zapping gates

[Jim Thompson]

On Thu, 25 May 2006 18:04:10 -0700, John Larkin

[snip]

For a *really* fast hv edge, use a relay.

John

Mercury-wetted. I used to make fast edge, short-width pulses by
driving a shorted-termination transmission line with a mercury-wetted
relay.

...Jim Thompson

 

[Terry Given]

On Thu, 25 May 2006 18:04:10 -0700, John Larkin

[snip]

For a *really* fast hv edge, use a relay.

John

Mercury-wetted. I used to make fast edge, short-width pulses by
driving a shorted-termination transmission line with a mercury-wetted
relay.

...Jim Thompson

Noted, thanks guys.

with 1uF of capacitance, a 47N60S5 switching fet and an IRFBC40 DUT,
with a Tektronix AFG310 signal generator, it happily ate 30 or so 60-V
gate pulses. A single 90V pulse wrecked it. The edges are a bit slow,
90V in 250ns, but thats cos I've got Rg = 50.

So: if the dynamic abs-fail voltage is lower than Johns measured 80V,
its not by much. I've run out of FETs I want to break (that was a
left-over lying on the bench).


Cheers
Terry

 

[dalai lamah]

Un bel giorno John Larkin digitò:

Prove? I just thought the numbers were interesting. The gate leakage
numbers are cool, too. Who needs a reason?

The numbers are interesting indeed, and you don't have to give any reason,
with "prove" I meant "prove to yourself". My english has its limitations,
please be patient.
I was meaning: why did you take the time to pick some MOSFET and zap them?
Were you just curious or did you think this information will be useful for
you sometimes?

Gosh.

Are you saying that if you will use that component for a serious design,
you are going to do one or more of the above?

1) Willfully use Vgses >30V
2) Not put any protection circuit to avoid Vgses>30V, because "if I'm
lucky, it's going to withstand at least 100V"
3) Choosing this MOSFET instead of another, because "in the single test
I've made, it withstood a greater voltage than the other"

 

[dalai lamah]

Un bel giorno Paul Hovnanian P.E. digitò:

Darn! I thought this was going to be another anti Microsoft rant.

What?

 

[John Larkin]

Un bel giorno John Larkin digitò:


The numbers are interesting indeed, and you don't have to give any reason,
with "prove" I meant "prove to yourself". My english has its limitations,
please be patient.

No, your English is fine.

We're having one of our products, a 17 KW peak power MRI gradient
driver, blow up in the field now and then, so we were investigating
failure mechanisms. It doesn't look like even severe drain dv/dt
transients can couple into the gate enough to damage the fets, so that
ain't it.

John