Modding the SSM2167 compressor module

[Synthlurker]

I thought that I might as well write about this board as nobody else seems to have mentioned it here before. These boards can be found on Aliexpress etc. for 2-4€ so quite cheap as just buying one of these chips costs practically as much (unless one buys +100 at once). It would be nice to see SSM2166 modules too as it is clearly superior chip compared to this SSM2167 but this should still be good enough for basic stuff. This chip also has been in the market for 25 (!) years now, hopefully it will not be phased away any time soon. There are only some rather basic articles on this chip floating around at various ham radio pages, so here is a more in-depth writing on how to get the most out of this chip.

The boards are already well marked with the basic connections (+5V, IN, OUT) and it does hint that R1 and R2 could be changed from resistors to potentiometers. I have marked the components that can be removed in blue and lugs for connecting the mods in yellow. Comparing to datasheet, R1 is Rgate, R2 is Rcomp, and C2 is Cavg.

The Rcomp could be just 200k linear potentiometer or alternatively 250k linear pot with 470k resistor between the lugs to make it easier to set low compresison values. Higher values do not give more compression, 10:1 is practically the internal limiting ratio of the chip.

Rgate is best with just 5k linear pot. Higher values do not drop down the gate due to noise according to datasheet.

The Cavg affects both the compressor and gate speed quite nonlinearly as the chip has some kind of auto-release circuitry. The SSM2167 article reads “The rms detector filter time constant is approximately given by 10 × Cavg milliseconds where Cavg is in μF.” Reasonable values for Cavg range from 1μF to 22μF (and higher), but if you are interested in more interesting behaviour 100nF will result in rather funky limiter performance. The gating becomes rather unnatural at low Cavg capacitance values (but might still sound cool in some cases!).

The chip does have some kind of internal limiting, but just to be sure I would recommend some kind of clipper input circuit before the module. Here is my zener clipper with nice soft knee limiting. Alternatively one could just bias the input signal to +2,5V and add signal diodes to +5V and GND to get more hard knee input limiting.

Output level pot is useful as increasing comp ratio increases gain. However, the chip seems to get rather noisy with high output resistances. A simple buffer before output level should improve performance (I haven’t tested this yet).

One could also increase the supply voltage up to 5,5V to increase headroom. But there is always a risk of burning or shortening the life of the chip.

As the gain increases with higher resistance, one could turn this chip into a pseudo sidechain-VCA with LDR-potentiometer pair and for Rcomp. Furthermore, for Rgate a FET optocouler would make more sense as the resistance needs to drop very low to effect the threshold. Maybe optocoupler could also make sense for Rcomp as well… I might actually try to implement this in next iteration of my modded circuit

I will try to send some performance results at the start of next year when I have my Eurorack PCB version built

[krztoff]

Do you mean RComp in the beginning of the 3rd paragraph? (200k pot)

[Farabide]

Great post!

I think I may have used the 2166 as a compressor for a bass sans amp years ago. I am often amazed when I see a huge modular set up and no compressor(s), especially on bass and drums. George Martin raved about how compression allowed you to hear every track in detail and still create a balanced ensemble in a track. My main compressor is patched into my rig allowing me to route single tracks and mixer outputs.

As they are so cheap and small can I suggest your planned build include 3 or 4 chips and some kind of patched or mixer matrix. Blending compression is divine. I look forward to seeing your results.

[K.ostas]

I think it is not surprising that we do not see mny compression modules. In order for compression to work, there needs to be a wide dynamic range. Unless one is able to dynamically chage the amplitude of the bass/drums, compession will not make a difference. If you think of it in reverse, you don’t need a compressor. You just reduce the dynamic range by not modulating the VCA of the bass/drums

[Synthlurker]

Fixed the typo

But anyways this chip is quite nice for cleaning up drums and acid bass sounds with the noise gating.

[Farabide]

Yes. Commonly compression is used to take a wide dynamic range of natural sound and narrow the range down to improve clarity etc.

Natural sounds like a microphoned drum kit, bass amps, vocals etc can each produce a huge and diverse dynamic range and benefits to a mix of song dynamics. Synths mostly operate at line level and their dynamics are handled at source or in a mix however they are not immune to noisy clutter. I’ve used compression with older mono synths to remove artifacts like gate click, line noise, unwanted harmonics and more. You can ‘tuck’ one synth voice into a track without it bleeding into another or deliberately cause dynamics to clash. The algorithm is very versatile and has many uses. I’ll have a dig around for the book I got in my 20’s which was like the music producers anarchist cookbook and filled my head with what ifs. On the old analogue desks applying individual compression (via insert not fx loop or post mixdown) was/is often used to condition a track before adding it to a final mix. Again, see George Martin.

FM synthesis with compression is a perfectionist swamp that can produce unique timbres though it’s almost impossible for me to explain to people what it actually does. (Like the presence knob on a power amp. Where’s @analogoutput when you need him )

If you need an example of perfect compression I’d suggest checking Peter Gabriel’s album “So!” which has been taught at Berkley as a practical guide to compression use.

There are so many ways to incorporate compression into music; use it like a noise gate, a filter, a pre-amp, a sub mixer, conditioner, on stage monitoring, the list is extensive.

Now it is easy to overdo and turn your opus into dull flat mush so use it sparingly but do experiment as sometimes the results can be spectacular. Percussion can be punchy, pads less ethereal and more choral and so on.

All compression does is manage the dynamics of audio signal and while that’s the usual use case, it’s not the only use in making music.

Do experiment. Feed a microphony module into compression on an FX loop where you can mix clean and compressed sound. Use a compressor on a synth voice to provide more clarity for gates and envelopes.

A compressor, especially more than one, on a track is a great toolkit to help you shape the audio.

Waffle over (sorry Rich!)

I may add one of these chips to an active mixer and also see what happens when you ‘circuit bend’ the board…

Enough so! All the best.

[K.ostas]

I got one of these myself months ago with the intention of bulding something, but it is currently 6th in my to-do list after three built modules that need debugging and another two that are half built.

I don’t think that I will do any modifications to work with modular synth levels, though, as these do not give a dynamic range that warrantees such compression. My intention is to turn this into a microphone (dynamic/electret) preamplifier/compressor. Apart from adding the 5V regulator (tap off the 9V supply that I have), potentiometers, and electret bias switch, I was thinking of taping the signal off pin 6 to drive a simple VU meter.