JFET as a switch for 48 volts!

[valvedoctor]

Hello everyone,

I'm looking for suggestions for a JFET that would be suitable for a switch for over 50V. I'm getting small transients in a current valve pre-amplifier design due to relays engaging on primary of the input transformer. These are for impedance selection and 48V phantom power. These transients are amplified by the pre-amplifier.

So to start, I'd like to find a JFET for switching in the +48V from Drain to Source. I'd like to be a low ON resistance so I can use it elsewhere. I'm not into SS design as much as tubes, so I'd like some suggestions!

I then want to experiment with 4 JFETs in a DPDT arrangement to replace the DPDT relay. I've looked into SSRs but I've only been able to find DPST, and buying two is more costly than the relay!

[Benta]

Why JFET? For low-ohmic switching a MOSFET would probably be a better choice.

[valvedoctor]

No discrimination against MOSFETs. Some similar circuits I've looked at use JFET, but I'm open.

[Kleinstein]

I have not seen many higher voltage JFETs - and the ones I have seen so far are rather expensive (GaN / SiC based ones). I would guess MOSFETs might be the better choice, especially if there is a defined polarity.

There are a few 50 V Silicon JFETs, but one usually needs to add the turn off voltage to the signal voltage, thus 50 V JFETs are likely no enough for a 48 V signal.

[David Hess]

They used to make high voltage JFETs which might have been suitable but no longer.  Teledyne Semiconductor used them in their solid state FETRON tube replacements.

Depletion mode MOSFETs could be used and may be the easiest solution.  Enhancement mode MOSFETs would be less expensive.

[Gyro]

Just a couple of random thoughts...

If the relay is working for you in other respects then maybe you can quieten it by slowing the coil drive. Try an RC network and screen the coil wiring, make sure you have a snubber across the coil too. You could also look at magnetically screening the relay, if that is actually the problem.

A MOSFET will have gate capacitance too, this will couple in switching spikes too unless you take similar precautions.

On balance, I'd work on improving your existing solution first, it provides the best galvanic isolation of the signal.

[danadak]

https://www.onsemi.com/pub/Collateral/2N7000-D.PDF

Just a thought, they are cheap.


Regards, Dana.

[valvedoctor]

Thanks everyone for the help.

danadak, I'm looking into the 2N7000 MOSFET.

Whats happening is there is a transient created when phantom power +48 is applied to the input XLR of a mic preamplifier.
A SPDT relay switches between +48V and circuit common, through the 2 6.8K resistors. Most designs I've seen just use a SPST switch, and the input XLR floats.

Now I'd like to use the 2N7000 as a series switch, and apply a small gate voltage to control it.

Does this sound plausible?

[Ian.M]

Probably not, as the gate charge has to go somewhere whenever it switches and will cause current injection, mostly to the source, but some to the drain.  Also the body diode means that unless you use two back-to-back (in series, source to source), with a pulldown on the source to pass the gate return current, you'll have problems switching AC signals.

Look at Photo-MOSFET optocouplers.  The control circuit is fully isolated, they don't inject charge, and typically already have a back-to-back MOSFET pair internally with no need for a pulldown.

[macboy]

Maybe the transient is caused by the break-before-make operation of your relay. You are floating the line briefly during the change.

Try using two relays, one with 6.8 k resistors to +48 V and the other with 6.8 k resistors to common. Switch the alternate one in briefly before switching the other one out. Maybe even add a small (some 10's - 100's pF) capacitor to ground on the connector side of the resistors. (not too high as you don't want to load the mic output). It might also help to have cable connected to add its capacitance which will limit the rate of change of voltage via the series resistors.

[darrellg]

Whats happening is there is a transient created when phantom power +48 is applied to the input XLR of a mic preamplifier.

Are you sure? It's common to mute the channel prior to engaging phantom power, as condensor mics pop when they are powered up regardless of the switching method.

[David Hess]

I was thinking it was a problem to be solved with precharging before switching.  Besides muting, I might try controlling the transition time of the applied phantom power but that is going to be more complicated than muting.