PWM Compression...

Advertise with us
shop deals on amazon
Shop deals on ebay
GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
T

taliska

Well-known member
Joined
Aug 16, 2013
Messages
136
Location
London, UK.
Hi all,

I don't know whether it's been posted before, but I stumbled across a DIY PWM compressor that I hadn't seen before.

It's from an italian guy who makes plug-ins and some hardware units too. The PWM project seems to have been run in collaboration with a music mag of some sort and the page for it is here:

http://www.sknote.it/MS4_en.htm

I found it because I randomly googled for a sta-level plugin (which this guy also makes) after finding a sta-level schematic online, but with too many unfinished builds as it is, I figured I'd go the plug-in route on the sta. (Won't be building this pwm either as I'm already working up to building the AC Sound pye, but it might be interesting for someone to have a look at for curiosity's sake...).

Cheers,

Kaz

P.S. The sta-level schematic was this one if it's any use to anyone: http://www.audioschematics.dk/downloads/schematics/amplifiers/sta_level.jpg
 
gyraf said:
Unfortunately no schematic or other info...

Jakob E.
Click to expand...

Yeah, I didn't look further in to it myself and unfortunately I don't speak or read Italian.

For those that do, look at:

http://www.compmusic.it/costruisci-il-compressore-ms4-stereo-in-pwm-primi-passi/

(looks like there might be a link to an online copy of the magazine at the right hand side...perhaps that will contain more info?)

And the almost non-existent discussion of the project at:

http://mixincafe.com/index.php/Thread/5776-Compressore-MS4-con-Computer-Music-Studio-Thread-di-supporto-in-italiano/?postID=70088#post70088


As a last resort, could always try e-mailing the guy himself at [email protected].

Of course, who knows whether the juice will be worth the squeeze, but it seems like it might be of interest to those who want both feedback and feedforward...


<edit>
Just googled to try and locate a schematic, but no info, so perhaps this is more of a closed project. Came across a thread on GS of people discussing his plug-ins, and there is a brief mention of the MS4 which suggests it's actually very new...(looks like it was released around May/June...)

Actually...just found the support thread on kvr: (no entries yet, but English at least...)

http://www.kvraudio.com/forum/viewtopic.php?f=49&t=412438&sid=0db833080421d2214b9aca60165d38f9

</edit>

Cheers,

Kaz
 
I did a lot of work on PWM compressor / limiters for broadcast use.  8) PWM as a means of control is very elegant.  As long as the switching frequency is well above the audio range, and you include lowpass filters to eliminate any possibility of aliasing or control breakthrough to the output, the results can be superb.  You aren't at all concerned about the linearity of the control elements and you don't need to match them for stereo use - they are either on or off!  My designs (mostly) used the 4066 CMOS switch IC as the control element.  Each IC has four switches, and these are very accurately matched (within the same IC).

The sidechain was done in the usual way, with buffer amplifiers from the input combined to mono, then feeding a full-wave rectifier and timing circuit using a diode pump.  ;) The derived DC voltage was then used to control the pulse duration of an oscillator. Again, I used a CMOS IC for this function.

The audio path for each channel used a "Superbal" input stage (a nifty way of getting superb CMRR) using an NE5532, driving another NE5532 stage for pre-emphasis, then feeding a pair of Sallen and Key low pass filter stages which turned over at 15 kHz (this was for broadcast, remember).  The Control oscillator ran at 120kHz (well outside the broadcast bandwidth), and the pulse width varied from 3% to 90% (roughly).  The CMOS switches were wired with one in series with the audio and then a second switch to ground.  The switches in each channel were switched with a pair of complementary square waves from the PWM oscillator.  The series switch was "on" when the one to ground was "off", allowing audio through, and vice versa......

The switching stage was followed by a further pair of 15kHz lowpass filters, and then on into a current-pump balanced output stage.

I found that it was necessary to shape the frequency response of the sidechain input amplifiers for best results, and I built a series of these units with BBD "analogue" delay lines before the gain control stages, so that the "attack" phase was eliminated, thereby giving a broadcast limiter that didn't overshoot and didn't require a nasty clipper after it to catch the peaks!  I built about 200 of these devices, and many of them are still in use today, over 25 years later!
 
Mictester said:
I did a lot of work on PWM compressor / limiters for broadcast use.  8) PWM as a means of control is very elegant.  As long as the switching frequency is well above the audio range, and you include lowpass filters to eliminate any possibility of aliasing or control breakthrough to the output, the results can be superb.  You aren't at all concerned about the linearity of the control elements and you don't need to match them for stereo use - they are either on or off!  My designs (mostly) used the 4066 CMOS switch IC as the control element.  Each IC has four switches, and these are very accurately matched (within the same IC).

The sidechain was done in the usual way, with buffer amplifiers from the input combined to mono, then feeding a full-wave rectifier and timing circuit using a diode pump.  ;) The derived DC voltage was then used to control the pulse duration of an oscillator. Again, I used a CMOS IC for this function.

The audio path for each channel used a "Superbal" input stage (a nifty way of getting superb CMRR) using an NE5532, driving another NE5532 stage for pre-emphasis, then feeding a pair of Sallen and Key low pass filter stages which turned over at 15 kHz (this was for broadcast, remember).  The Control oscillator ran at 120kHz (well outside the broadcast bandwidth), and the pulse width varied from 3% to 90% (roughly).  The CMOS switches were wired with one in series with the audio and then a second switch to ground.  The switches in each channel were switched with a pair of complementary square waves from the PWM oscillator.  The series switch was "on" when the one to ground was "off", allowing audio through, and vice versa......

The switching stage was followed by a further pair of 15kHz lowpass filters, and then on into a current-pump balanced output stage.

I found that it was necessary to shape the frequency response of the sidechain input amplifiers for best results, and I built a series of these units with BBD "analogue" delay lines before the gain control stages, so that the "attack" phase was eliminated, thereby giving a broadcast limiter that didn't overshoot and didn't require a nasty clipper after it to catch the peaks!  I built about 200 of these devices, and many of them are still in use today, over 25 years later!
Click to expand...

This is some great info, thanks for registering and telling us about them!
 
Yes, thanks Mictester, fascinating stuff.

Can you/would you share a schematic?  I understand that it might be IP you need to protect.  I'd be interested in seeing how it all goes together.

 

Similar threads

StarTrucker
Scratch 2ch Sta-Level Build
2
Replies
35
Views
2K
StarTrucker
StarTrucker
F
Bogen M330A Tube PA
Replies
14
Views
2K
abbey road d enfer
A
heatwalk
First Time Racking Questions
Replies
14
Views
3K
MidnightArrakis
MidnightArrakis
U
Shure M67; Restoring and Ideas
Replies
5
Views
3K
Scodiddly
Scodiddly
U
Motown Direct Amplifier-inspired Preamp?
17 18 19
Replies
372
Views
66K
MidnightArrakis
MidnightArrakis

Latest posts

Back
Top