fet for automatic gain control?

Many audio compressor circuits use jfets as voltage dependent resistors. a search for "compressor fet schematic" in google images outputs some references.

In this application, the problem with jfets is the need for a negative control voltage. The board is small with little room for more circuitry. Good thought otherwise.

Hul

asdf wrote:

A MOSFET stops behaving like a resistor once the voltage across it exceeds Vgs - Vth. For low distortion, maybe you want to keep it to tens of mV or something like that, not 30V.

Best regards, Spehro Pefhany

Maybe you can ground the gate of the jfet and use the source to control it. That way, as the source moves positive, the jfet turns off since the gate is at ground.

-Bill

Would a photofet based optocoupler do the trick. Not cheap, but simple otherwise. Fairchild makes a few.

And there's some gimmick where you add 0.5*VDS voltage to VGS to linearize. ...Jim Thompson

Not to mention the substrate diode!

Unfortunately not for a 30V swing. One option Hul would have is LDRs but his name sounds Scandinavian and Europe has outlawed those for most applications.

Cadmium?

What do they do about GaAs? PbSe? PLZT?

I see devices with LCDs being advertised as "mercury-free backlight" and "arsenic-free glass".

Of course the backlights may not be completely benign:-

".. according to California regulations, excessive levels of copper (up to 3892 mg/kg; limit: 2500), Pb (up to 8103 mg/kg; limit: 1000), nickel (up to 4797 mg/kg; limit: 2000), or silver (up to 721 mg/kg; limit: 500) render all except low-intensity yellow LEDs hazardous".

Best regards, Spehro Pefhany

Copper is hazardous? Maybe we can go back to using oxen for our prime energy source.

Silver? No more jewelry?

How come car batteries get a pass? They are huge masses of lead

Rum is, by California law, a toxic substance. It contains *alcohol*

Variants: PMOS, NMOS, DMOS, JFET.

esistence element in an automatic gain control type circuit? Power involved is a bare whisper; worst case voltage is probably 30 volts. There may well be a better remidy than an mosfet, but it looks good at this point.

of mV or something like that, not 30V.

e.

More a circuit design technique than a gimmick.

and it was around before 1997.

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

Too bad it still only works to fractions of a volt, at least if you want a wide range in R(on)/R(off).

Use a single balanced mixer instead (three BJTs).

Tim

At 30V? ...Jim Thompson

That seems to be a lot closer with JFETs than with MOSFETs.

I did some measurements a few years back that showed that 2N7002s kept improving as I cranked the V_GS feedback up from 0.5*V_DS to 1.5*V_DS. See

Cheers

Phil Hobbs

Can you use range switching? Failing that, current-mode is your friend. A BJT diff pair can have an enormous range of gain depending on bias, so you can use one pair as a variable current divider to get the maximum input signal down to something manageable, then maybe another pair to get some gain. If you run the two in cascade, you can get probably 40 dB of gain range easily, and 60 dB if you work at it a bit.

You can improve the linearity by ~20 dB or a bit more by using matched diode-connected transistors on the second diff pair's inputs. (See the LM13700 data sheet.)

Any simple AGC design using elements that respond fast enough to follow your signal (as opposed to light bulbs, thermistors, or CdS photoconductors) is going to cause some distortion.

One low-distortion method that may work in a restricted ambient temperature range is a few cascaded voltage dividers whose shunt legs are NTC thermistors, with external heating applied for control.

Cheers

Phil Hobbs

Pretty soon they'll even outlaw dihydrogen monoxide because of all the environmental hazards and because it contributes to "global warming":

:-)

But seriously, if one of your products contains a GaAs FET, you spell that out in writing as "gallium arsenide" and customs in Europe sees that, be prepared for a major imbroglio.

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

Sure. It'll need unconventional supply voltages, but why not?

If the source can supply a little current (or the bandwidth demand is small), only two transistors would even be needed.

Anyway, FETs are utterly useless here. Curiously no references mention it, I dervied that myself.

Tim