"Pooh Bear" wrote in message
...

Arny Krueger wrote:

"Pooh Bear" wrote in message
...

Yes - noise does drive up the figure but simple examination of the
analyser
output on a scope wil show shedloads of crossover on most 'run of the
mill'
amplifiers.

An experienced eye ( such as mine ) can interpret the waveform rather
well.

Look Graham I was probably looking at crossover distortion from SS before
you were born. ;-)

I rather doubt that actually.

Name a date.

I'll name one: 1962.

Furthermore, crossover distortion shows up like a sore thumb on a
spectrum
analyzer - all the high order garbage.

The 'order' varies with design actually.

Of course, but its there.

You're very mistaken to think it's all high-order.

I never said it was *all* high order. I simply said that the presence of
high order distortion is a common indicator.

The only circuitry I've seen that's effectively free of the effect is a
mosfet
output stage - optimally biased.

I just don't know what srt of amps you've been looking at.

Well........... If you want one of the worst examples I can think of -
think
QSC !

It was obviously a different QSC than the three I've measured.

The RMX is better than the MX but it's still pretty rubbish. I'd expect
the
USAs to be even worse. The 'criminally' deficient biasing method deosn't
help !

Interestingly enough our experience does not overlap. I've only measured
QSCs from the small end of the USA series.

This includes cheap IC chips, cheap receivers, and cheap
autosound amps.

The IC chips are frequently the worst examples.

Please name a readily available modern chip that is offensive. I'll run
right out and buy one and test it.

LM386 !

Fails the criteria of being "modern". My best info suggests that the 386 is
a 1982 design. In fact I've never tested a LM386. I started testing chip
amps with the LM 2002. The LM2002 was far from being fault-free but it was
OK for crossover distortion. This was the late-1980s.

I would consider the LM1875 to be a modern chip.