Page 1 of 1

variable gain amplifier using analog optoisolators

Posted: Mon Nov 23, 2020 11:30 am
by optodyne
I´m Horst from Bavaria, been lurking for some time.

My current project is an optical dynamics processor to manage my audio, the main goal is to prevent the level going higher than 0 dBu.
I purchased the fastest optos available, the Tesla 3WK160 which has an almost zero attack and release.
By biasing the LEDs and trimming the on resistance I managed to get a total gain range of 40 dB. Left-right balance isn´t perfect yet but usable.

The control signal from the input feeds a fixed threshold gate/expander circuit. While this already works quite well, the limiter which gets its control signal from the output produces some heavy artifacts. Despite being a beginner I know why this is happening but not how to get rid of it. A friend recommended to set attack and release to 10 ms or higher. This reduces the audible distortions but introduces a lag which prevents the limiter circuit to react in time. As already mentioned, I need a consistent level of 0 dBu at the output.

My question is how to design a sidechain which gives a smooth response yet fast enough to catch those nasty peaks ?

Re: variable gain amplifier using analog optoisolators

Posted: Mon Nov 23, 2020 5:45 pm
by JR.
welcome to the hard part...

Over the decades Dbx (now That corp) have published multiple papers about dynamics processing.

Simply any gain control is a multiplier so the control voltage modulates (multiplies) the signal so the side chain modulations are audible... Too fast of a release time constant actually is visible as distortion on low frequency tones.

Of course it is more complicated than my brief explanation...

Have fun..

JR

Re: variable gain amplifier using analog optoisolators

Posted: Wed Nov 25, 2020 7:19 am
by mediatechnology
I purchased the fastest optos available, the Tesla 3WK160
Hi, thanks for your post.
Do you have a data sheet link for the opto?

Re: variable gain amplifier using analog optoisolators

Posted: Wed Nov 25, 2020 8:06 pm
by optodyne
mediatechnology wrote: Wed Nov 25, 2020 7:19 am
I purchased the fastest optos available, the Tesla 3WK160
Hi, thanks for your post.
Do you have a data sheet link for the opto?
http://www.tesla-blatna.cz/en/products- ... rezistorem

Re: variable gain amplifier using analog optoisolators

Posted: Thu Nov 26, 2020 3:51 pm
by optodyne
some progress:

as already mentioned, I need both fast and slow response, so I decided to split the control voltage and feed those into different RCs which are then buffered and finally summed at the LED driver. Atm there are nine timeconstants and a DC bias which makes sure the LEDs are always on.
The response is not perfect yet but I think it goes in the right direction.
What started as a kinda distortion generator is slowly becoming an envelopeshaping beast.

Any thoughts, is it worth going further ?

Re: variable gain amplifier using analog optoisolators

Posted: Thu Nov 26, 2020 7:53 pm
by mediatechnology
The Tesla "Vactrols" are very interesting.
I hope they come out with a VTL 5C3 replacement.

I would keep going with your project though I'm not sure you'll find an CdS-based photocell to be fast enough for peak limiting.
You might want to consider a hybrid approach using both an Opto and a FET.
The Opto would give you a smooth response ideally suited for average-responding compression and the FET a fast response for peak limiting.

Re: variable gain amplifier using analog optoisolators

Posted: Sun Dec 06, 2020 11:43 pm
by optodyne
mediatechnology wrote: Thu Nov 26, 2020 7:53 pm The Tesla "Vactrols" are very interesting.
I hope they come out with a VTL 5C3 replacement.

I would keep going with your project though I'm not sure you'll find an CdS-based photocell to be fast enough for peak limiting.
You might want to consider a hybrid approach using both an Opto and a FET.
The Opto would give you a smooth response ideally suited for average-responding compression and the FET a fast response for peak limiting.
the mentioned opto is very fast, probably the fastest ever made (on-off 0,2 ms). The driver circuit doesn´t allow the LEDs to go completely off.
Plus the expander/gate section prevents excessive peaks before the audio reaches the output. Atm there are twelve different timeconstants working in parallel and I´m totally happy with the result so far. The compressors attack is in the microsecond range, release is kinda adaptive.