Attack (Recovery) in an MXR DynaComp or Ross compressor

[Manolo Dudes]

New interesting article (collaboration) in www.pisotones.com:

Attack control in an MXR DynaComp

Go get those web translators!!!  :wink:

[Edit: modified the title of this topic to make it a bit more precise  :roll: ]

[Jay Doyle]

Not to beat a dead horse here, but this is a Decay control, not attack.

The transistors in the FWR are set up to have a high gain and to be always on. The diodes to ground connected to the base ensure that each transistor of the FWR has it's base above it's emitter to keep it on, with a huge gain, meaning with no signal, the signal at their collectors is at V+. A signal can only serve to reduce this voltage.

Think about it at the extremes, without signal, when the comp is running at it's highest gain and with a large signal that turns the compressor gain all the way down.

Before signal is applied, the 10uF cap is sitting at V+. This forces the FWR output transistor to drive the 3080's Iabc port as hard as it can; the circuit is running at it's highest gain. When a large signal is applied, the collectors of the FWR input pull hard to ground, causing the cap to discharge through the collector-emitter channel in the transistors (it's the only place it's charge can go), rather quickly as the parallel resistance of the two channels won't be much more that 50ohms. This reduces the current into the Iabs port of the 3080 to be at the minimum, it's lowest gain. This is the attack, the time it takes for the cap to discharge through the collector-emitter channels.

The decay comes when that signal is removed and the collectors of the FWR move back up to V+. Now the cap can only charge up through the 150k resistor on the collectors. Changing the resistance on the collectors changes how long the cap takes to charge back up once the large signal has disappeared. As it moves up to V+, the gain increases. This is the delay. A pot on the collectors adjusts the decay.

To adjust attack you need to put a pot from the emitter's to ground or insert a standard a/d network after the FWR.

At first it seems like an easy mod, but you are modding for adjustable decay, not attack.

[Mark Hammer]

Nice technical explanation Jay.  Much appreciated.

I'm just guessing here, but the control is perhaps more properly thought of as a "recovery" control.  It *DOES* adjust the decay time, but given the rate with which new events are thrown at the envelope follower for detection, a longer decay time means the envelope follower is essentially unready to fully detect another event, such that it gets blended in with the previous event's envelope (and gain reduction), such that the attack transient of note N+1 is effectively blotted out.

The long and the short of it is that the mod creates the illusion of a change in attack time by slowing down (or speeding up) decay time, which ends up having an impact on the capacity to recover for the next note.  "Recovery" is probably too big a word for legending on most of the pedals that use it, and "Decay" is not sexy enough, so I suspect they used "Attack" because it was short and that's what it felt like.  

Having added one of these onto a Ross clone, I can confidently report that:
a) it works
b) you notice it more when you place lots of notes close together
c) you may not find yourself doing much more than setting it to 1 or 10, since the effect of small variations is not readily apparent
d) it adds no noise or anything unpleasant

Personally (and especially given how difficult it would be to find a suitable pot), I think the smarter way to do it is to stick the max value fixed resistor on the board (160k), and use a 3-position SPDT on-off-on minitoggle to select either none or one of two parallel resistor values to get slow, medium, fast decay/recovery.  Fast = 10k, slow = 160k, and I'm gonna guesstimate medium as around 50k.  If you go with a fixed resistor of 160k, a 68k resistor in parallel will yield a combined resistance of just under 48k, and a 10k resistor in parallel will get you 9.4k....close enough.  

Beyond going straight to the settings you want, doing it this weay saves you the cost of a pot and knob, and also makes installation into a 1590B with room to get at the controls that much more feasible.

[Jay Doyle]

Mark,

I see your point.

I just wish that it was properly labeled. As it is, it just confuses the issues involved, IMO.

Jay

[Mark Hammer]

In a perfect world, Jay, there would be one unique term for a given control, and it would mean exactly what the control does.  So close would the correspondence between term and function be that every EE prof would sign off on it.  This may result in a lot of firings at DOD or Danelectro ( :wink: ) but so be it.

[R.G.]

Seems to me that you could insert a resistor between the pulldown transistors' collectors and the 150K pullup resistor to limit how hard they could pull down, and hence how fast they could discharge the timing cap. I think that's a true attack control.

[Manolo Dudes]

As Mark said, it is probably an "illusion" of Attack. The fact is that it indeed "sounds" as an attack control.

Also the final solution was indeed not to use a pot but a toggle to switch from "Hard" (10K resistor) to "Soft" (150K) as there were no significant use for the "in betweens".

Jay, I'm going to talk with the author of the article for a change in the title.

R.G.'s suggestions are also very interesting as well as Jay's explanation.

Thank you all for your interest.

[Manolo Dudes]

In a perfect world, Jay, there would be one unique term for a given control, and it would mean exactly what the control does.

I've got a compressor, the DOD FX-84 "Milkbox". Not a bad compressor albeit a tad noisy. The controls are labeled "Quarts", "Pasteurization", "Cream" and "Spill"... go figure!  :lol:

[Boofhead]

Not to beat a dead horse here, but this is a Decay control, not attack.

Theres quite a few topics that keep coming up - we need a link page for answers to common questions.

[Mark Hammer]

Manolo,

The "geniuses" at DOD who come up with cute terms to label their controls drive me up the wall, across the ceiling, and down the other side.  It's as if they have some sort of sacred mission to keep the unemployed people who frequent music stores permanently stupid and confused.

I don't expect EVERY musician to be a techie, and there is certainly something attractive about those who don't get tech-obsessed and prefer to just plug in and play, but at least don't stand in the way of people understanding how things work.

[R.G.]

Theres quite a few topics that keep coming up - we need a link page for answers to common questions.

That's the motivation for a lot of the things I write up. Sadly, I notice that not all that many people go read the info at GEO before firing off questions here.

I think that no amount of link pages, FAQs, etc. will ever keep these things from coming up - beginners will keep on blasting in without reading the preparatory material. It's what beginners do. I've largely quit popping in with "go read the FAQ" notes, certainly less than I used to. It saves me a lot of typing.

[Manolo Dudes]

Well, the DOD naming affair was evidently "tongue in cheek".

The real reason for naming it as "Attack" is that, as stated in the page cited above, the idea was originaly extracted from the BOSS CS2 compressor. It was indeed labeled "Attack" but, ASMOF, you're all right, it's a "Decay-control-performing-as-a-recovery-and-giving-us-the-illusion-of-attack"  

Have a look here:

http://www.pisotones.com/Articulos/DynaMod/CS2original1.jpg
http://www.pisotones.com/Articulos/DynaMod/CS2original2.jpg

See what I mean?  :wink:

[Jay Doyle]

Seems to me that you could insert a resistor between the pulldown transistors' collectors and the 150K pullup resistor to limit how hard they could pull down, and hence how fast they could discharge the timing cap. I think that's a true attack control.

R.G., you mean to put a pot from the junction of the 150k and the cap to the collectors of the FWR transistors?

I can see this working very well, but the FWR won't be able to go fully to ground. Though the small resistance needed to adjust attack, which is usually kept real fast anyway, with a 10uF capacitor may not affect performance very much.

Good point, never thought about that.

[Boofhead]

That's the motivation for a lot of the things I write up. Sadly, I notice that not all that many people go read the info at GEO before firing off questions here.

Unfortunately that's true.  There's a lot of stuff on your site which puts people's minds on the right track.  No doubt it's a source of frustration for those who bother to put the effort into such pages - especially when they hang out on the forum (grrr...it's already on my site...).  Many questions can be answered by typing a few words into google too.  The main problem is not everyone is savy about tapping into the info on the web - nor the forum archives for that matter.

I think that no amount of link pages, FAQs,

Very true, I guess we can only try to minimise the damage so to speak.  The main angle I had in this case was to have a "common questions" section which is small in size, say one page of links with clear titles.  If it's too large and too comprehensive it will just bamboozle people - especially beginners who don't really know what they are looking for.

[Dongle]

Hi guys!
Sorry if it is unusal to bring old threads back again, but I think it is better than making several similar threads...
I have a question: I did this mod with a pot (250k, lin in series with a 10 k resistor)
My problem is, that it also controls the level from 0-max... Is this normal, or did I make a mistake?
Thanks guys!
Dongle

[Mark Hammer]

You likely made a mistake.  It should not alter the volume level at all.  It will alter how quickly the circuit can resume max gain for the next/first note picked, but if you are simply strumming lightly in a way that should not elicit gain reduction (and make sure the compression amount is turned down from max) you should not hear ANY change in level.

[zombiwoof]

On the DOD comp pedals, they alternated between calling the pot "release" and "attack".  The terms were used at different times for the same control.

Al

[zombiwoof]

In a perfect world, Jay, there would be one unique term for a given control, and it would mean exactly what the control does.

I've got a compressor, the DOD FX-84 "Milkbox". Not a bad compressor albeit a tad noisy. The controls are labeled "Quarts", "Pasteurization", "Cream" and "Spill"... go figure!  :lol:

I have a Milkbox, and it doesn't seem noisy to me, compared to other comp pedals I have.  On that one, "Spill" is the attack (or release, or decay, or whatever) control:

http://www.americaspedal.net/fx84/

I think it is a very under-rated compressor pedal.  The four controls really let you dial in the compression.  You can get them for small change on Ebay.   I think I got mine (which is in mint condition with the box and papers), for around $25-30!. 

Al