Class B distortion

parts cost with labour as unskilled and ill equipped as possible, and using only scrap parts with no data. I've got a simple class A amp, now want a B for battery apps. The point of the exercise is to provide designs that som e people in poverty might learn to build to sell.

What kind of nutty requirement is that?

No, not really; in particular, LM358 and LM324 are anomalously bad because their class B output stages are intended for low quiescent power. It was a deliberate decision to insert a dead band in the output drive.

You really want to find a class AB op amp instead, like TL072 or TL074, if you care about distortion (and there's a bit of a quiescent current penalty). In a pinch, you can put a pullup resistor onto an LM324, and make it operate in class A.

OK here is my results:

distortion at 1k was under 0.5% and at 10k under 5% - I didn't filter out the output above 20kHz so the figures are better than that. Output was about 3 watts, too much for the output transistors for a continuous signal but OK for music and voice.

The VBE shift will be present in the swing of the amp. How you are seeing a VBE type error in the output doesn't make sense. I suspect the problem is your circuit and not the part.

You do know that 175 is beyond mil spec. I had figured this was a well rig project, but now I don't know what to think.

That's almost identical to what I simmed, but distorted. When I get time, which probably won't be this week, I'll try & see what change fixes it. Thank you :)

I had it running on lower voltage, 30v is a big mismatch for an 8 ohm 0.5-3w amp.

NT

Most purchasers won't want raw class B. I'll include the option though for the few that do.

NT

een

d.

gns

I assume so

Maybe I didnt explain well enough upthread. Its not designed to hit 175, bu t if I can pick an option that's bias stable over the full range then its o ne less cause of failure in the real world, where what people do with the d esign is outside of my control. The helper tr circuit, with trs biased by t he opamp's quiescent current, falls down on this point, otherwise it would be a very promising option.

A circuit that works reliably for end users is going to sell far better tha n something that gains a bad rep for frying itself.

NT