What could cause a short in a jfet?

[trjones1]

I used amz's splitter/buffer to split my guitar signal 3 ways as part of a clarinot build I did that had one dry and two effect outputs.  Here's the article for the splitter: http://www.muzique.com/lab/splitter.htm

It worked for months before suddenly giving up the ghost.  I tracked the problem down to the splitter/buffer setup and found that the gate, source and drain were all shorted, giving me 9V on all three pins.  I replaced the jfets and the pedal is back to working fine now. 

My question is what could have caused this internal short in the jfets?  I would really like to figure out how to prevent this in the future.  Could one of the jfets have been faulty and it took the others out with it, or is there likely to be some external cause for this problem?

[PRR]

> what could have caused this internal short

High voltage large current input surge.

Plugged into a Speaker Out?

Add "belt and suspenders" resistors:

[DiscoVlad]

Ungrounded (i.e. 2-pin plug) switch-mode wallwart?
The output of those tends to float at half the mains voltage, which is pretty good at zapping circuits especially grounded ones.
To fix this - either plug everything in BEFORE connecting the power, or get an old, inefficient (though much less damaging, I'll take that over so-called "efficiency" any day) Linear power supply, or a switch-mode supply which is actually grounded properly.

I had the same problem with a wah buffer, JFET gate junctions are very fragile. PRR's suggestion of current limiting resistors will certainly help though.

[trjones1]

Thanks for the responses.  Is the voltage spike something a zener diode could take care of?

[deadastronaut]

i blew up a fet the other day.it went bang..and took the whole rounded back completely off it..exposing the inner metal prongs

it was a shock when it happened, but it was funny too...first time for me..and probably not the last...

[PRR]

> something a zener diode could take care of?

JFETs are not easy to kill. Something big hit it.

Since you do not know "what" blew it up, you can't know how big a Zener is needed, or what a big-enough Zener will do to whatever hits it.

OTOH, the 100K at the input protects the Zener up to 1,000V inputs (though long-term 160V inputs will damage a small resistor). 5K at the output will be safe for any speaker-line; if you are jacking into wall-outlets then don't do that.

The resistors won't harm whatever you are hitting it with, and the effect on your signal is too small to care (much better than a dead FET).

The resistors are cheaper than a Zener.