Very simple 20:1 compressor (sustain circuit) - no optics..

DrVilches · Post by **DrVilches** » 08 Dec 2008, 11:50

This is a simple compressor that will boost small voltage inputs in an audio BW of around 20kHz by upto 20x and attenuate larger inputs to a maximum of 1x gain.

Distortion is remarkably low for such a simple circuit, which is nothing more than a highZ buffer (first 2N2222), linear amplifier (BC337) and rectifier based around any two Si diodes (use low VF diodes for reduced output / greater sustain) and second 2N2222 device.

Schematic, PSPICE simulations and PCB artwork for use with 1206 SMD resistors (I prefer them..) included.

Feel free to replace any resistor with a pot to vary the effect, etc.

Keep playing and MERRY XMAS

Dr. Vilches.

PS. built and tested, ready for my daughter's soon to be delivered shiny new red and white guitar (not pink, OK Boo?!).

RnFR · Post by **RnFR** » 08 Dec 2008, 14:18

interesting simple circuit. a couple things-

would it be possible to label the resistors on the PCB layout? what resistors would you vary, and for what effect?

thanks Dr.!

DrVilches · Post by **DrVilches** » 08 Dec 2008, 15:28

Sure, I've added another pic with resistor values to this.

The circuit works by loading the amplifier (middle transistor and 500Ohm load) with the impedance of the third transistor (RHS device), its impedance being a function of the amount of base bias it receives from the rectifier network (higher input signal = greater DC bias = lower 3rd transistor's collector to emitter impedance).

Making the collector resistor (currently 500Ohms) variable will allow the voltage gain to be varied somewhat.
Making the emitter resistor (currently 70 Ohms) variable, will allow the collector current to be varied and hence the gain also.

Making either value too large will reduce the transistor's VCE to the point where it clips the output signal, causing distortion. So I guess one could easily turn this into a variable sustain + distortion effect.

Making the 100 Ohm resistor into the base of the 3rd device variable, will allow the amount of bias received by that device and the time taken for the bias to rise in response to an input signal to be varied too.

I guess, at the very least, one should add a foot-switch to the battery terminal and use a 500Ohm pot to tap the output in order to avoid over-driving successive stages but I built and tested it as is and the sustain achieved is quite remarkable (many seconds) with a sound that is quite bright and clean.

Build it on a piece of vero with a 741 wiredup as a voltage follower (unit gain) and wire a set of stereoheadphones across its output (to get a 60 Ohm load at least) and you will have one unbelievably loud mono-headphone practice rig that no-one else (blessedly) can hear..

soulsonic · Post by **soulsonic** » 09 Dec 2008, 05:18

This looks fun. I will experiment with it.
Thanks for sharing!!!

RnFR · Post by **RnFR** » 09 Dec 2008, 14:26

thanks for the description! helped a lot. there shouldn't be a problem subbing standing resistors in for the SM components as well. if need be, the solder pads could be increased in size.

so it sounds like by varying the gain of the first transistor, you could have a threshold control, varying the 100 ohm resistor will give you an attack control, and a pot on the end and you'll have a level control. that's about all you'd need unless of course you want to add some distortion as mentioned above.

thanks again, i'll have to give this one a try sometime soon.

intricatecode · Post by **intricatecode** » 11 Dec 2008, 01:20

is it possible do to a fet version of this using say j201 instead of those transistors?

borislavgajic · Post by **borislavgajic** » 11 Dec 2008, 09:40

Very nice...thanks for sharing

could it be done vith tubes???

boris

DrVilches · Post by **DrVilches** » 11 Dec 2008, 10:09

Both tubes and FETs can be used if adequate biasing is provided. Tubes are not something I have ever dealt with but FETs typically will have less gain and a higher VDS requirement than their BJT equivalents. BTW, most SS trannys will work but use the 2n2222 (or something very similar) for the HighZ input buffer for there it outperformed a number of other BJTs during simulation. Then again, a FET would work better than the 2n2222 there (higher input impedance possible) yet would be worse at amplifying the smallest signals (middle stage).

The beauty of it as it is, is that those BJTs (and equivalents) are commonly and cheaply available and the thing can be built for pennies (or free, if you happen to have electronic components at work..). The greatest impact on dynamic range is likely to come from the diodes used - try to get some very low VF at low current types like the BAT54 Phillips devices: http://www.konkoly.hu/staff/racz/Szider ... BAT54S.pdf.

The idea is to reproduce the input signal as closely as possible whilst sustaining its amplitude. Also, the use of discrete devices instead of ICs means that the circuit will typically be less noisy and that is something that is quite noticeable when some compander chips are used.

Silent Fly · Post by **Silent Fly** » 11 Dec 2008, 12:00

Interesting circuit DrVilches - thanks for sharing!

intricatecode · Post by **intricatecode** » 11 Dec 2008, 19:20

DrVilches wrote:Both tubes and FETs can be used if adequate biasing is provided. Tubes are not something I have ever dealt with but FETs typically will have less gain and a higher VDS requirement than their BJT equivalents. BTW, most SS trannys will work but use the 2n2222 (or something very similar) for the HighZ input buffer for there it outperformed a number of other BJTs during simulation. Then again, a FET would work better than the 2n2222 there (higher input impedance possible) yet would be worse at amplifying the smallest signals (middle stage).

The beauty of it as it is, is that those BJTs (and equivalents) are commonly and cheaply available and the thing can be built for pennies (or free, if you happen to have electronic components at work..). The greatest impact on dynamic range is likely to come from the diodes used - try to get some very low VF at low current types like the BAT54 Phillips devices: http://www.konkoly.hu/staff/racz/Szider ... BAT54S.pdf.

The idea is to reproduce the input signal as closely as possible whilst sustaining its amplitude. Also, the use of discrete devices instead of ICs means that the circuit will typically be less noisy and that is something that is quite noticeable when some compander chips are used.

Cool

the thing is i have a bunch of j201 lying around so i guess when i have time i'll try and work out a jfet version and post it here.

Might even try a valve version with a jfet input buffer and a 12au7

intricatecode · Post by **intricatecode** » 10 Jan 2009, 18:01

would something like this work?

if you can see the image go here http://picasaweb.google.com/lh/photo/HT ... directlink

I just substituted the buffer section with a jfet on using a j201 and the rest of the schematic i used the two halves of a 12au7 instead of the transistors. I haven't tested any of this, i was just thinking about what this circuit would look like using a tube.

Since i already use a 18v power suplly for the dual valvecaster i use on my pedalboard if i ever build this i would run it at 18v too

floris · Post by **floris** » 11 Jan 2009, 11:12

Interesting design, thanks!
It looks to me that the high input pic shows distortion on the positive tops of the sine wave:
https://www.freestompboxes.org/download/ ... &mode=view

intricatecode · Post by **intricatecode** » 11 Jan 2009, 14:31

intricatecode wrote:would something like this work?

if you can see the image go here http://picasaweb.google.com/lh/photo/HT ... directlink

I just substituted the buffer section with a jfet on using a j201 and the rest of the schematic i used the two halves of a 12au7 instead of the transistors. I haven't tested any of this, i was just thinking about what this circuit would look like using a tube.

Since i already use a 18v power suplly for the dual valvecaster i use on my pedalboard if i ever build this i would run it at 18v too

hum sorry i guess i posted the image wrong tell me now if you can see it:

or go to http://picasaweb.google.com/lh/photo/HT ... directlink

I'd really like some feedback on this

mikebike · Post by **mikebike** » 13 Jan 2009, 03:10

a couple questions about you schematic.

is e ohms? as in 22e6 @ r2 and r3?

also what are the grey switch symbols with the v's next to them?

also what is v3?

thanks!

RnFR · Post by **RnFR** » 13 Jan 2009, 03:53

v3 is the signal source.

i think the v's are places where he was checking voltage in the sim program.

not sure what 22e6 is. e is used for ohms though. 22K6? that seems like a weird value to me. it looks like it's just a voltage divider, so maybe that's why he didn't label that on the PCB layout either. there are a number of values that will work right? not sure about that one myself.

i've got a couple of questions myself. it's obvious where the input is. but is the input cap SMD?

also regarding the output. on the PCB layout there is a 10uF cap for the output capacitor. this doesn't correspond with what i am seeing on the schem.

Dr.? are you in?

DrVilches · Post by **DrVilches** » 13 Jan 2009, 10:38

Hi all - sorry, should have checked back here earlier..

2.2e6 is simply 2.2 Mega Ohms (2.2e^6). Spice accepts the exponential term directly.

You can use any type of input or output capacitor, so long as the input is around 100nF and the output around 10uF.
In practice, make R5 a variable resistor and tap the output from its middle connection using one end of the 10uF capacitor.
That gives you a volume control.

The grey pointers are PSPICE probes indicating what nodes where simulated and displayed in the graphs.

Sorry I cant help with the valve idea as I have no experience with that technology.

And yes, the top end of the sine wave out is slightly distorted (not that its noticeable in practice) but that can be remedied by increasing VCE and adjusting values accordingly (I wanted to run it off a 9V supply) and/or using lower VF diodes.

Happy New Year

RnFR · Post by **RnFR** » 14 Jan 2009, 04:42

alright! thanks a lot for the clarification.

mikebike · Post by **mikebike** » 16 Jan 2009, 21:49

just to clarify:

the out put is at the inter section of c5 and the emmietr of q5?

so if i used a pot for volume pin 1 would be at that intersection, pin 2 to the jack and pin 3 to ground?

floris · Post by **floris** » 16 Jan 2009, 23:11

I think the output is at the collector of Q4, meeting with R5 en C2.
... where it says "out" ...

mikebike · Post by **mikebike** » 17 Jan 2009, 07:19

it seems that obvious but look were the "in" is. and thats not afterr a coupling cap...