On 7/28/2011 12:09 AM, RussellMc wrote:
> First suggestion is a speedup capacitor across RB.
> Comment on that and we can go from there.
Played with different values and basically increased the value until=20
there was no appreciable difference in turnoff speed.  Basically right=20
now I have a 0.022  or thereabouts across the 1K base resistor.   And no=20
gate resistor on the big P channel mosfet (to eliminate that variable).

That moved me down into the 2.5uS range, which is about half what it=20
was, and maybe a little better than that... Almost tolerable.

It's late so I'm going to ask the potentially stupid question my mind=20
isn't processing right now very well.. since this produces a negative=20
gate pulse as the 3.3V input goes to zero... is there a 'too big' of=20
value for this that will cause damage to the level shifting transistor?
> FET turnoff is of course passively driven by R1 so it has a crucial
> effect on results. Smlleer R1 should give improved results at the
> expense of dissipation (all of which you know).
>
Just for kicks I played with the voltage divider values.   Reducing the=20
dividers to the point where I have a Pq of around 1/4 of a watt shaved=20
another uS or so off of it.  But I'm not willing to burn a 1/4 of a watt=20
per driver... the existing values (10K top and bottom) with a 'on state'=20
power of just over 1/20th of a watt is bad enough.

I guess I should add that turning the FET off is the only time critical=20
piece - on, well, if it takes a few uS to start, that isn't a big deal.

I keep hoping to find a level shifter which is a bit more 'active' - and=20
can permit use of even higher-value resistors.   So far, about the only=20
options I've found are the utilized circuit, or overly complex (and=20
expensive) high-side driver IC's.

-forrest
--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.