Very simple compressor. Works well IMO

[brett]

Hi
For both the JFET and the MOSFET, the source is to ground (via 10k for the JFET). 
The gates are connected to 1M resistor and 3.3uF cap, respectively.
Drains go to V+ via 33k, and to the 386 input for the JFET and MOSFET, respectively.
cheers

[markm]

Okay fellas,
Here's the new Aussie Comp II Layout.
see what ya think;

PNP is in the gallery here for the few that may be feeling brave 
Keep in mind, it is unverified at this point.
http://aronnelson.com/gallery/album15

[brett]

Hi
I did some testing and was disappointed with a few things.  But they are probably fixable.

Sometimes it is possible to hear a "warble" which probably means that the 3.3uF capacitor and 33k release resistor (33k) are too small (RxC = 0.1 seconds, whereas it is 1.5 seconds for the popular Dyna compressor.)  A smaller cap (1uF) and larger resistor (470k) should be a good combination.  (I want to decrease the cap because when driven hard you can already hear a brief delay before compression kicks in.  A smaller cap will charge faster and initiate compression faster.).

This schematic shows the 3.3uF cap and 33k resistor, but it shows a much better compression control:


Thanks for everyone's patience while I get this working.
cheers

[b2thea]

looks nice, gonna give it a try some day..

still a prototype or good enough for building for myself?

[markm]

I'm going to update the layout tonight with the changes made by Brett and will post it up this evening.
PCB layout + PNP transfer and a perfboard layout as well.
Stay tuned folks 

[Noplasticrobots]

Am I right in assuming a non polarized 1uf cap can be used without problems?

[markm]

Okay fellas,
The new and current Layout for this project can be found here;

http://aronnelson.com/gallery/album15

Scroll down in the layouts section if you'd like to try a perf build of this circuit!
This layout is UNVERIFIED for the moment but if anyone wants to give it a whirl, please post a report!

[brett]

Hi

still a prototype or good enough for building for myself?

A working prototype.  It will go through a few changes for sure.

Am I right in assuming a non polarized 1uf cap can be used without problems?

Sure.  Or you could try a 0.47uF/1M  or 0.1uF/10M combination.  Can't say for sure how well they will work.

cheers

[moosapotamus]

Since you no longer uses pin 1 and 8 for controlling gain(it's set at 20), could a transistor or Jfet be used instead of LM386?

I think that's an interesting question, too.
Add a cap and a couple of resistors for the jfet amp and you might not be saving much space on the PCB, tho.

~ Charlie

[markm]

Since you no longer uses pin 1 and 8 for controlling gain(it's set at 20), could a transistor or Jfet be used instead of LM386?

I think that's an interesting question, too.
Add a cap and a couple of resistors for the jfet amp and you might not be saving much space on the PCB, tho.

~ Charlie

I wouldn't really care about the space, I like the idea of a discreet compressor with no LDR.
That's kinda why I was crazy about building the Walco but, could never get it to fire or work correctly 

[mac]

A stupid question,
can a transistor be used instead of the 386?
Looking the last schem the way the 386 is connected is like a bjt with emitter to gnd , a feedback resistor (the mosfet and other parts) a say a collector resistor to vcc.
I know opamps work different but I'm too lazy to read the 386 datasheet   


mac

[Meanderthal]

 OMG, DO NOT count on the Walco sustainer as a usable alternative to this! I have one, it must be the single most noisy device ever devised by mankind besides a niose generator, and the compression is so extreme it makes the old EH soul preacher seem like a true bypass in comparison! And I still have yet to hear of anyone actually succeeding in fixing these problems, although there have been a few ideas on how to go about it...

[markm]

OMG, DO NOT count on the Walco sustainer as a usable alternative to this! I have one, it must be the single most noisy device ever devised by mankind besides a niose generator, and the compression is so extreme it makes the old EH soul preacher seem like a true bypass in comparison! And I still have yet to hear of anyone actually succeeding in fixing these problems, although there have been a few ideas on how to go about it...

No, I wasn't looking at it as an alternative but, I agree with you.
The one I built was garbage.
Sounded GREAT on the tonefrenzy site but, other than that.....JUNK 

[Noplasticrobots]

So I just breadboarded this from the newest schematic and a NP 1uf and 470k to ground as suggested. I'm getting a pretty large gain boost (as I'm sure I should) but as far as sustain I'm not really getting a whole lot.

[markm]

Have you tried the actual schem values?

[Noplasticrobots]

I don't have a polarized 3.3uf cap. I've got a polarized 2.2uf, but I don't know what size resistor to stick in there.

[Noplasticrobots]

Ok, I just tried the .1uf/10M combo and that didn't sound too good to me. I tried the 2.2uf cap with the 33k and it sounded much better to me than the 1uf/470k combo; a lot less gain but I don't hear as much compression. I think my ears need to be trained to hear compression better!

[brett]

Hi
Regarding using a transistor instead of the op-amp: that is a brilliant idea.  The 20dB gain of the op-amp seems a bit too high, anyway.  In theory, both the input buffer and the gain stage could be single JFETs with just a few other components.

I'm getting everything from no compression to huge amounts.  When turned up, the effect sounds more like limiting than compressing - the main effect is on the really large signals.  Unfortunately, I connected my compression pot the wrong way around.  Other may have done this, too. 
The correct orientation (compression increases in a clockwise direction) is achieved by joining lugs 2 and 1(CCW and middle when viewed from the front).  Compression now goes up and output goes down as the pot is turned clockwise (pot resistance going down).   Don't forget that in this circuit the compression/limiting works to cut the large signals, so when compression is maxed (the pot is at 0 ohms) output is very low and the volume needs to be turned up. 

You can easily adjust the amount of compression by playing the guitar about as loud as the maximum output that you want.  Then, slowly increase compression until you hear a volume drop.  Once you've set this "limit", you'll find that no matter how hard you play, you can't get much more volume than this level.  You can also keep compressing and "squish" loud playing down to a low level, but you'll need to increase the volume to compensate for crushing the signal.

I'm using a 1uF cap and a 470k release resistor worked much better for me than the 3.3uF and 33k resistor.  The "wobbles" are gone.

The next thing I'm going to look at is a loss of bass when compression is turned right up.  A quick calculation indicates that the 0.1uF cap between the JFET and MOSFET needs to be more like 1uF to keep the bass in there.  (At R=10k and C=0.1uF, fc=160Hz, which is way too high).  If using a polarised cap, keep the (+) towards the JFET.

What I've got working quite well at the moment is:


cheers

[Noplasticrobots]

Brett, thanks for putting so much time into this, everyday you've got a new version posted!

I've gone over the schematic, changed the pot lugs according to your post and reverted back to the 1uf/470k release resistor and I'm still not getting the amount of compression you're describing. Is this a subtle compressor? I've double checked my pinouts for my transistors and everything appears to be hooked up right. Just in case, here's what I have:

JFET: J201

D: 33k, 0.1uf
S: 10k to ground
G: input 0.1uf, 1M to ground

MOSFET: 2N7000

S: direct to ground
G: 470k/1uf cap in parallel to diode/4.7k to pot
D: to pin 3 of LM386

I'm interested in the transistor version too. A discrete compressor sounds very interesting. Maybe it'd give API a run for their money.

[Barcode80]

Hi
Regarding using a transistor instead of the op-amp: that is a brilliant idea.  The 20dB gain of the op-amp seems a bit too high, anyway.  In theory, both the input buffer and the gain stage could be single JFETs with just a few other components.

I'm getting everything from no compression to huge amounts.  When turned up, the effect sounds more like limiting than compressing - the main effect is on the really large signals.  Unfortunately, I connected my compression pot the wrong way around.  Other may have done this, too. 
The correct orientation (compression increases in a clockwise direction) is achieved by joining lugs 2 and 1(CCW and middle when viewed from the front).  Compression now goes up and output goes down as the pot is turned clockwise (pot resistance going down).   Don't forget that in this circuit the compression/limiting works to cut the large signals, so when compression is maxed (the pot is at 0 ohms) output is very low and the volume needs to be turned up. 

You can easily adjust the amount of compression by playing the guitar about as loud as the maximum output that you want.  Then, slowly increase compression until you hear a volume drop.  Once you've set this "limit", you'll find that no matter how hard you play, you can't get much more volume than this level.  You can also keep compressing and "squish" loud playing down to a low level, but you'll need to increase the volume to compensate for crushing the signal.

I'm using a 1uF cap and a 470k release resistor worked much better for me than the 3.3uF and 33k resistor.  The "wobbles" are gone.

The next thing I'm going to look at is a loss of bass when compression is turned right up.  A quick calculation indicates that the 0.1uF cap between the JFET and MOSFET needs to be more like 1uF to keep the bass in there.  (At R=10k and C=0.1uF, fc=160Hz, which is way too high).  If using a polarised cap, keep the (+) towards the JFET.

What I've got working quite well at the moment is:


cheers

I'm a real newb, so i want to make sure I am reading the schem correctly. it appears that two of the pins of the mosfet are wired together on the left side of the schem. is that correct? am i reading this wrong?