another opto-amp

Watch out for C3 current at start-up....

RL

Small series R or drive belly-to-belly thru an R from output.

Add Zeners across opto transistors, protects them _and_ the gates. ...Jim Thompson

Not a problem. The power supplies will ramp up at a reasonable rate and the opto LEDs can stand a lot of current. Fet currents can't exceed Idss.

That's not needed.

I have seen a Supertex depletion fet pull its open source up, but it apparently doesn't do any harm. The current must be tiny and the gate ESD protection seems to absorb it.

A resistor would work if this is a concern.

I suppose I could test an LND150 and see what it does.

You'd probably have a few nanoamps of leakage in the phototransistor anyway, by the time it got to 20 V C-E.

Cheers

Phil Hobbs

You can instruct the voltages to start from zero in the simulation command edit window, without explicitly modelling their ramp, just as an indicator.

Depending on the model, you can't always count on supply currents being accurate, either.

I'd be wary of parameter spread in the Idss, or 'improvements' in the dmos fab process. This might be a factor in zero crossing or zero signal performance as well.

Any reason why you're just using a unipolar signal source in the model? The whole point behind the expense of bipolar supplies.....

RL

People seem surprised when the Apex amps heat up or pop when driving inductors or capacitors in a test bed, not figuring on the phase shift of load current.

RL

In a cascode with a faster interface, you might be surprised at the voltages that can show up on the power switch's source or emitter terminal, depending on the configuration.

...but no real switching done here, in the analog LF signal app.

RL

I just recently Spice modeled their latest... PA99, featuring +/-1250V supplies. ...Jim Thompson

I have *never* blown out an optoisolator in Spice.

The Supertex parts have been very consistent over the years. Idss is always close to 1.6 mA. But this circuit should work with min/max opto CTR and min/max Idss.

I can't simultaneously post all the test cases!

There's also a sinewave generator that can be connected to the input. Just draw a wire.

They don't protect themselves?

Apex used to do hybrids and monolithic amps, but they are mostly now little surface-mount PCBs. May as well make your own.

Just tested an LND150. The gate zeners at about +-40 volts without apparent harm.

I'm not sure what protection is built into what Apex type, nowadays, but the thermal stress would be the same as driving the resulting load current into a short circuit; supposedly a closely coupled thermal limiter would suffice.

Shouldn't be too hard to model an 'arc', which seemed to be a concern elswhere.

I wish things WOULD blow up in spice (in a realistic manner of course

- and not just erratically in the engine). Would be wonderful to be able to peruse the waveform record at leisure, and to replay them iteratively. Spent an expensive week once, working out of a motel in Fort Worth, getting a single boom to record itself on a scope. Surprised how obvious it all was, to so many instant 'experts', after the fact.

RL

One of the things that's terrible and wonderful about electronic design is how many things can go wrong. There's a simple explanation for that.

The explanation is usually simple; getting one often ain't.

Just because it doesn't go kaboom, doesn't mean it's not wrong.

By the way, all those nice-looking trimmed grass verges and fields on either side of the road in a Fort Worth industrial park - no evidence they'd ever seen a human footprint, on close inspection. Makes you wonder.

RL