Class D Crest factor?

[eem2am]

Hello,

Can you confirm, that it is physically impossible for a Class D amplifier to be able to draw more than peak_power/2 continuously?

[Cristi]

You ask who? :-D

negative. it is possible, as long as you drive the amp with max. amplitude of the input signal. in fact the amp will draw the output power / eff. i do not comment weather or not can withstand this for long term, this depends on the particular design. imagine that you use a class D amp to drive whatever load needs a continuous wave, sine or other shape, a particular example is a DC-AC inverter, who

[ontoaba]

Originally Posted by Cristi

You ask who? :-D

negative. it is possible,.... who

Most of them has over current protection at its supply. So high current that higher than the current protection treshold just work until capacitor discharged and protection work in.

[iand]

Originally Posted by eem2am

Hello,

Can you confirm, that it is physically impossible for a Class D amplifier to be able to draw more than peak_power/2 continuously?

On an unclipped sinewave, this is obviously the case by definition for any amplifier.

On a squarewave, peak and average powers (and voltages, and currents) are the same.

[Cristi]

Originally Posted by ontoaba

...WHO

I was writing the reply while the forum went to sleep for the daily maintenance. when i pushed submit, just part of the message was posted. can be deleted. anyway i forgot my idea.

[eem2am]

-ok, i see your meaning, though i used to work on SMPS for flat screen TV.

I did loads of thermal tests on the flyback which supplied the Class D audio amplifiers.

I distinctly remeber that at peak audio power, the flyback failed the thermal test by a long way.

However, no managers were concerned.

-Because all they were bothered about was that the flyback could continuously supply peak audio power divided by three....which it could.

Given this fact, on these TV's which are sold the world over, how can we be saying that Class D amplifiers can supply more than peak power divided by two.?

[Saturnus]

Originally Posted by eem2am

-Because all they were bothered about was that the flyback could continuously supply peak audio power divided by three....which it could.

That's a very conservatory specification. If I were in the finance department of this company I would demand why these engineering managers agreed to a too high specification. Continuous output is typically -9/-10dB or 8/10 times lower than the peak output, so 3 times lower is a very conservatory specification and it probably should have been even lower at peak power divided by 4 or even 6, to give at least 50% headroom. 200%+ headroom seem exaggerated.

Naturally that could be tempered by the opposite cost of needed large capacitors but it was something I would definitely address if I was in your finance department.

[eem2am]

OK thanks, in my post (2 back) i accidentally said the Class D was supplying.....of course, i meant the smps (the smps that supplies the Class D
amplifier.)

or rather , that the class D amplifier "draws" power

Anyway, Heatsinked fets heat up in about 2 seconds, so i wouldnt want to de-rate that much.

[Saturnus]

Originally Posted by eem2am

Anyway, Heatsinked fets heat up in about 2 seconds, so i wouldnt want to de-rate that much.

They're only at peak output for a few microseconds or picoseconds at a time though, or roughly 10% of the time as noted above so why should it be a concern?

[Eva]

Detecting a too hot power supply is far easier than detecting a too hot voice coil.
Replacing a shorted diode or transistor is easier than replacing a couple of TV speaker drivers with custom sizes and specs.