Envelope filter diodes

[alanp]

In pedals like the Nautilus, or the Meatball, or that Doctor Duck one (I forget the name, sorry), I normally see a 1N4148 or similar in the envelope detection part of the circuit. Vf for these is, what 0.7volts? Would the performance of this type of circuit improve with a germanium diode in them, since IIRC they have a lower Vf ?

[Thomas_H]

The diodes in an envelope follower are usually there to rectify the signal to generate the envelope. I dont think that it does make a difference which diodes are used.

[midwayfair]

Lower Fv will generate a larger envelope signal.

This sentence is very different from suggesting that it will "improve" performance. There are some situations where the size of the signal source before conversion to DC or some circuitry afterward could require a lower Fv. E.g. if you source the guitar signal directly at the input rather than after a buffer or amplifier (in which case you are using a particularly weak signal source and loading it more heavily) before the envelope or in a situation where the circuit following the envelope blocks a lot of DC before it does its "thing".

Of the top of my head, the first case either would be extremely rare and would just be poor design if it mattered. It sort of occurs in the Nurse Quacky, where the input signal sees the op amp input for both the dry path and the envelope. (This different from Orman's Dr. Quack it's based on, which has an input buffer.) Even then, though, the NQ is capable of boosting the signal enough to use LEDs for the diodes, which lose an entire volt above what a generic silicon loses! Which I suppose just goes to show just how good ICs are at making tiny signals into huge ones ...

The latter case occurs in the Rothwell Lovesqueeze. They set it up so that the source signal for the envelope is already very large, so it can't be amplified too much further without hitting the power rails. But then a bunch of DC is blocked by a 1M/1M voltage divider before the variable resistance element AFTER the envelope conversion to DC. In such a situation, a super low Fv was required of the diodes to keep the envelope signal size effective. (Note that they could have avoided this situation and used a cheaper diode just with a slight change to the post-envelope voltage divider ... that's basically the change I made to that rectifier in the revisions to the Bearhug's version of that envelope circuit.)

So where's a legitimate example of a circuit design that absolutely requires a low Fv in the envelope's diode? The only one that comes immediately to mind is the Orange Squeezer -- the 1N100 has a Fv of ~0.25v and is noticeably better even than a 1N34A (Fv ~ 0.35), and that's only 1/10 of a volt difference. The envelope in that situation is entirely dependent on the size of the amplified signal because it doesn't get amplified again on its own. This is why sometimes people use the (unused) second half of the op amp to create an actually decent envelope circuit in OS derivatives.

If you want to experiment, don't use germanium. Use Schottkys. They're cheaper and more durable.

[alanp]

Thought that might be the case, Thomas and Jon, but if you don't ask you don't know Thanks