Level shifter /opamp voltage booster?

OK I've done something like the second high side one... But then everything gets referenced off of the positive voltage... I guess that's where Tim W.'s voltage dividers come in.

George H.

and

mp.)

t the

p.

op-

he

Yeah in this case 100C should be the max... But I've used a TIP120 (to-220 pac) for days at 120 C and hours at 150C. (yeah the data sheet says I shouldn't do that.)

George H.

Just the opposite, HFE/beta rises with temperature in... a _bipolar_ device ...Jim Thompson

I'm pretty sure that at red hot HFE would be zero.

NT

Those two pics just illustrate the proper way to draw mosfets.

Your best bet for the heater PWM is probably just one n-fet and one gate resistor.

Depends on your definition of red >:-}

In the early days of my development of a chip-based alternator regulator I did get back a field failure where the chip had melted into a sphere shape ;-) ...Jim Thompson

IBM would use insulators between the heatsink and chip for cool chips and heater chips for cool modules for higher HFE (and Ft).

You can do that with just a gate series resistor.

Jon

I've had a few get to yellow hot! They didn't last very long, and I was damn lucky that the fragments missed me!

Jon

Eek. Someone should teach you about safe operating areas ;)

NT

SOA is fine, but when some component fails, the SOA can be exceeded very suddenly, expecially with 350 V DC available. I took to testing with ear protectors and a switch under the bench. I'd crouch down there and turn on the switch, if there was no bang I'd then pop up for a look at the scope. I only had a couple explosions before I figured out the problem.

Jon

I used to work with a guy that had worked at place where they made big VFDs he said they could be bit "jittery" with loud noises :)

no kidding :) Carbon resistors are more fun than silicon IMHO.

NT

ar

Carbon resistors are banned in instrinsically safe electronics.

The negative temperature coefficient of resistance of carbon allows the for mation of "hot channels" when they overload - I've seen a 10k carbon film r esistor exhibit a resistance of less than an ohm while carrying about half an amp for a few minutes. The resistor still measured as 10k+/-1% after th e experiment, but there was in thin dark line along the resistor surface ab ove the path of the hot channel.

This was a party trick to illustrate the problem, but it was a very persuas ive party trick.

Years later, I saw the same effect at work in NTC thermistors - use an mill iamp to measure their resistance and the resistance varied slightly from mo ment to moment as the warm channel moved around. Dropped the test current t o test current to 100uA and the resistance was dead stable (to five decimal digits).