Q&D Compressor Builders......any opinion?

[markm]

Hello fellow DIYers,
Just did a search about the Q&D from Jack's site and didn't really find any
concrete opinions about it's sound and performance.
Just wondering if I could get some opinions on it's performance and tone/sound
and if it's a worthwhile build.
Please help me out here.
   Thanks all,
       MarkM

[Mark Hammer]

Open to more sophisticated control than is shown, and can probably use something like a FET input buffer (default input impedance is stated as 180k, which is low-ish for a first pedal after the guitar/bass), but the basic circuit works great, and offers excellent performance for the small number of parts involved.  Note that it also serves as a noise-reduction unit.  If I gigged, I'd gig with it unhesitatingly.

[markm]

Hi Mark,
I was kinda hoping you were going to respond to this one!!
I tried to read your build report over at AMZ but, it's a dead link.
If it meets your approval it must be pretty good.
You'd gig with it in it's stock form?
Thanks again,
    MarkM

[Mark Hammer]

It's an extremely good idea to get the SSM2166 datasheet and read it over as fully as you can before attempting this project. http://www.analog.com/UploadedFiles/Data_Sheets/83095497SSM2166_b_.pdf

I don't say that as a caution to warn against problems with it.  Rather, this chip has SO much packed into it that there are a lot of things you can tap, and some priorities will need to be set.  The SSM2166 was intended for use in high-end "control strips" that would encompass a bunch of functions, most likely in tandem with a decent mic preamp, some EQ and metering.  I imagine the specs are at least as good as the BA662 ota chip that Roland has specially made for them, and probably subtantially better than any 3080 you are likely o run across.

Most people will want to use the Q&D2 as a compressor.  It does that very well.  (Of course, like just about any compressor there is some loss of high-end bite, and that can be compensated for.)  But what doesn't get touted as often are its exceptional properties as a downward expander and noise reduction system.  These are linked to the "Rotation Point" function....which, of course you need to read the data sheet to fully grasp.

The "all-dressed" Q&D will be anything but (q & d), and can include the following on its front panel:

• Compression and gain controls (more compression = greater average level so gain should be turned down)
• Noise gate and Rotation Point set for adjusting where the compression starts and where the gating/downward-expansion starts
• Treble emphasis to compensate for "bite-loss" under heavy compression
• Cavg change to produce different attack/decay characteristics

That's 4 pots and a couple of toggles, so while you CAN build it into a 1590B easily (which I've done), a 1590BB is probably desirable to give you room for all possible controls, while having a choice of what pot/knob sizes you can use.

For treble loss, the schem on the data sheet shows a resistor/cap pair (10k/1uf) between pin 6 and ground (These are R2/C2 on Jack's schem).  If the internal schem of the chip is to be believed, this gives the input op-amp buffer stage a gain of x2 when you take the 10k resistor between pins 5 and 6 into account.  Increasing the vaue of R2 will reduce the gain, though never lower than x1.  We can put this to use by sticking a second RC combo in parallel with R2/C2, in exactly the same way it is used in the Rat...except with much less gain involved.  If you were to increase R2 to, say 18k, the input buffer gain for whole spectrum would be about x1.55.  If the R2/C2 combo were paralleled with another RC combo consisting of a 6k8 resistor and a .01uf cap, that would provide for a gain of about x2.47 for content above 2340hz.  You might also consider a 5k6/.015uf combo, which will give a gain of roughly 2.8 for content above roughly 1.9khz.

That's not a huge treble boost, but probably enough to offset what typically happens in compressors.  Between that, a decent high-impedance unity-gain FET buffer added in front of C1, and appropriately set compression, rotation, and avp cap values, you should be able to get some especially transparent compression/limiting.

[markm]

Mark,
Thanks so much for your suggestions and advice.
I do see that this SSM2166 is got alot going on in it!
I'm going to go strictly with Jack's Layout, see how it suits me
and perhaps try a couple of the additions you reccommend.
I wonder if this circuit is similar to Toadworks Mr. Squishy?
Seems as though it may be but, the last time I looked inside one
of those was quite awhile ago and I'll be darned if I remember it.
Thanks again Sir,
    MarkM

[Aharon]

Hey MArk ,is the chip still available,where did you get it?
aharon

[markm]

Yeah,
They are still out there it seems.
Got mine from Steve D.

http://www.smallbearelec.com/Detail.bok?no=234

It's almost ready to fire up and play with.

[Phorhas]

Are 5 volts enough headroom for a compressor?

[Mark Hammer]

Apparently, in this case.  As long as one is not expectingto put out huge signals and the sidechain has enough to work with, no problem.

[Phorhas]

Good to hear... kinda wierd, the mainstream conviction is that the higher the voltage - the better, with compressors

[Mark Hammer]

Good to hear... kinda wierd, the mainstream conviction is that the higher the voltage - the better, with compressors

Well generally, yes.  One usually wants lots of clean headroom if the desire is to have a more transparent sound, and a higher supply voltage can help that.  But not everything needs huge supply voltages to do that.  Keep in mind that some stompbox compressors attempt to have a few stages do an awful lot.  If the stage that does the boosting for the sidechain is also the stage that produces the gain for the audio path, then even though you don't need pristine sound for the sidechain you DO want as much in your favour for the audio path, so assuring that the gain stage can swing wide is a good thing.  The cleanliness is wasted on the sidechain, but it is needed for the audio path.

Good case in point is the stock Orange Squeezer.  Here you have one op-amp gain stage that is both accomplishing the level recovery after the FET attenuator AND providing enough signal gain for the sidechain that controls the FET attenuator.  I would imagine that more clean headroom is a good idea in that instance, although obviously increasing the supply to the op-amp does not necessarily result in less distortion in the FET.

However it does so, if the 2166 is able to produce gain reduction without having to crank the audio signal all that much to do it, then there should be little problem running it off 5v.  Keep in mind that supply voltage lore is typically for discrete or op-amp based circuits.  IN this instance, all the compression functions etc are internal to the chip itself so presumably all the requirements are already worked out.

It's a bit like bringing food to a potluck whose attendance you have no clue about, vs making dinner for yourself and your sweetheart.  In the one instance, making more may be necessary for insurance.  In the other case, it's a known circumstance and you can do with less.

[Phorhas]

Thank you Mark

[Mike Burgundy]

fwiw,
the QD2 is the only comp that's always been and still is on my board. It sees just bass nowadays (I think I slowed down the response slightly, and did the bright-mod but I'm not really sure unless I open it up), 4 and 5 string, active and passive. Worked great on guitar (tele, 335) too. I now use it for added "hello-I'm-in-the-mix" when I'm doing stuff that just lacks mass (such as tapping on a neck-through bass) and needs serious oomph. If however I want something thats a more expressive effect (think dub, or anything else that's obviously processed) I actually go for a tinkered Ross. Situations differ, opinions differ, so do tastes and thank the lord for that.
Obviously 5V is ample to work since I get no distortion of any kind, although I *think* the QD2 gets a bit lazy when you go really low (low B anyone?) and really whack it. It's like the attack is more compressed and the release lags a bit, but it might be suggestion. I'm not sure where the difference is between real artifacts and psycho-acoustic suggestion on this one - there is a very strong expectation of how something sounds, and when it behaves differently...
The controls take a bit of getting your head around it, since they interact differently from a "standard" studio comp. The datasheet and Jacks article explain all.
All in all - build it. Even if this is not your main squeeze (pun intended, ahem) it will be a very valuable addition to your arsenal.

[markm]

I'm still wrestling with getting it to function properly.
THEN I'll consider some mods!
Right now it is in "limbo" on the corner of the bench.....
we really needed some time away from each other if ya know what I mean!

[calculating_infinity]

I know this is an old topic, however since there was a recent group buy of SSM chips this may prove useful.  I too tried to read Mark Hammer's build report only to find a dead link.  I would like to hear your opinions on the Q&D Comp 2 now MarkM?  I am wondering why the impedance is so low on the input.  I have a AMZ jfet buffer sitting around ready to pop in if needed.  Any thoughts?  Thanks!

[MartyMart]

Jonathan, there's a follow up thread on this with loads of info, I built one a few months ago
and think that it's a GREAT comp/NR circuit.
It's very simple to build as everything's on the chip, you only need a few selected components
and pots/jacks etc.
Well worth building IMO
MM.

[Mark Hammer]

I know this is an old topic, however since there was a recent group buy of SSM chips this may prove useful.  I too tried to read Mark Hammer's build report only to find a dead link.  I would like to hear your opinions on the Q&D Comp 2 now MarkM?  I am wondering why the impedance is so low on the input.  I have a AMZ jfet buffer sitting around ready to pop in if needed.  Any thoughts?  Thanks!

This chip is intended to be part of a "control strip" for a mic signal.  Consequently, it essentially ignores the world of guitar impedances, having a fairly low input and output impedance.  It also assumes the input signal will need a helping hand as far as level goes.  You will note that the appnotes say "typical speech application" for a circuit very similar to Jack's.

These two occurences suggest that Jack's application of the chip could be improved upon for guitar if: a) one uses a high-impedance buffer in front (any FET buffer will do), and b) drops the gain of the first op-amp stage by increasing the value of that 10k resistor to ground.  Currently the op-amp stage has a gain of x2.  Dropping it to x1.2 by swapping the 10k ground resistor for a 47k resistor would help a bit.

You will note that the envelope follower derives its control signal from the output of that op-amp stage, so having the "right" signal level there is critical to having the right feel for both the compression and downward expansion.

I've found that the gain control can provide a LOT of output, and unlike many commercial pedals more compression = more volume, so it might be better to have the gain control be an internal trimpot and add a 10k-50k volume pot on the output, so that you can get high compression with little or no volume change when you want it.

[markm]

I performed some of the mods suggested by Mark H. with fantastic results.
There is another "Mystery Thread" that goes deeper into taming the output of this circuit as it has ridiculous amounts of Gain.
However, I searched and cannot find it as of right now but, I may have saved the info in Notepad at home.
I'll have to check and see.
Overall, it's a very good circuit but, it's a bit more complicated than the standard Dyna-Ross type compressor which is why I don't use it all that much. It's like the Trem-Lune of Compressors...... 

[Mark Hammer]

Although the chip offers a great many options, it actually doesn't have to be built with anything more than a compression-amount and volume control.  Clearly, for folks like us all those other potential adjustments are too good to resist, but there is no reason why one could not simply replace some of the variable resistances (like gain, rotation, etc.) with suitably chosen fixed resistors, and have a circuit that is no bigger than a Dynacomp, no more complicated, and with a lower parts count.

[MartyMart]

.... to follow up, I too used Marks and markm's reports to get a superb result, bufferd, gain tweaked
and external vol control add up to a superb and flexible circuit.
MM.