Schematic

Using The Unit

Sound Samples

Reference
Update History

Smash-2 / SM-2 (revisited): an up/downwards sustainer (dual mode compressor) (click on the picture to load larger image) Nicholas Boscorelli’s Project #38 (Sustain-O-Matic) [1] takes compression for the purpose of prolonging  tone to another level. The obsolete dynamics processor has been replaced by THAT4305, which involved several adaptations. The amount of gain boost resp. reduction can be dialed in individually, which at the same time avoids a couple of problems usually associated with such circuitry. Threshold is common for both compressor modes. The side-chain can take its clues from either the signal itself (feed-forward) or from the output (feed-back). Both settings have distinct virtues. A special gain reduction indicator has been devised by the Author, comprised of an RGB LED and driver circuitry, which makes dialing in of the sustainer a breeze.

It was soon becoming clear that a universal (read: swiss-army-knife compressor) was something a recording studio needs, but was useless for live guitar playing. Consequently, all musicians I know that bought such a thing, did not know what to do with it.

A specialty compressor, like a sustainer, on the other hand just does one thing, and this hopefully well. With SM-2 (Sustain-O-Matic II) Boscorelli has hit the nail on the head again. He cleverly combines a traditional downwards-compressor (that squashes all program material above the threshold) with an upwards compressor (that boosts all program material below the threshold), while circumventing the pitfalls usually associated with such circuitry.

But what is an upward compressor?

"In downward compression, threshold defines the point above which gain reduction occurs. In upward compression, threshold defines the point below which boost ensues. Threshold may define both points in dual-mode compression (…)"    - Nicholas Boscorelli, The Stomp Box Cookbook, p. 156

"A more aggressive-sounding sustain uses upward compression. In this case the compressor runs at unity gain above threshold, but boosts signals that fall below threshold. (…) The two types can work together. Downward sustain is easier to implement (…)"    - Nicholas Boscorelli, The Stomp Box Cookbook, p.158

Up/down/dual compression modes compared (click on the picture to load larger image) The first three charts show equal compression ratios for all, while the last one  employs different rations for up and down compression, as SM-2 can do. This is where SM-2 shines.

In the book, Project #38 (SM-2) logically comes after Project #35 (SOM-2), and it is pure coincidence that I built SOMA earlier. Both units in the original use the obsolete SSM2120 dynamics processor, which I replaced with THAT4305. Several adaptations had to be made, which will be described in the following. Since many subjects are already described in my article on SOMA [2], I suggest the kind reader to study that before proceeding.

Same for the original document. I can thus restrict my comments and only address specific problems hereafter.

Boscorelli was not exaggerating when he said:

"SM2 produces very natural sounding sustain (…)"    - Nicholas Boscorelli, The Stomp Box Cookbook, p.252

I can fully recommend building it.

Noise-wise, VCA compressors are far better than other architectures, since they are only attenuating. But when we speak of compression, we normally associate this with downwards compression, meaning that everything above the threshold gets progressively knocked down.

But here we also have upwards compression, meaning that everything below the threshold gets increasingly boosted. But this is where VCAs do not really excel.

"Upward sustain exposes what may be the chip-VCA’s worst flaw."    - Nicholas Boscorelli, The Stomp Box Cookbook, p.169

"Where downward sustain is naturally quiet, upward  sustain magnifies system noise. (...) Higher gains are useful for "rad" sustain, or to establish feedback, at the cost of  conspicuous operation and overt noise."    - Nicholas Boscorelli, The Stomp Box Cookbook, p.167

"(...)when not playing, either turn the axe's volume all the way down, or unplug it from the box."    - Nicholas Boscorelli, The Stomp Box Cookbook, p.252

SM-2 PCB board: (click on the picture to load larger image) Here are the schematic and PCB that incorporate THAT's dynamic processor into the signal part of SM-2.  The large IC in the bottom center is representative for the DIL-16 to SMD adapter.

As found with SOM-2, the side chain is laid out perfectly. There is basically nothing to change, except:

There is a fatal error in Boscorelli’s schematic. IC3a (designators according to the original document) is supposed to do the following:

"The resultant control voltage is buffered by IC3-a, whose output swings negative in the absence of a positive voltage from the level detector, due to the bias applied through R4."    - Nicholas Boscorelli, The Stomp Box Cookbook, p.249

Unfortunately, as drawn, it cannot possibly do that. The inverting and non-inverting pins are swapped.
That circuit does nothing, it just sticks to one rail. Fixing that, it does its thing exactly as described. Nick, you're challenging me.

The feedback resistor on that buffer has been halved (22k).  The reason for this measure was that this buffer only does (well) what it is expected from, if both supply voltages are of equal magnitude. In our case (using a charge pump), the deviation of the negative rail from the ideal value, caused by the high loss in the charge pump, kept the positive half wave from going all the way to zero, which I suspected to cause problems. A computer simulation revealed that a reduction of the gain resistor mitigates that problem. This just changes the range of the pots slightly. Also, a rail-to-rail op-amp in this position might help.

A further reduction of this resistor to 18k poses a limit to how far negative the CV can float. In other words, this sets a limit to the maximum idle boost, which does not hurt because it tames noise a bit.

The change in the drive circuitry following the CV divider, which is required for THAT4305, has been described in SOMA.
The auto variable attack  circuit used has been described on p.180 of the book [1].
The resistors around the threshold pot have been adapted for a subjectively better adjustment range.

---

Using The Unit

We've mentioned it already:

"SM2 produces very natural sounding sustain (…)"    - Nicholas Boscorelli, The Stomp Box Cookbook, p.252

"A feedback control path yields upward sustain that sounds so natural that only A/B comparison betrays it (...)"   - Nicholas Boscorelli, The Stomp Box Cookbook, p. 169

... great for tone, but that bears the danger of over-compression again. An indicator is being asked for again.
Since the (unipolar) gain reduction indicator [3] used with SOMA does not do all we want, a suitable unit has been derived off it, which is described here [5]. I strongly recommend adding this because it helps a lot.

Assuming the kind reader followed my recommendation, we can turn the unit on. The adjustment procedure is somewhat unusual at first.

• Set the upwards ratio CCW initially. Then play with THR and downwards ratio. This part is a stock downwards compressor. You will get the hang soon.
  
• turn the upwards ratio up, until you see the green LED coming on and until you hear some noteworthy sustaining. Now you understand the thing.
  

Back To Index

---

Sound Samples 
 

A video is worth a thousand written words. A short demonstration of the unit with the bipolar gain reduction indicator. The quality has been reduced to keep the file size small. The camera goes blind on the bright LED. The following equipment has been used: Proprietary speaker cabinet with built-in omnidirectional capsule microphone Recording device: PC interface into DAW Digital camera Video editing in Mixcraft 10 all recordings untreated

Back To Index

• Mar. 12, 2026: added schematic (forgotten!)
  
• Mar. 11, 2026: first release