pwm compressor ??'s

[wilebee]

How hard would it be construct a pwm circuit with a pic chip and some memoriy that we could design and progam some compression curves. the pwm could control maybe a combination of gain control devices like some jfet's and or a vca to mimic the gain reaction of a varible mu compressor or a variety of Jfet style compressors ?

I'd love some feedback (no pun intended)

Wil

Wilebee

 

[gyraf]

http://www.groupdiy.com/index.php?topic=780

 

[wilebee]

oh yeah I think i do now remember this thread going round. I'm interested in wraping my brain around this . Wil

 

[Svart]

wow, i ahd forgotten about this. ... still interested in working on one with someone.. i have plenty of PWM drive experience..that is if you call blowing power IGBTs for fun experience.. :green:

 

[wilebee]

I'm still trying to digest this schematic. The idea I had was to use the pwm to varing the on/off time on say either a single jfet or a pair on input devices I'm thinking that with pwm shouldn't we be able to kind of talor the rediction charteristics. I think it would take some pretty crazy detector circuit thoughts ??

Wil

 

[Svart]

I still opt for a specialized PWM ic based compressor.. they work much better than any osc circuit i've ever prototyped. You can get near perfect square waves with FET gate driving ICs as buffers. from there you can adjust pulse width by inputting your signal into the error amp or comp inputs through a yet to be determined control circuit..(RMS detector/averager?) I still have interesting thoughts about how an A/D converter will like a chopped signal being fed into as it samples.. you'd better hope that the sampling isn't near (or any multiple of) your chop rate or you will get a lot of weird effects as the sample is taken of a chopped portion... i think you'll get a weird pumping effect as the sampling times overlap the deadtimes of the heavy chop times..(how many times can i say this to make more sense to myself?? :guinness: :guinness: :green: )

maybe we should think about 1/2 attenuation of the signal during chop periods instead of full attenuation? this might be more pleasing to the ears i think...

just my .02$

 

[wilebee]

svart I'd love to see your ideas in print it sounds like you know alot about this topic Wil

Wilebee

 

[Rob Flinn]

For you guys who are looking at this design I just found some info that maybe of use to you in an old book I picked up "The technique of the sound studio" by Alec Nisbett who is/was a BBC engineer.

One modern compressor/limiter reduces the power of a signal by earthing (grounding) it momentarily at intervals of 250KHz. At 6dB above the onset point the selection of a 2:1 compression ratio means that half the power must be remooved from the signal to reduce it by 3dB. It is arranged, therefore, that the control switch spends half of each successive 4 microseconds open & half closed. Such a high rate of switching has no audible effect on the signal, but simply reduces the total power transmitted. A signal 12dB over the threshold has to be reduced by 6dB, so that the switch must be closed fo 3 microseconds of every 4; for signals at or below the threshold it does not operate at all.

I hope this helps with ideas for your PIC program Svart

 

[wilebee]

Hey Rob is there any other jems of wisdom in this article? this is exactly what I was thinking about about with the on off cycling of the gain reduction stage. My question is if you have several sets of fets that are set up for different gain reduction slopes couldn't you consevably dictate which devices were on and off in which order thereby being able to taylor you compression slope and or signature ?

Wil


Wilebee

 

[Rob Flinn]

Wilebee

Unfortunately that was all there was about PWM limiters.

What struck me about this quote was that to have the SAME ratio operating on differing signal levels you have to vary the pulse width.

I`m not quite sure I understand what you`re driving at with multiple FETS. It strikes me that anything you suggest with multiple FET`s could be done by controlling just a single FET.

 

[wilebee]

my though was that like in multi band limiters you have 3 seperate gain reductions stages that are working in different frequency ranges why not look at several different devices for different slope characteristics and figure out what part of slope (operating range) is most desirable and then incorperate pulse width of the different stages. what you are saying makes sense buy just using one device and pulse width control of it. I was just taking it out to the insane idea of making say 3 parallel stages, 3 contol circuits, and 1 master control circuit that translates front pannel controls into preset pulse width data control the three stages. I think I'm going a little over board oh well

Wil

Wilebee

 

[Rob Flinn]

Wilebee

When you talk about slope are you talking about the curve for the individual FET?

If so this doesn`t really enter into the equation as far as I can see, because the PWM circuit is set up so the FET is switched fully off or fully on. i.e the grey area, where the curve of the FET is doesn`t get used.

If you`re talking about some other curve please can you expand on what you`ve said ?

 

[wilebee]

Ahaa :idea:

What I'm thinking isn't possible with pwm unless the on stages are biased to theses indivdual sweet points and then combined so

I've gotta think on this some more

Wil