VCA compressor design

[lefthandRED]

Hey all, I am designing and building a VCA controlled compressor for an honours year music technology project.

I'm currently in the research stage and starting to think about building my schematic.

As part of the project I want to also do a case study on a VCA compressor, sort of a reference point for developing my user spec and choosing the topologies for the various subsections in the design (rms detection, attack/release mechanism). Does anyone know of any VCA compressors that have easily accessible schematics and a lot of information on them (a dbx or API maybe?). i've had a search here and on gearslutz but not a great amount of success.

This is my first electronics design/build project, i have the basic background knowledge of electronics and the research skills to learn as I go.

Ideally I'd like a discrete a design as possible, aside from the VCA and op amp ICs.

Any advice is appreciated.


Nick

 

[JohnRoberts]

Well you could even make your own VCA using discrete transistors if your want to pursue a discrete design.

A good source of VCA comp 101 data is THAT website http://www.thatcorp.com/Design_Notes.shtml

I'm sure there are plenty of schematics around if you search a little harder.  The DBX engineers published a number of papers in AES journal over the years about dynamics processing.

There was also some discussion about basic VCA design in the trade press back in the '80s when there were several competing approaches. These days it's THAT or THAT for VCAs, and there's little reason to roll your own RMS circuit when you can buy a cheap off the shelf IC solution that works good enough.

JR

 

[strangeandbouncy]

Hi,



  if you must be discrete, you could always look for some DBX 202(black can), 202c(Gold can), or 2001(silver can) vcas. I believe there are also some discrete offerings from THAT, but I havent seen any. I have built G-SSLs with 2001 and 202c, as well as 2151 and 2181. I personally think that the discrete vcas are worth their weight in gold, but I am sure you will find many that dont. I KNOW which versions get used the most round these parts! THe Class A 2001s . . . . .


  Take a look for schematics of Barth and also Polygram compressors. These used Gold cans, iirc, and I have fond memories of using them in the past. Also, I have a DBX903 that has a gold can. It get used a HELL of a lot! I also have an NTP limiter (179-470) that had discrete vcas in the original version. Mine has a later version using iirc 3 2151 in parallel. Perhps you might be lucky, and find some info on that discrete vca . . .

 

[jdbakker]

Suggested search terms (for the greater Google, not necessarily just here):

"Blackmer gain cell"
"quadrant multiplier" (more specifically: one/two/four quadrant multiplier)
"Gilbert cell"

Possibly useful:
http://www.thatcorp.com/History_of_VCAs.shtml
http://www.groupdiy.com/index.php?topic=21475.0
http://www.groupdiy.com/index.php?topic=40622.0

Hints: at the moment you have a very broad scope, easily worthy of a Masters thesis or three. Spend a week or so researching and then narrow it down. Also, hand-building a VCA is an easy recipe for lots of frustration (as discussed in http://www.groupdiy.com/index.php?topic=40622.msg503352#msg503352). Then again, for a Honors tech project that may be (partly) the point.

Good luck, and do let us know what you come up with.

JD 'learning experience' B.

 

[lefthandRED]

Cheers for all of the help so far, that's a good starting point. I think I'll go with a IC VCA for this one anyway and take it from there, I can easily justify it in the write up. I'll take a look at those sites though, never like doing anything unless I know exactly how everything works.


I'll let you know how i get on in a week or so.



Thanks again, Nick

 

[EvLoutonian]

Hi there.

How did you go with this project in the end?

I partly ask because it's a interesting topic generally, but also because I have an old Australian made compressor by a company called OPTRO which seems to be based on the Gilbert Cell VCA approach.  I'm currently trying to work out more about it!

 

[EvLoutonian]

Another pic (showing internals)
(the unit comprises mic+line inputs, compressor, and 10-band inductor based EQ section)

 

[RuudNL]

I have built many compressors in the past.
In fact, with the (THAT) components available today, it is not too difficult to create a VCA compressor.
But what makes a compressor sound 'good' or 'bad' is mainly the attack/release network.
Some manufacturers use a special circuit with an 'auto' setting, that sounds good on most program material.
It seems that the 'dangerous compressor' (http://www.dangerousmusic.com/products/compressor) uses a very sophisticated timing circuit, that gives an almost inaudible compression on a lot of different signals. They also use the THAT (2181) VCA's in their products.
Any ideas about the principle they are using?

 

[JohnRoberts]

Hi there.

How did you go with this project in the end?

I partly ask because it's a interesting topic generally, but also because I have an old Australian made compressor by a company called OPTRO which seems to be based on the Gilbert Cell VCA approach.  I'm currently trying to work out more about it!

The Gilbert cell was used as the front end to linearize and give the CA3280 IC OTA (operational transconductance amp) a little bit better S/N. It was still not comparable to modern VCAs (like THAT corp makes).

JR