to be grounded.
This would eliminate the opamp in the chip.

Hi Markm,

I'm a bit confused since the updated schematic by Arn C showed 5 and 6 connected together, and you responded with a "Yes Sir." This makes the op-amp a unity gain stage.  From the block diagram of the SSM2166, I can't see how pin 5 should be grounded if it is connected to the 10 uF cap to pin 3. The audio signal must pass through pin 3 (VCA input) to get any audio out!

From the datasheet:

"To configure the SSM2166's input buffer for gain, provisions
for R1, R2, and C2 have been included. To configure the input
buffer for unity-gain operation, R1 and R2 are removed, and a
direct connection is made between the –IN pin (Pin 6) and the
BUFOUT pin (Pin 5) of the SSM2166."

[markm]

Here's what I've done so far;

replace the 10k resistor (R2) and 1uf cap to ground (C2) with....nothing.  You could replace the 10k feedback resistor (R1) with a wire link but the key thing is to make the resistance to ground as large as possible so as to turn it into a unity-gain stage.

I did this modification to set the op amp section to unity gain.
This completely took care of the over abundance of Gain/Boost/Output/Volume that it had originally and made the comp very useful on it's own. I do plan on trying some of the other options available that Mark H. mentioned in this thread as well.
You are correct abou the info in the data sheet, there are alot of possible "adjustments" to be made to this circuit but for now, I'm just so tickled with getting rid of the Overkill Volume that this circuit had that I've not gone any farther......I've been just playing my Tele through it most of the night playing some killer Chicken Pickin' country licks and they sound absolutley fabulous through it!

Unless you have a different layout, doing exactly what Mark H describes here will do it.
Are you using the Jack Orman layout, Perfboard, or your own?

[theman]

I perfed my own layout, but it's similar to Orman's latout. The suggestions Mark made essentially make the input buffer a unity gain stage, which is connecting pins 5 and 6 together. This is also what the datasheet suggests.

In any case, the volume should get louder as you increase the comp ratio.  This is what I get. I'm just trying to figure out a way to incorporate a level out (which Mark also suggested) without adding another pot. I am thinking of reducing the GAIN resistor which sets the minimum resistance (in the schematic, it's 1K) to below this. This will make the gain less than unity (less than 0 dB). Thus, making the gain pot both attenuation and amplification. If I can set it such that unity gain of the pot is set at noon, then you can use it to back off the volume if needed, or boost. This would be ideal for me.

[markm]

Ahhhh,
I see what you are saying.
I'm still looking at tweaking the limiter a tad.
Hey, Good Luck with it, it really is a great Comp!

[theman]

Okay, I figured out how to get a volume control out of this thing without adding an extra pot! Use the GAIN pot. Change the 1K resistor (R9) to something less than 330 ohms. Let's choose 220 ohms. Then put a 10K resistor in parallel with the GAIN pot. That's easy ... just put a resistor from the 220 ohm resistor to GROUND.

When the GAIN pot is at noon position, and the COMP is all the way CCW (1:1), the gain is a bit greater than unity. Fully CW, and you get a boost of about 13 dB.  When fully CCW, it is nearly completely silent (muted).

Now, when you increase the COMP pot, the volume will go up. Use the GAIN knob and rotate CCW and this will reduce the volume. So you can always get a volume close to bypass volume at ANY COMP or LIMIT setting. Or you can completely mute it. Or you can boost it by a lot.

This chip never ceases to amaze me ...

[markm]

Mine is like that now.
The Gain control is perhaps a bit more sensitive than yours but, it seems about right.
I have to agree, there sure is alot packed into that chip.
It is amazing!