DOD 280 A Optical Compressor, underrated or undiscovered ?

[Krinor]

I'd like to draw some attention to the DOD 280 A Optical Compressor. The schematic can be found at GGG (pdf project file):

http://www.generalguitargadgets.com/pdf/ggg_opcomp280a.pdf

I have built two of these, and I find them to be very nice and subtle compressors which serves a lot of different purposes in my home studio. Searching this forum for more in depth discussions about this effect, possible mods etc, yields almost no information of much value. Below are links to all the threads which can currently be found:

http://www.diystompboxes.com/smfforum/index.php?topic=64395.0
http://www.diystompboxes.com/smfforum/index.php?topic=54046.0
http://www.diystompboxes.com/smfforum/index.php?topic=64416.0
http://www.diystompboxes.com/smfforum/index.php?topic=62155.0
http://www.diystompboxes.com/smfforum/index.php?topic=56859.0

Could someone with an interest in compression please review the project file and maybe suggest some possible improvements which can turn this simple compressor into an even more usefull unit ? Of particular interest would be possible ways to control attack, release and threshold.

Thanks!

[guitarman89]

I've built it three years ago (my third built), and I test it only some weeks ago. It is a very nice compressor. I have a behringer cs100 (i think a boss CS-3 clone) but it isn't a good compressor. Too noisy,almost like a microphone into a desert storm! It kills my sound. I don't like it, and when I've tried the dod280 I said "this is my compressor!". Some mod can be very useful on this project; however if you are searching something complicated it rocks!! Only two knob! 

[chi_boy]

Seems there is no love for the DOD 280a.

There really is not much info available, but yet it's still called a "popular" compressor at GGG.

I bought the board anyway, just to check it out.

More info would be nice though.

Cheers,
G

[Mark Hammer]

It has a few things going for it.

1) It doesn't use an OTA which means the signal down not have to be attenuated to accommodate the OTA, then boosted back up again to accommodate the needs of the rest of the pedal chain.

2) It uses an LDR as control element which makes it more immune to distortion and envelope ripple.

3) It uses a rather unusual rectifier circuit that has no series diodes to introduce "threshold" effects.

4) It uses a low-power op-amp.

5) It's basically a one o-amp booster that just happens to have some automatic gain-control built in.

Essentially the transistors draw current from the supply, which then lights up the LED, reducing the LDR resistance and the gain of the op-amp.  C6 is the time-constant cap, and changes in the value of it will alter the rate that the stored charge accumulates as the transistors pull current out of the supply.  Larger values make it take longer to fully charge the cap, and longer to discharge it.  I imagine that R10 is functionally equivalent to the famed 150k resistor found in the Dynacomp/Ross and so many others, which s often co-opted into a recovery-time control.  Since the LED has to have enough current passing through it to light up, you probably don't want to tinker with R10....at least not very much (a 2k pot?).  If you feel like playing, try a 10uf cap instead of 47uf and see it that gives you anything different.  It should probably push the circuit more in the direction of something like an Orange Squeezer.

The Compress pot simply adjusts the gain of the op-amp (whose output is inverted without gain by IC1b) to drive the rectifier circuit harder or softer.  When you adjust the gain via the "ground leg", higher gain settings (reduction in the ground-leg resistance) result in loss of bass.  In this case, the bass rolls of at around 7hz at minimum compression, and around 153hz at max compression.  I was going to suggest possibly having two ground legs with a little more gain for the treble, but because the same op-amp that passes audio signal is also used to determine the control signal, it is unwise to start tinkering.

[Johan]

the rectification is interesting..rectify through bias(or lack of..)..cool..
I guess the interaction of R10 and C6 could be a timeconstant for the release...with the 4k7 and 47uF, you have a releasetime of 220 mS...
redusing R10 to 1k5 will make it even shorter and probably not far from the LDRs responsetime..try a switch for different (bigger)values of C6 instead
R10 is also a current limiter for the LED and therefor it also controlls the "ratio". try a 5k pot (reostat)
inserting a smaller pot ( say 1k connected as reostat) between the junction of Q1/Q2 and R10/C6 could give you a attack controll.

..PLEASE NOTE THOU...all of these mods would be interacting and it might be hard to find the sweet spot...

...oh, well...
j

[Mark Hammer]

..PLEASE NOTE THOU...all of these mods would be interacting and it might be hard to find the sweet spot...

...oh, well...
j

Well that's just it.  One has no idea about what sort of current-limiting resistance is possible or tolerable.  That's why I woldn'ty play with the current-limitin series resistor much, if at all, and would focus on the cap alone.  Not the greatest degree of fine control, to be sure, but it's not like the pedal is to be matched up to a wide array of signal sources.  krinor is essentially looking for some variety in "feel", and changing between a few cap values might deliver that.

[earthtonesaudio]

I would experiment with the transistors.  You could add some resistance in series with both the bases with a dual-ganged pot, which would reduce sensitivity.  Also you could provide some positive bias to the transistors, preferably adjustable, to be able to bias the LED "barely on."

Or replace them with MOSFETs... 

[Ben N]

I'll have to pull this one out. I haven't played with it in a long time, but I thought mine had been modded--it really squelched the attack, almost like a slow gear, IIRC.

[matt239]

When you adjust the gain via the "ground leg", higher gain settings (reduction in the ground-leg resistance) result in loss of bass.  In this case, the bass rolls of at around 7hz at minimum compression, and around 153hz at max compression.  I was going to suggest possibly having two ground legs with a little more gain for the treble, but because the same op-amp that passes audio signal is also used to determine the control signal, it is unwise to start tinkering.

Wouldn't a much larger cap prevent roll-off in the audio range with out changing the overall gain? (So the comp knob would still function as expected..)

Couldn't we also design that section for the same gain range without such large resistance values, thus lower noise?

[Fp-www.Tonepad.com]

280 is a nice and simple compressor, it's bad rep comes from DOD's infamous crappy pedals, I've hear so many bad ones.. they used to be ok.

[Mark Hammer]

When you adjust the gain via the "ground leg", higher gain settings (reduction in the ground-leg resistance) result in loss of bass.  In this case, the bass rolls of at around 7hz at minimum compression, and around 153hz at max compression.  I was going to suggest possibly having two ground legs with a little more gain for the treble, but because the same op-amp that passes audio signal is also used to determine the control signal, it is unwise to start tinkering.

Wouldn't a much larger cap prevent roll-off in the audio range with out changing the overall gain? (So the comp knob would still function as expected..)

Couldn't we also design that section for the same gain range without such large resistance values, thus lower noise?

The resistance values are likely selected to complement the anticipated resistance of the optoisolator.  Remember that the LDR is expected to drop down to some given range, and the gain (and especially gain drop) produced in that range depends on the components like R5 and R14.

A person could use a different optoisolator, or even roll their own, but the values of those resistances would need to be adjusted to yield the same degree of gain reduction in response to input level.  And, depending on the degree of change in those components, C2 might need its value adjusted.  Still, though, a pretty flexible design that allows for a wide variety of substitutions.

I'll put in a word here for using a "hi-lo sensitivity range" switch to select between different values of R5.

[matt239]

Actually the noise performance seems pretty ok, so I guess there's no need to fix it.

Francisco's layout (Tonepad) already calls for 1M at R5 instead of 3M, apparently this works ok. - Seems reasonable.

Isn't C2 just there as DC blocking? - We don't need it to function as a filter in the audio range right?

So we could leave the resistors, and just use a very large value for C2, then have no audible filtering right?

(Edit) - To answer my own question: Changing C2 from .047 to .47 effectively keeps filtering out of the audio band for all settings of R4 "compress" pot.
- I think..

[DavidM]

The DOD 280 is indeed a great compressor. I built it with a couple useful mods. HAven't got my modified schemo near so I will try and remember:

1) The 1M resistance in the feedback loop of the op amp: I made it switcheable between 1M and 3M. The latter gives lots more sustain, but will lead to a quite high "quiescent" gain, very noticeable especially when you hit the first note after a longish silence (i.e., when no notes are being played, the gain tends to be quite high, so the first pick or finger stroke will give a very audible compression effect, which then kinda normalizes while you play). Still very very useful.

2) Input and output capacitor: raised them both to 100 nF since I play bass, for extra safety.

3) I added a switcheable resistor (I think it was around 120K) in series with the LDR to have a "softer" action option.

4) Decoupling capacitor between transistors' collectors to ground: I lowered it to 10uF or 22uF, cannot remember now, to have a sharper effect.

5) 47nF capacitor in series with Compression pot and then to ground via resistor: I upped this one to 100 nF so as to have less bass loss when compressing. Probably omitting it and connecting the resistor to Vref instead of ground would not cut off any bass frequencies at all, but this would imply using some cable jumper.

6) LED resistor: I put a minimum safety resistor in there (say 470 ohm, don't remember) and subbed the fixed 4.7 resistor for a 10K pot. This enables to sharpen/soften the attack.

I used matched 2n3904 for the LED driver and a TL072. I rolled my own Vactrol with a 5mm LDR (do not know the specs) and a normal, opaque green LED. All in all I really like this comp, it has now 3 knobs and two switches, it doesn't pump, it is quite transparent and provides me with many options.

Best,

David