anyone ever try this.... easy envelope sensor!!!!

[Brian Marshall]

not sure if this would work.  i sort of stumbled upon it accidentally.

Ok we start with a jfet set up as a drain follower amplifier.  for argumetns sake, we'll say an mpf102 with a 20k resitor at the drain to the pos supply, and a 5 k resistor, and a 10 uf cap or larger parallel to ground. for arguments sake, we'll just say it is propperly biased, and we have a large gain with some prominent harmonics.  you could base a simple overdrive off of this all by its self.  i breadboarded this a few nights ago, and was driving it with an opamp to increase gain.  I wasnt really impressed with it that much tonally, but at one point i put my scope on the source rather than the drain, and noticed something interesting.  it totally made sense after i though about it, but wasnt something i would have necissarilly thought of without seeing it.

as i played my guitar i saw the trace on my scope move with the amplitude of my playing.  the jfet acts as a voltage divider, and the cap charges and discharges as the voltage changes at the source.

has anyone ever tried to use something like this for an envelope controll voltage.  I dont see why it wouldnt work in theory.  the only thing i can think of is that since jfets dc off set is often different from fet to fet with the same part number it may be really hard to design a circuit arround a specific output voltage.  seems to me you would need at least two trim pots... one to bias the fet, and one to bias the output voltage.

I hate using trimpots if i dont have to, but this seems like it might be a little easier than a lot of other full wave rectifier circuits, and a lot cheaper than an led/ldr.

[gez]

Not too sure what you mean Brian, but caps are routinely used in enveloped followers to smooth out a rectified signal (see Mark's article on the subject).  You DO need to rectify your signal though.

You might be able to achieve this by overdriving a common-source FET amp with a diode/LED at the source instead a resistor (and some sort of limiting resistor plus DC bias on the gate perhaps?).  It would be convoluted though - lot easier to do this the conventional way with an op-amp!

Apologies if I've got the wrong end of the stick, was a little unclear in your post what you meant.

[Tim Escobedo]

Sounds like you may be observing the biasing sag when the FET is driven hard, if I read your post correctly. I've wondered if it could be used in a similar way, since it is a rather smooth drift in proportion with the input signal.  Depending on the circuit, it can be annoying, too. I used that effect in my Jinx circuit, where it does a kind of "soft attack" effect.

[gez]

Brian, did you mean you're taking the reading from the junction of the 5k resistor and 10u cap comming off the drain?  If that's the case then yes the FET would work as a half-wave rectifier if the source were grounded and you took the gate a diode drop above ground (DC bias it) AND used a current limiting resistor to prevent you frying the device.  This is what I was on about when I mentioned using a diode/LED instead of a source resistor, but just grounding the source would give the same result - severe clipping one side, normal amplifying action the other (it would be biased into saturation though, which is what you'd want)

Seeing as you're already using an op-amp in the first place it still seems convoluted.  But sometimes the 'feel' (compression etc) that doing things like this creates makes it worth while.  I recently used CMOS inverters in a precision-diode set-up to get a nice compressed feel (even though the parts count was higher).

[Brian Marshall]

Brian, did you mean you're taking the reading from the junction of the 5k resistor and 10u cap comming off the drain?  If that's the case then yes the FET would work as a half-wave rectifier if the source were grounded and you took the gate a diode drop above ground (DC bias it) AND used a current limiting resistor to prevent you frying the device.  This is what I was on about when I mentioned using a diode/LED instead of a source resistor, but just grounding the source would give the same result - severe clipping one side, normal amplifying action the other (it would be biased into saturation though, which is what you'd want)

Seeing as you're already using an op-amp in the first place it still seems convoluted.  But sometimes the 'feel' (compression etc) that doing things like this creates makes it worth while.  I recently used CMOS inverters in a precision-diode set-up to get a nice compressed feel (even though the parts count was higher).

yes that is what i meant....

i was just mentioning it as an idea.   I'm not sure that i would use it.... the only reason the opamp was being used was to overdrive the jfet... i was actually working on a distortion circuitat the time.  i think that there is one big problem with the idea though, and that is the slow attack.... the faster the attack the more ripple, but i guess to some extent this is always a problem with any envelope follower.  Also i think it would lack a lot of the adjustablity you can get with an opamp circuit.

My point was more that i stumbled upon it while i wasnt looking for it.  Seemed like it might work, so i though i'd mention it.  I'm sure someone could put it to good use.  it certainly could have a really low parts count.