Development of a "reference" class D starting point

[Kenshin]

[i]
I know what you mean about the rising bootstrap assembly adding drive to the base of the gate drive input transistor--regeneration. [/B]

what do you mean by?

the drive is supported by the voltage on the capicitor--when the output is low,it get charged by 12V PSU at the low side through the diode.

the 12V PSU also powers the low side driver.

no requirement for another boosting PSU.

the total circuit(plus low side & output L )is once simulated,but its not on my computer now.

[Kenshin]

I have a guess on why you use inverting drive at the high side:
the high side circuit has a reference to the output stage,so feedback is involved.

using voltage follower:
Even with no any output device connected, when the input voltage becomes high,the MOSFET turns on,and the output going up, and the input voltage is now low compare to the output,the MOSFET tuning off and the output voltage go low...this negative feedback could cause extremely high frequency oscillations.

use inverting drive at the high side,it will be a positive feedback,causing hystresis and quick switching instead of oscillations.

[classd4sure]

Originally posted by Kenshin
classD4sure:

do you mean the circuit you posted before,with a diode across the PNP transistor's emitter? I don't quite catch it. could you explain it to me?

I mean all the discrete drivers discussed.

Anyway,a pull up resistor plus a voltage follower plus the diode is simpler. why not use it?

You seem to understand it fine... why not use it? Simulate it more, you'll see, it is a VERY, VEEEEEEERY weak driver. When I say weak I mean the only way it can work is to achieve a very delicate balancing act between having it so starved for drive that it is just on the verge of not working at all. If it is any more loaded down than that you lose drive voltage to your mosfet (smoke), with any more drive than that.... your drive signals overlap (smoke). So to be honest I think it sucks.

do you want to make a inverting drive at high side,so you need another one?

Inverting? No need. I want a driver with total control over the gate under all conditions, minimal deadtime (non) and zero overlap!

...prevent the base of two following transistor be pull under the output(the source of high side FET).

I'm not sure what you mean?

[subwo1]

Kenshin,

Ah yes, the regeneration, an old-timer's term for positive feedback, especially in switching circuits, snaps the high side on, and your reason for including it in the circuit sounds good.

what do you mean by?

The emitter of the drive input transistor is pulled high along with the MOSFET source by the bootstrap capacitor.

[subwo1]

Chris,

Thanks for the warning about the non-inverting MOSFET driver. I am sure I was lax in forgetting a previous warning I had received about that method.

I think this diode is the one.

[Kenshin]

if not with the diode,the b will be pulled to GND quickly,so the transistor's bc junction will be positively biased.

to make things worse,if not the follower ...the gate pulled to GND...and the source is on high voltage output...smoking...

Originally posted by subwo1
Chris,

Thanks for the warning about the non-inverting MOSFET driver. I am sure I was lax in forgetting a previous warning I had received about that method.

I think this diode is the one.

[classd4sure]

Originally posted by subwo1
Chris,

Thanks for the warning about the non-inverting MOSFET driver. I am sure I was lax in forgetting a previous warning I had received about that method.

I think this diode is the one.

Oh wow, you guys had me right confused there! zzzzzzzzzzzzzzzzzz


I didn't realize you were talking about that circuit again, sorry.

[Kenshin]

It may be happening in the famous L6203 motor drive.It can run WITHOUT boosting. so the gate driver must be not coupled with the output.

2 years before,i'm using it in a robot. when the MCU go wrong,the L6203 broke --even with a small current lamp serialed at the power. The chip first leaked,the motor turning strangely,than the current go up slowly...after 1 minuit...

now imagine,it may be this.

[subwo1]

...The chip first leaked,the motor turning strangely,than the current go up slowly...after 1 minuit...

Indeed, reverse breakdown can cause internal hot spots and then failure.

[phase_accurate]

Regarding the reverse recovery problem: In the near future there will be output devices specially aimed at class-d amps. They are in fact already available though not in small quantities.

Oother example that isn't specifically targeted at class-d is a newer version of the IRF540:

http://www.irf.com/product-info/data...ta/irf540z.pdf

It's body diode has a reverse recovery time in the range of good fast-recovery diodes.

The other solution would indeed be an adaptive dead-time control. I have a Tripath patent somewhere dealing exactly with that but I haven't read it yet.

Regards

Charles