Mosfet Querry: Surviving Vgs overvoltage

[Winfield Hill]

Roy McCammon wrote...

DN3535 is more like the impedance I want, but the sucker
leaks a whopping 1mA at 125C. That's easily measured; can't
believe its just a convenient spec because of the measurement
problem.

Of course you've put us at a disadvantage, by not describing the
circuit you're considering. But using a 100mA depletion-mode FET
for GATE protection makes little sense, especially when a stated
alternate is a 1M gate resistor! The LND150 is a 1mA FET, with
Ron = 1k, so two in series can make a solid proper connection to
your power MOSFET's gate, yet still limit the 500V fault-voltage
gate current to 1mA, which can be easily dealt with by a variety
of means, without going beyond your FET's gate-voltage max spec.

See Supertex AN-D11 app note for the 1mA fault-limit concept.

Thanks,
- Win

whill_at_picovolt-dot-com

 

[Roy McCammon]

Of course you've put us at a disadvantage, by not describing the
circuit you're considering. But using a 100mA depletion-mode FET
for GATE protection makes little sense, especially when a stated
alternate is a 1M gate resistor! The LND150 is a 1mA FET, with
Ron = 1k, so two in series can make a solid proper connection to
your power MOSFET's gate, yet still limit the 500V fault-voltage
gate current to 1mA, which can be easily dealt with by a variety
of means, without going beyond your FET's gate-voltage max spec.

Thanks for the input. I was thinking of the Depletion Mode Mosfet
has the main element.

Still, my question is, in light of the IR app note, do you think
the typical power mosfet would survive overvoltage on the gate
if the current were limit to 100 uA or so.

But tell me about zener leakage well below the knee.

 

[Winfield Hill]

Roy McCammon wrote...

Still, my question is, in light of the IR app note, do you think
the typical power mosfet would survive overvoltage on the gate
if the current were limit to 100 uA or so.

The heat-dissipating layer is too thin, I wouldn't do it, period.
Jim said, "No visible effect until the gate abruptly shorts."

But tell me about zener leakage well below the knee.

Basically zero. For all zener diodes above say 10V, which are
totally avalanche diodes, if you're more than two volts below
the avalanche voltage, there's no normal mechanism for current
to flow. Go measure a few parts to evaluate what any "defect"
mechanisms may produce in the way of leakage currents: nearly
zero. Did you say something about 125C, what was that about?

On Semi's datasheet for the 1n5221B series (fig 3, page 6) says
the typical leakage for a 1n5242B (12V zener), measured at 10V,
is 5pA at 25C and 200pA at 125C. It'll likely be even lower.

Thanks,
- Win

whill_at_picovolt-dot-com

 

[Roy McCammon]

Roy McCammon wrote...



The heat-dissipating layer is too thin, I wouldn't do it, period.
Jim said, "No visible effect until the gate abruptly shorts."
thanks



Basically zero. For all zener diodes above say 10V, which are
totally avalanche diodes, if you're more than two volts below
the avalanche voltage, there's no normal mechanism for current
to flow. Go measure a few parts to evaluate what any "defect"
mechanisms may produce in the way of leakage currents: nearly
zero. Did you say something about 125C, what was that about?

just quoting the spec on the par. I don't intend to go to 125C.

On Semi's datasheet for the 1n5221B series (fig 3, page 6) says
the typical leakage for a 1n5242B (12V zener), measured at 10V,
is 5pA at 25C and 200pA at 125C. It'll likely be even lower.

I've always assumed that typical meant "you won't get one
that good, except the sample you use in the breadboard."
If I could a typical one, I wouldn't have any problem.

thanks again for the info.

 

[Roy McCammon]

Roy McCammon wrote...



Of course you've put us at a disadvantage, by not describing the
circuit you're considering. But using a 100mA depletion-mode FET
for GATE protection makes little sense, especially when a stated
alternate is a 1M gate resistor! The LND150 is a 1mA FET, with
Ron = 1k, so two in series can make a solid proper connection to
your power MOSFET's gate, yet still limit the 500V fault-voltage
gate current to 1mA, which can be easily dealt with by a variety
of means, without going beyond your FET's gate-voltage max spec.

OK, I finally stumbled across Supertex app note AN-D11

http://www.supertex.com/pdf/app_notes/AN-D11.pdf

I see what you mean. Its a good idea. It might be just the thing.

mercy buckets

thanks a lot

-Roy