Low saturation voltage output transistors

Mosfet or GaN. Bipolars will have a lot of switching loss, not to mention base current. 2 amps from a SOT23 fet, or a BGA GaN, would be easy.

Sorry, linear amplifier, should have said.

How low are you looking for?

There are packages much better thermally but not much larger than SOT23. I was planning to use this part in a 2 mm sq package with thermal pads. Rds in the teens of mOhms.

There should be plenty of bipolar parts in the same type of packaging. Digikey is better than Google for this sort of search.

Clive,

It would be helpful to know much more about this linear wide-band design spec and load impedance (f) rather than the parts that don't work.

I suspect you are implying ~ 1 Ohm but what L,R values or datasheet.

This then poses problems for AC coupling single supply with a need for > 1mF

What non-linearity specs? THD, asymm. etc.

I suggest you need a list of design goals and specs clearly list first.

Implementation will be multiple choice TBD.

This is the hardest lesson to learn in design. Write good specs 1st. But you can learn by reading them from commercial datasheets.

Depending on your THD requirements, you could go class D on it

That will run cool

Cheers

Klaus

Oh. "Saturation" sounded like switching. But you still might consider mosfets. Even little ones turn on to a few milliohms.

PBSS303NX, PX and friends?

Is there a power opamp with RRO and slightly above average power bandwidth?

Class D isn't completely ruled out as it could be buzzing at low MHz, possibly GaN based. Would have to shop around. Downside is, such chips are often highly specialized, so maybe it only accepts I2S audio, or has internal filtering, or something, or maybe you just don't want to proto with a This is for a low voltage power amp, say 10kHz to 100kHz, maybe 2A peak

The gotcha is "mounting pad for drain 6 cm2"

Sylvia.

What is the topology ? A single ended class-A common emitter stage ?

With that low supply voltage an inductive collector load should help increasing the voltage swing. Either a transformer coupling or a choke to V+ and output voltage from the collector.

Why is that a "gotcha"??? That's the heat sink. Much less expensive than bolting to a hunk of aluminum. How else would you expect to get rid of heat???

You do realize you can put the copper area on the other side of the board in the form of a ground plane, right? A dozen or so vias and you have effective coupling.

Interesting, thanks.

Anything complicated usually has thermal shutdown, which rules it out. This is for 180'C ambient...

Thank, yes I've looked at those. Maybe two in parallel, and their thermal shutdown is optional for a change. It's a higher risk but a simpler solution, so may be worth testing some at higher ambient temps.

-------------------

Yes, gas mark 4. Limited lifetime, but it's not usually 180'C for more than a few hours at a time, but it does take some hours to get there and back.

That rules out most plastic devices. They're usually rated for

They may well be /rated/ for 150'C (or even 85'C) that doesn't necessarily stop them working at 180'C. People who do this work have a jealously-guarded list of parts which are known to operate at high temperatures, often with some limitations. Plastic packages aren't a problem particularly, in fact they may even be better than ceramic in that they can flex just a little.

Nice part, similar to OPA547; pity it's basically 5 volts.

Interesting how they avoid discussing capacitive loads. That's usually a flag that they don't want to reveal something, and want thousands of users to discover it independently at great expense.

Why does TI use such crazy file names?

If a thing like that is hard switching a few amps at low rates, power dissipation will be tolerable. But getting any amount of power out of a package that small is a pain. Thermal spreading resistance of a copper pour, from the tiny power pad, will be a limit on cooling.