P channel class d anyone ?

[Tekko]

Today in lack of anything to do i decided to revisit a cheapo class d topology that i've seen in manuals for flat screen tv's and such.

This is the circuit:


I call it the "Cheapo s**t class d"
Since i dident intend this as a project, i dident put down much time in layouting, so heres a video of what i ended up with:
YouTube - Simple quick and dirty class d amplifier

Its running at around 600kHz, but started out at over 1MHz so i had to slow it down a bit. The fets are IRF9540 and IRF540.

Sound ? Pretty darn good for such a lashup, the low frequency response is wicked, it goes right down to DC!!!

In which thread was the similar circuits posted ? I can't remember.

[savu]

fredos did something similar a couple of years ago

[stocktrader200]

I have been thinking of a P channel high side to replace the
high side driver and N channel fet.
I'll bet the specs are better then average due to your high switching rate.

[Tekko]

Not quite, with a LM311 you cant get below around 0.1% THD due to its internal construction.

Above 400kHz is really just a waste of switching instants, since 300-400kHz is fully enough to produce a 0.01% THD or better full range class d amplifier.

Also, the P channel fet class d is not a good design, it was simply used because high side gate drivers dident exist back then and noone had yet figured out how to drive a high side N channel fet.

Why this kinda circuit is still in use in commercial products is that it dramatically reduces production costs.

[phase_accurate]

Just wanted to remind you that the IR2110 already existed more than 20 years ago ! The current versions have better specs though.

Because I didn't like how it behaved back then I did an amp with complementary output stage as well. It had a DC path parallel to the capaciator coupling. And it wasn't self osciallating. The schematic can be found somewhere on this forum.

Regards

Charles

[Eva]

This circuit demands clean, damped and very low impedance supply rails. Otherwise signal-dependent and load-dependent parasitic turn-on of the wrong MOSFET can arise.

In fact, the power supply is in series with gate drive.

[Tekko]

Yeh its not a good design. I'm not gonna waste time on making any boards for this one, just play with when nothing else to do.

[Sch3mat1c]

It works better from the same rail, low voltage only. That way you don't have to worry about Vgs(max) or DC restoration hacks (as shown).

For most MOSFETs, you need shorting-mode commutation, because Vgs(th) is small and shoot-through occurs when both transistors are switching.

My block-model class D amp seems to be working fairly well, by the way. "Complementary" H bridge output.

Tim

[Tekko]

I have been messing around with it some more and eventually damaged something which caused the amp to temporarily stick to one of the rails above a certain volume, so i decided to short the speaker terminals and see what would happen, volume on max and heavy clipping and rail sticking, whilest still on a heatsink it just kept on going, got hot but nothing else.

I've never before had an amplifier survive full volume into a dead short!

[m-tech]

A year ago I made similar amp ( on single sided PCB ). Sound is quite nice ( for me ), but I never done any measurements.