Ross Compressor - initial attack

[Unlikekurt]

I've searched the internet and the forum and haven't come up with an obvious solution.
Is there a way to tame the initial attack of the Ross circuit.
I've had one built on a tonepad pcb for quite a while now and its great.  I adore the little coloring it gives the tone and it certainly is helpful at evening out the playing (i use it with bass) as well as allows the amp to get a little more driven when i really dig in.
That said, with it engaged, the first clamp down always has a sharp transient pop/click to it.  It's not the end of the world in the live environment, but when recording it can be an obnoxious pain to have to edit all the pops at the beginning of passages.
Any ideas?

[Mark Hammer]

It's quite possible that the little pop/transient you hear at the beginning is a result of the attack not being fast enough.

Like many compressors, this one works by pulling down the gain in response to a big input.  But it takes a moment to do that, so what you hear is essentially the signal - as is - followed immediately by a rapid gain reduction.  The "as is" signal sticks out in stark contrast to the gain-reduced version, sounding like some sort of imposed pop....which it isn't.

I suppose if the rectifier circuit was super-fast, you might be able to pull the gain down fast enough that the initial transient would be tamed before it ever reached its peak...but that's unlikely to happen.

Alternatively, if the unit has a much slower onset of compression, such that the contrast between uncompressed and compressed was not quite so sudden, you'd eliminate the pop.  But then you wouldn't really have compression, would you?

I suspect the real solution lies in using less than maximum compression, so as to reduce the contrast, while still having a relatively responsive rectifier circuit.

[Unlikekurt]

Thanks for the response Mark.

A lot of what you said makes a lot of sense.  I'm not really dialing the "sustain" up all that hard.  That said, I do play very heavy handedly.  Ultimately I guess I'm expecting 1176 attack speeds, or a look-ahead, from the ross and really shouldn't be. 

[Mark Hammer]

More and more, people are turning to a "blend" feature in stompbox compressors.  This is done primarily to avoid losing the brightness of the input (which tends to disappear, because of how compressors alter the frequency balance via their action).

I have not had any experience using such features.  I'm curious as to whether it has any ability to reduce the audible pop/thunk you describe.  My hunch is that it doesn't, but I'll defer to others who have used such a feature.

[karbomusic]

I sort of don't care for compressor stompboxes unless there is a blend. That doesn't make it sound like an 1176 but it gives the sustain minus some of the smash/pop factor. This is my version of the orange squeezer in such form which ended up on my pedal board and another for my acoustic...

http://www.diystompboxes.com/smfforum/index.php?topic=110545.msg1014321#msg1014321

[Unlikekurt]

Mark, the pop i'm hearing is 100% without a doubt the initial transient spiking through.  The attack time would be perfect for giving a snare just a hair more snap.  But in my case, its just a hair too slow.  I started looking into the Keeley Pro and the Cali 76 pedals due to their abundance of controls and descriptions.  That said, what's interesting about the setup for me, is that I like the colour the Ross imparts.  It isn't transparent. 
This brings me to the idea of a blend.  Although I can see the merits of the parallel compression option, for me, it's as much about the color as it is about the compression from the ross.

[Fender3D]

Hasn't it a trimmer to defeat compression pop/click?

[merlinb]

More and more, people are turning to a "blend" feature in stompbox compressors.    I'm curious as to whether it has any ability to reduce the audible pop/thunk you describe. 

In a compressor, 'blend' is functionally the same thing as a 'ratio' control, so it will reduce the attack transient inasmuch as it reduces the the amount of compression for a given input level.

You should be able to reduce the attack time by reducing the 10uF cap between Q3/Q4. Perhaps try 4.7uF. (This will shorten the recovery time too unless you compensate for it). However, with too small a value you may begin to notice distortion on bass notes, though this could be perhaps be supressed with some heavy-handed bass cut before the compressor, e.g. a smaller input coupling cap.


(And what's happened to the look and feel of this forum? I go away for a few days and when I come back it's gone all ugly and generic... )

[Mark Hammer]

I was going to suggest reducing the 10uf averaging cap, but it struck me as introducing more complications than solutions, and thought better of it.

Hasn't it a trimmer to defeat compression pop/click?

My understanding was that the trimmer was for reducing "control-voltage feedthrough" or something, which does produce a thunk, but is not the same thing as the sharp contrast between an initial sharp attack, and instant gain-reduction.

[Keppy]

Since only the first note thunks, the problem may have less to do with the attack time and more to do with the threshold at which compression begins. When you go from NO compression to LOTS, you notice the attack a lot more than if you go from SOME compression to LOTS, even with the same attack time. If I'm recalling previous threads correctly, raising the threshold can be done by changing D2 & D3 to a higher forward voltage, such as LEDs or two diodes in series.

[samhay]

More and more, people are turning to a "blend" feature in stompbox compressors.    I'm curious as to whether it has any ability to reduce the audible pop/thunk you describe. 

In a compressor, 'blend' is functionally the same thing as a 'ratio' control, so it will reduce the attack transient inasmuch as it reduces the the amount of compression for a given input level.

In a feedback compressor, I don't think that blend and ratio controls are completely interchangable, especially not if the mix comes after the point that feeds the sidechain.

[midwayfair]

Thanks for the response Mark.

A lot of what you said makes a lot of sense.  I'm not really dialing the "sustain" up all that hard.  That said, I do play very heavy handedly.  Ultimately I guess I'm expecting 1176 attack speeds, or a look-ahead, from the ross and really shouldn't be.

There's nothing I can see in the Ross circuit that prevents it from producing attack speeds every bit as fast as that rack unit. They both use a rectifier feeding a transistor as the variable resistance element, unlike an optical compressor which has to contend with the turn-on time of the LDR, which may be many times longer than the attack in mS of the rectifier. The transistor in the Ross isn't different from the FET in the 1176 in that regard. The main difference in the rack unity is the precision of the rectifier and the complexity and noise performance of the dry path (which doesn't use an OTA chip). In any case, both are capable of attack times faster than a half cycle of most notes, which might as well be immediate.

You may want to check out the Engineer's Thumb. Its recifier is a bit easier to tinker with, and while the attack in the stock unit is wired to be a little less than immediate, it's easy enough to modify to be faster.

Keppy: I don't think the treshold matters much. If it did, then turning the compression pot up higher would have gotten rid of his problem.

Mark: Decreasing the averaging cap could actually reduce the attack time depending on what the driving impedance is. Unfortunately, the Ross doesn't have a parallel resistor (IIRC) to increase to bring the decay time back down. This is another reason I think it's worth considering a slightly modified version of the ET.

[PRR]

> nothing I can see in the Ross circuit that prevents it from producing attack speeds every bit as fast as

It is not obvious, but the impedance in the Base leads of the rectifying transistors (10K resistors in the phase-splitter and the 0.01u caps) limits the rectifying transistors' current and prevents infinite attack.

This is easily fixed with brute force, though when you lower the 10K you throw more load on the '3080 and have to re-ponder a lot of stuff.

> Ross doesn't have a parallel resistor (IIRC) to increase

It does; that's the red-mod in the drawing in Reply #7. (It bleeds up to V+ instead of down to ground because of the way the '3080 works.) Again, you can only up this resistance so far before being loaded-down by the 527K-27K network in the buffer to the '3080's Iabc pin.

I do think the Ross was good for its time and place; but today that Thumb is a much better plan and possibly easier to build.