Protection diode (Zener) on jfet gate?

[lapsteelman]

I have been playing with some JFet preamp circuits lately. I've noticed on some Mosfet circuits (SHO for example) they have the 9.1 volt Zener diode on the gate. But I can't recall seeing this used with Jfet circuits. Are they less sensitive to the static damage?

[PRR]

A MOS-fet gate is a nano-thin layer of glass, which can be fatally punctured with negligible current at rather ordinary voltages.

A J-fet gate is a junction diode, slightly smaller than a 1N4001 rectifier, and very-very hard to kill.

J-FETs in hostile situations (like on-stage where cords get plugged wrong) probably should have 10K resistor between any jack and the Gate, so that large audio signals do no damage.

MOS-FETs *need* that resistor AND a protection diode (such as a Zener). Actually many MOSFETs have reasonable protection diodes inside, but it is hard to be sure, and for our abuses a bigger (external) diode is wise.

The un-protected MOSFET "will" work, UNTIL a static shock. Depending on the situation this may happen in building, first gig, not until next winter (dry air), or maybe never. But since the threat is non-obvious and instant, best to have protection on all MOSFETs.

[GibsonGM]

Paul, is there some 'standard' for that diode?  We see the 9.1v Zener all the time...will an ordinary diode do any good?  Or several in series?  How 'bout a lower value zener?   I've wondered about this question, too, and have always just built 'by rote' rather than understanding...thanks.

[Keppy]

Paul, is there some 'standard' for that diode?  We see the 9.1v Zener all the time...will an ordinary diode do any good?  Or several in series?  How 'bout a lower value zener?   I've wondered about this question, too, and have always just built 'by rote' rather than understanding...thanks.

R.G. has posted a few different times that the glass layer Paul mentioned can withstand about 20v for the MOSFETs we commonly use. In order to avoid changing the sound of the circuit, you want the zener to be larger than any signal the circuit can produce (>9v to be safe), but significantly less than the 20v that would fry the MOSFET. Anything 9-15v is good. 9.1v or 12v is common. Lower values will work but may clip the signal, depending on the circuit.

The other benefit of a zener is it limits the voltage both ways: Forward biased (which does not happen in normal circuit operation but may happen in cases of static discharge) it acts as a normal silicon diode with a .5-.7v voltage drop from source to gate, while reverse biased it limits the voltage from gate to source to whatever its rating is. To use ordinary diodes for this type of protection, you need a string that drops 9-15v gate to source, plus a single diode in the opposite direction.

[GibsonGM]

Sweet, thanks Keppy.   Makes perfect sense.  I considered that it would want to be higher than anything you'd be throwing at the gate, wasn't sure if there was a "rule", and now I know! 

[lapsteelman]

Thank you Paul, always very helpful.

[duck_arse]

seeing as these poor blokes have been typing this up for years and years, I did a collection of threads. there is some interesting reading included.

http://www.diystompboxes.com/smfforum/index.php?topic=106637.0
http://www.diystompboxes.com/smfforum/index.php?topic=100760.0
http://www.diystompboxes.com/smfforum/index.php?topic=97215.0
http://www.diystompboxes.com/smfforum/index.php?topic=62555.0
http://www.diystompboxes.com/smfforum/index.php?topic=82749.0

[GibsonGM]

Thanks, Duck!

[tca]

The Zener diode *does* mess up the quiescent working point of the mosfet.

P.S.
The effect is bias-network dependent.

[PRR]

The diode probably should be larger than the Vto of the MOSFET, to assure the diode never limits "normal" signals. On modern MOSFETs, and certainly in 9V battery work, 5V may be ample.

The 0.6V of an "ordinary" diode is too small. And also works one-way; we need both-way protection. (A "9V Zener" will clip at 9V one way but 0.6V the other way. A 1N4007 will clip 0.6V one way and >1000V the other way.) (*Two* back-back plain diodes will limit 0.6V either way, which may be enough swing for some circuits, not for others. The brain-pain of figuring out the circuit is greater than the cost of a Zener.)

And as said, less than 20V. (Breakdown is usually higher, but we can't be sure, and we never need to throw >20V at a gate in our circuits.)

BTW: the Base-Emitter junction of a Silicon BJT transistor "is" a 7 Volt Zener. Good to know if you don't want to run out and buy a Zener.