JFET 2n4393 as a voltage controlled resistor

[xambo]

JFET 2n4393

Hi, I am new with JFETs. I want to incorporate a JFET 2N4393 to work as a voltage controlled resistor. I have two questions:
a) can I have it to work as a linear resistor? I learned from somewhere that if I put 1M resistor in the drain and a 1M resistor between drain and gate then the Vgs controls the Rds reasonably linearly. I think this is approximately true?
b) Just checked the resistance between Drain to source without connecting the component into the board and it shows 47Ohm! Is it supposed to be the case?

please let me know your valued comments soon.

 

[Audioguru]

JFET 2n4393

A JFET is turned on as hard as it can when it has no gate bias voltage. Negative gate bias voltage reduces the conduction of an N-channel JFET.

A JFET is non-linear unless the gate voltage has some feedback from the output and the output voltage change is less than about 100mV.

 

[FvM]

Re: JFET 2n4393

Some results have been presented in this thread
The voltage range, that can be considered linear strongly depends on the acceptable distortion level. To my opinion, the linearized FET circuit can achieve good results. It's often used e.g. for audio effects, but it can't hardly compete with dedicated audio variable gain amplifiers and least of all with today's digital solutions.

 

[xambo]

Re: JFET 2n4393

Thanks guys. One more question! I have been using 2N4393- an N-channel FET. It did work quite well as a voltage control resistor up to quite high frequencies, say upto 1MHz. But obviously it kept on deteriorating as it departed the audio domain. It got worse at even higher frequencies making its usability non-existent for my purposes as I wanted to use this as a possible option to replace my POTs at around 30-40MHz.
Have you got any idea if this is a viable option at RF? Can I get away by changing the JFET to something else which can work at such high frequencies?

Have a look at my circuit. I actually used a 6k resistor instead of a 1.5k.

 

[FvM]

Re: JFET 2n4393

You didn't mention the intended resistance and voltage range and didn't specify your linearity requirements, e.g. in terms of acceptable intermodulation.

Generally, the transistor capacitances are the most serious restriction of resistive behaviour. Also the linearization circuit is working as a compensated RC divider in this frequency range. The two transistor capacitances can be expected to match by the symmetric chip design. But also the external circuit capacitances should be symmetric for Cgd versus Cgs.

A better high frequency behaviour could be possibly achieved with lower capacitance FETs. However there's basically a tradeoff between low capacitance and low Rds,on. For higher frequencies, pin-diode variable attenuators or gilbert-cell multipliers are preferable, I think.