DIY Audio Projects Forum

Analog compressor, feedback loop
Page 1 of 1

Author:  philtronics [ 27 Jun 2018, 13:32 ]
Post subject:  Analog compressor, feedback loop

Hi everybody, my first post so hope I'm doing it right. I've worked up an audio compressor design using dual matched JFETs in a bridged attenuator, with Very Low harmonic distortion (LTspice says -200dB, lol).

Current perplexation is about the gain of the feedback loop. Vout gets up to about 500mVp-p and the control voltage immediately ramps up, and it snaps into regulation. The effect is Hard Limiting beyond a certain level, and nothing below that point.

Instinctively you might want to reduce the loop gain for a broader transition from zero to near-infinite compression. The "side chain" works by Vout feeding a PNP rectifier, which dumps current into a cap, the voltage of which controls the JFET gates. If you scale back the current, the time domain response suffers (hopelessly, beyond a certain point).

I hope this makes sense! Can't really find anything credible on Google. Maybe it's somewhere in some antique schematics with a variable ratio.

Author:  Geek [ 27 Jun 2018, 21:47 ]
Post subject:  Re: Analog compressor, feedback loop

Check out the manual/schematic for the JBL/Urei/Universal 1176 compressor... particularily "Rev E". It has compensation for the -6dB distortion bump.
(too large to attach here)


Author:  philtronics [ 28 Jun 2018, 17:01 ]
Post subject:  Re: Analog compressor, feedback loop

Very interesting! I was looking at this schematic, it's Rev. N or something.

The compression ratios are set up with different voltage dividers from Vout to the peak-detector amp. I think with my design the transition from zero to infinite compression would still be pretty sharp, just at a different Vout levels. There's another bank of switches that appear to control the DC offset of the peak detector. What does it all mean??? Guess I'll spend the next 80 hours running the simulator.

Author:  philtronics [ 28 Jun 2018, 17:59 ]
Post subject:  Re: Analog compressor, feedback loop

Coolness! It works perfectly, after I changed to a voltage envelope follower (traditional peak detector) instead of the current source charging a cap, where the JFET gate voltage is set only by the feedback loop.

Author:  philtronics [ 08 Jul 2018, 10:53 ]
Post subject:  Re: Analog compressor, feedback loop

Curve time! I can set it up somewhat normally, as in the 1st picture, sort of like the 1176, or for constant max output level as in the 2nd picture. Seems like the 2nd way might be preferable, possibly could delete the Output Level control. Any "thoughts" or recommendations? Thanks!

Author:  Geek [ 09 Jul 2018, 00:12 ]
Post subject:  Re: Analog compressor, feedback loop

It depends on your intended purpose. I see useful curves for a variety of uses.

Author:  philtronics [ 14 Aug 2018, 10:35 ]
Post subject:  Re: Analog compressor, feedback loop

Still chipping away at this, for some reason. Does anybody know how the Gain Reduction panel meter works in this schematic?

The control voltage connects to a FET on the left - is that to convert voltage to dB?

Not to mention, the output AC voltage appears to be powering the meter.

Please advise!


Author:  philtronics [ 18 Aug 2018, 10:29 ]
Post subject:  Re: Analog compressor, feedback loop

My understanding is inching along. That JFET in the meter circuit is nothing fancy, just a buffer and level shifter. Also - the meter can be switched to monitor the output voltage.

In other news... just when you thought that Junction Field Effect Transistors were obsolete - the guys at Linear Systems are coming out with some new FETs for Voltage-Controlled-Resistors! This is on the advise of none other than Bob Dobkin, co-founder and CTO of the former Linear Technology Corporation. The first one will be a dual 2N4391, which is PERFECT for this application.

Sooner or later I'm going to have to actually build this suckah! Will advise...

Page 1 of 1 All times are UTC - 6 hours [ DST ]
Powered by phpBB® Forum Software © phpBB Group