THAT (was Fet) Compressors

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
analag said:
Just how good are they?
Extremely vast question.
FET as VCA's have two weak points.
They don't have a large linear operation zone; in fact they are NOT linear, they are exponential in nature so a significant part of the trick of using them as gain control element is to restrict the signal. Obviously, doing so exacerbates noise problems, so the amplifier must have very low noise.
The control voltage feedthrough is quite high, which means that some clever circuitry must be added, that mixes a fraction of the signal with the CV, all this while not impairing the timing characteristics of the rectifier.
In both these respects, they are not much different than the diode-bridge types used in the 2254.
And in addition, FET's must be selected; pairing them for stereo operation can be a very tedious task.
The UREI 1176 adds some trickery by putting the FET in a bridge, which allows somewhat higher gain-reduction for a given CV.

On a different POV, FET compressors are another flavour in the arsenal of studio gear, less predictable than VCA types, different than optos both in timing and colour (distortion) aspects. They are definitely far from neutral; for that reason, they are not one-fits-all and may not be really suitable on some sources.
If I had to have only one compressor, I would take a VCA type, but when you want choice of taste and flavour, one should have an FET type and an opto type at least.

And as always, more than half the performance is in the design, not only in the components.
 
I like their warm, vintage, punchy, tone for kick, bass, guitar, vocals, kazoo, in my API, NEVE, SSL, Studio.  FETs Domino.

How good is anything really?  Good to a tracking/mix engineer may be entirely appalling to a mastering engineer.  

FET comps that I have used tend to be fast and aggressive.  Which isn't always welcome, but always seems to have a place.  If you're asking how good a FET serves as a gain control element, than I say probably not that great.  They need input attenuation and a good chunk of makeup gain.  Maybe not the most efficient setup from a design perspective.  Like a gal with fantastic breasts and a good looking face, coupled with hefty emotional needs and bad breath.  Maybe a bit much to deal with for some boobs, but maybe you really like boobs.
 
analag said:
Just how good are they?

As with anything it depends on your definition of "good" and the actual design execution.

If "good" means crude, inexpensive compression for guitar effects, with relaxed demand for low distortion, they can be very good.

They can also be linearized somewhat with circuit tricks (like adding a fraction of the drain signal back into the gate), and interesting topologies to keep signal amplitude small or constant.

About the cleanest jFET gain element design I've seen was a low distortion agc loop for a sine wave generator designed by Robert Cordell (Published in Audio magazine years ago as part of his distortion analyzer), but he took advantage of the constant level output target to further optimize his performance.

In a general audio application, if the distortion caused by the gain manipulations are large and acceptable, the jFET nonlinearity can probably be manageable. For very high performance audio designs where gain change modulations are managed, I don't expect jFET to be truly competitive for linearity and dynamic range with SOTA VCAs.

Another factor working against jFET gain elements is the control law. For a gain application where you want to develop a predictable amount of gain open loop, these can be challenging. Back in the '70s I did some bench work with dual jFETs using one half as a control to linearize the other half in a predictable way. I abandoned this as not competitive with the lower performance than today VCAs available back then. Today we have lower noise jFETs, but we also have better VCAs.

So as usual the answer is it depends...

JR

PS: I've used the trick of adding some signal back into the gate, to reduce distortion in some low cost jFET based on/off gates. While the gates where basically on/off, if the switching between on/off is done slowly you can hear the distortion rise during the transition between on/off. Adding in some signal (-6dB) to the gate, reduced the distortion below being clearly audible.
 
Thank you gentlemen, I'm playing around with an idea. If I can get it right I'll post it.  
 
analag said:
Thank you gentlemen, I'm playing around with an idea. If I can get it right I'll post it.  

Got it  ;D

CompressorC.jpg


Five monolithic IC's and an FET with +\- 18V supply. No bells and whistles yet, maybe never.
 
Kingston said:
This has gone too far. You guys all have a compressor Fetishism.
Hahaha... ;D
MikoKensington said:
I like that you're getting somewhere.  When you're ready to celebrate its completion, I'll bring the con-FET-ti.  
Hehehe... ;D

Creativity abounds here.
 

Latest posts

Back
Top