Very simple compressor. Works well IMO

[Noplasticrobots]

Although the original showed a 1N4148/1N914 diode, but a Schottky or a Ge would be better because they have less voltage loss and make the compression more controlled.

I was going on Brett's suggestion. I wonder if a Zener would work or if the reverse voltage would cause problems?

[markm]

 
I failed to read that part so.....I guess it's okay in there huh?
You'd probably be better off with the Ge in this case I would think.

[Noplasticrobots]

Ok, so I have a few questions about this circuit and compressors in general. I've never built a compressor before and have no experience with them.

I built this thing to spec with a J201 and a 2N700. I used a 1k instead of a 1.2 k resistor. Wired in a 50k pot where the 47k resistor is. I'm using a SG copy and have read about the HB issues.

I can hear some "compression" at extreme ends of the spectrum and this circuit makes my tone controls very active. Is this normal. I don't really hear a whole lot of difference rotating the 1k pot, but I know compresssors can be very subtle. When I have the 47k pot rotated all the way to one end and play a noteI'll get a sort of very quick, almost distorted note and then it cleans up and sustains for a while. Is this the attack/bloom time? I'd really love to hear some samples of this circuit with explanations of the ot settings so I canknow what to listen for.

Also, this thing is very noisy on the treble side of things and I'm picking up radio stations like crazy. The strange thing is I'm picking up American radio stations which never happens; I always pick up Spanish radio. Is this circuit apt to pick up stronger frequenccies like this?

[markm]

I built this thing to spec with a J201 and a 2N700. I used a 1k instead of a 1.2 k resistor. Wired in a 50k pot where the 47k resistor is. I'm using a SG copy and have read about the HB issues.

Did you use my layout for your build?
On the original version, the Comp control wasn't overly noticable either.
I don't know about the new one yet as I haven't built it as of this writing.

[Noplasticrobots]

I'll take a look at your layout Mark, thanks.

[Noplasticrobots]

Alrighty, I checked over your layout and it seems to match the schematic brett posted. I still can't hear much compression in this circuit, but I replaced the 1k pot with a 50k and it will sustain a bent note forever.

This thing is still noisy as hell though, is this a problem with the JFET?

[billings]

Does the noise continue when you start playing, or does it drop off?  If it drops off, then it's probably cool because guitar compressors tend to do that anyway.

[Noplasticrobots]

That's the odd thing, sometimes it seems to stop almost like a noise gate, and sometimes it seems to still bleed through in the background. With all the pot swapping though, I don't know what I'm doing. I'm gonna sit down tomorrow and take some notes and report back.

[Bagge]

I don't know that much about IC's. But I seem to remember reading that LM386 was fairly noisy compared to for instance TL072. I don't know if its true? I am aware that they are two different IC's, but it should be possible to make this circuit with a TL072 and still fairly simple. I still need some basic knowledge about electronics before I'm there myself, but I intend to give it a try at some point. Maybe something between Orange Squeezer and this one.

[brett]

Hi
Hey, I found my old files.  The circuit has two significant differences.  
1.  There is a second 0.1uF cap between the mosfet and the 386.  (This will solve the diode conduction problem that Jack Orman mentioned, because the drain will be close to zero volts)
2.  The "compression" pot has its lug 3 is connected direct to pin 5 of the LM386 (ie before the output cap).  The pot in my old diagram was a 10kB, but I think I also used a 50kB at one time and it worked, too.

Markm: apologies for posting my untested ideas.

Although Markm's layout is generally excellent, it has a "bug" in that the wiper of the compression pot should not be connected to C4.  The wiper only connects to the diode.  If it was built as per the layout, the pot would do very little.

With the new pot connection (to pin 5) there should be lots, lots more compression available.  Note that the compression pot interacts with the volume pot a bit, because the compression is achieved by limiting the volume, so dialling in more compression usually requires a lift in volume. 

Also, the small voltage loss from the diode is now of no concern.  In fact, there is probably an advantage in using two diodes in series, or an LED, to remove some of the DC bias, which is quite large.

Regarding particular JFETs and MOSFETs.  The J201 and a BS170 were what I used first time, and they worked well.  The 2N7000 seems very close to the BS170, so it should be good, too.  Any "normal" mosfet will work, because there's nothing fancy happening here. 

RE: using a TL072 or similar
The circuit is designed for an op-amp that uses a single supply and self-biases.  You could make a regular, dual supply op-amp work, but it would need extra parts.  The LM358 is the main alternative, and it might be better in some respects.  Unfortunately, it has a different pinout to the 386 (it's the same as TL072, 4558, etc), so it's not a case of subbing one for the other.  Two advantage of the LM358 are that batteries will last longer, and it contains a second op-amp that could be used as the input buffer instead of the JFET.

Anyway, I'm off to the beach for 6 days.  No electricity, just folk guitar and rolling surf..
cheers

[markm]

I will address this situation with the layout and correct it ASAP.
Thanks again Brett, I am a compressor fanatic so, I really like playing with differnet comp circuits!
Thanks again for the help and guidance 

[markm]

I have a new UNVERIFIED layout in the gallery here;

http://aronnelson.com/gallery/Acid-Graphics

Check it out and if you build it, please post results here!

[Noplasticrobots]

Thanks for the reply Brett, I'm giving all your updates a shot right now on the ol' breadboard.

[Noplasticrobots]

Okey dokey, here's what I have going on:

Mark, using your layout I couldn't get it to work. I thought I had my original transistor pinout wrong looking at your layout, so I switched it according to yours and nothing happened when I plugged in. I went back to my original transistor pinout and it worked fine. As far as I can tell, the transistors I used (J201 and 2N700) have the same pinouts as the transistors called for in the schematic so maybe you made a mistake in the layout? I don't know for sure though, I can very well be making a mistake...

On my breadboard, here's what's working (following the schematic):

Input cap to Gate of J201 with 1M to ground, 10k to ground from Source of J201 with .01uf cap to Drain of 2N7000 along with a connection to pin 3 of the LM386, and Drain to +9v.

Source of 2N7000 has a 2.2uf cap and a 47k to ground and a 1N34A with the negative side connected towards the Source, Gate is grounded.

I'm using two 50k pots (the closest I have to 10k is 1k...)

Brett, I tried the mods regarding the diodes, I used to ge's in series and an LED. The LED wouldn't do anything to the circuit when I rotated the pot. The two Ge's in series didn't seem to mak an audible difference. Also, when you say the controls are interactive, does that mean at one extreme of both pots, they shut off? If I turn the volume down I obviously get no output, but if I put the Compression setting down, I also get no output.

I also tried flipping around my 2N7000 but to no avail. 

[markm]

Yeah, Looks like I gotta correct the pinout for Q1. 
Will do.

[markm]

I corrected the pinout and built the circuit but.....no-worky here either 

[Bagge]

Where is this supposed to be?

1.  There is a second 0.1uF cap between the mosfet and the 386.  (This will solve the diode conduction problem that Jack Orman mentioned, because the drain will be close to zero volts)

I saw your revised layout and otherwise it follows the schematic as I see it.

[markm]

Where is this supposed to be?

1.  There is a second 0.1uF cap between the mosfet and the 386.  (This will solve the diode conduction problem that Jack Orman mentioned, because the drain will be close to zero volts)

I saw your revised layout and otherwise it follows the schematic as I see it.

No clue!
I am a bit confused about the BS170 as to whether it is oriented correctly.

[Bagge]

I believe BS170 should be turned around. The pinout facing the front should be DGS and source should be connected to ground, right?

[markm]

That's what I was thinking but the symbol in the schematic has me a bit confused. 
I think you're right on the money though Bagge!
I just changed it.