Detectors in compressors and Envelop filters.

[Rob Strand]

Opamp version has far less parts.

But they don't do the same job.   The transistor one is non-linear and gives higher compression ratios for a feedback type compressor/limiter.    When you have a feed forward compressor you can get high compression ratios without the non-linearity:





This one is an interesting one.   The idea is when the amp clips it kicks in a limiter to stop the objectionable sound of clipping.   This is a Fender amp but I believe it is a Peavey invention.   
http://www.4tubes.com/SCHEMATICS/Music-amps/FENDER/bxr300.gif
The first stage has a lot transistors doing the clip detection.   The elaborate time constant circuit is interesting.

[PRR]

The opamp version you show does not compare the signal to a Reference, or filter the lumpy output.

That of course can be done with more parts.

[ElectricDruid]

The two circuits you show aren't entirely equivalent, for the reasons others have said.

But if you just take the rectifier part, I'd still go with the op-amp version. It's known as a "precision rectifier" for a reason. If you want performance, that's the circuit you'd use. The transistor version was significantly cheaper back when op-amps were still a newish idea. Now, it probably doesn't make as much difference as it once did.

T.

[PRR]

> It's known as a "precision rectifier"

But why would we want "precision"? Specifically "precision to very SMALL levels", which is what they mean.

The 0.6V of a naked diode can become "ideal to 0.001V" with NFB. But who limits 0.001V?

In most limiting we just want to know if the signal is "OVER". The DynaComp goes by the predictable 0.6V of a transistor Vbe. Less is unimportant, more is still "over".

Like your local version of football or tennis. The ref calls your balls "in" or "out". The line may be a precision call, but there's no call "97.6 inches out of bounds".

(And yes, the "precision rectifier" is usually consistent with temperature. This is important to missiles and oil exploration. The temp-range of musicians is usually "small".)

[ElectricDruid]

> It's known as a "precision rectifier"

But why would we want "precision"? Specifically "precision to very SMALL levels", which is what they mean.

The 0.6V of a naked diode can become "ideal to 0.001V" with NFB. But who limits 0.001V?

We're not only talking about limiting. Envelope filters and compression were also mentioned in the topic heading

To switch your question around, who wants an envelope follower that's doing really well until the input voltage drops below 0.6V, whereupon it falls off a cliff? That's not a great circuit, unless you sell a fault as a feature and call it a "noise gating action"!?!

I've seen passive rectifier designs that use germanium diodes for exactly this reason - reduced volt drop and better performance for quiet signals. Ok, we probably don't care below a few tens of millivolts, but that's still out of reach for any passive design I know of.

Furthermore, the precision rectifier circuit is easy to set up as a full wave rectifier, so you get reduced ripple over common half-wave designs, and the op-amp output provides buffering and won't get loaded down by whatever follows it in the circuit, unlike the basic RC filter passive version. Personally, I can spare an 8-pin chip to do the envelope following job properly.

[PRR]

> We're not only talking about limiting.

OMG. You are right. I can not read.

Yes, envelope followers may need to be precision down to "zero".