Using mosfet as a zener

You'd get more predictable performance by placing a proper zener between drain and gate, with a suitable gate shunt resistor for zener biasing.

The mosfet has better energy handling characteristics when not in avalanche.

RL

"legg" kirjoitti viestissä: snipped-for-privacy@4ax.com...

Yes, the avalanche voltage is not too precise

This is the more interesting part: It is actually just the other way!

I tried that zener (27 V) + resistor (1k) circuit, and it turns into a short circuit at about 50 W dissipation. 50 W at 30 V is way out of SOA Datasheet:

If the gate is grounded, then it can dissipate 100 W @ 43 V no problem I assume that at 100 W dissipation the die in a TO220 package is near max temp.

-ek

I'm unaware of any record of mosfet avalanche testing that would agree with your findings.

Have you scoped voltage and current waveforms during avalanche? Some metering is highly susceptible to the emi that accompanies avalanche or even improperly-biased controlled zenering.

I have one meter that can even be convinced to flash it's low battery indicator, if unproperly screened in such a test set-up. Under no circumstances trust internal shunt readings of meters with long leads.

RL