Development of a "reference" class D starting point

[classd4sure]

Hi,

I don't recall anyone ever talking about RMS power output since post #1. So, sailing right along with that convention, we can surely say it's a 900W amp, if the original half bridge was a 100W amp into an 8ohm load.

Regards,
Chris

[Konrad]

Interesting tread i must say!
But there is still some minor things to deal with:

I know the importace of swithing fast, but the mosfets are kind of prone to oscillate without gate resistance.

There is some info given: the upper halves seem to bee hotter than lower, as i've read it in the tread.
The upper halve's swithing network have a power suply and its bias current that passes the lover halve output mosfets. so is not happening in the upper halve.
Deriver bias from current sorces may bee better.

The common mode is not fully dealt with, but can bee fedback in the switching networks or at the balanse of the switching from input stage..

There is more work to bee done...

[JohnW]

Konrad,

Oscillation is only a problem when MOSFET’s are used in Linear mode (Class A/B etc) – and not when hard switched as in Class D. Gate resistors only increase switching losses, and have no place in a fast output stages – apart from maybe slowing Ton times.

The “heating” issue has been balanced by the use of Cascade level shifters.

The tiny extra “High Side” Driver current flowing thru the lower devices is insignificant compared to the switching losses / output current, and does not in reality cause a problem.

From what I see Chris and others have been working on, I'm looking forward to seeing the circuit - and hardware results

Cheers,

John

[JohnW]

I don't recall anyone ever talking about RMS power output since post #1.

So, you’re reasonable for all those Mid-system Output Power ratings!

John

[classd4sure]

Originally posted by JohnW


So, you’re reasonable for all those Mid-system Output Power ratings!

John

Hi,

450W RMS more than meets my personal amp goal anyway. If we start talking RMS power now I dont' think anyone will want the half bridge version, which is 100 peak as well, someone should have spoken up about that long before now. Tell ya what I'm flexible... I'll call it a 450W RMS amp.

Konrad,
The comment I made awhile ago about the upper mosfet running hotter was with respect to a stripped down, plastic proto board version which I'd just made for kicks. Since it's not the way to make an amp like this.... I wouldn't project it's failings onto this design in any way, I also certainly would not attempt to produce my current circuit in the same poor fashion, at 450W RMS.. ..We need proper techniques even just for a prototype.

Regards,
Chris

[classd4sure]

Hi,

For all the hardcores who've been following this thread, you can find the full bridge circuit I've been working on here:

UCD Based Fully Differential Full Bridge 450W RMS Amp

Regards,
Chris

[classd4sure]

Here's the little prototype amp I made, as requested, half bridge UCD 15V rails.

It's quite the little performer, kind of noisy on a plastic protoboard but the characteristic of the audio shines through! Say's alot for it I think, considering. I think it would be worth making a PCB for.

The Tip127 is of course extreme overkill ... but I ran low on 2n3906's, so it got the job done. It could be optimised more I'm sure.. but I was more interested with playing it than playing with it.

Here's how this simulates:

1.2volts input @ 20khz ~43Watts peak
DC COMPONENT = 1.437437E-02
TOTAL HARMONIC DISTORTION = 1.009981E-01 PERCENT
___________________
1 volt input @ 20khz
DC COMPONENT = 1.203865E-02
TOTAL HARMONIC DISTORTION = 4.605130E-02 PERCENT
_________________
0.5 volt input @ 20Khz
DC COMPONENT = 1.174075E-02
TOTAL HARMONIC DISTORTION = 2.375277E-02 PERCENT
___________________
0.5Volts input @ 1Khz
DC COMPONENT = 1.128617E-02
TOTAL HARMONIC DISTORTION = 1.332616E-02 PERCENT

So yeah, have fun with it.

Chris

[johnnyx]

Thanks for that.

So this is the softcore version.

I might just get my head round it. I'll try it in my simulator first.

Cheers

[classd4sure]

Hi Johnnyx,

I'll give you some more info to help clarify things for you.

The zeners are 10 volt zeners, again, it's what I had around. The zener on the upper driver is required in order to pre-charge the bootstrap capacitor to ensure it can switch when you first power up the output stage.

The lower zener is required to match the upper and lower driver voltage supplies, since it's a 12 volt regulator used and not a 10 volt one. So you can simplify things there as well.

Now that I think about it.. I didn't need the 12 volt regulator at all, but the original mistake was using the 12 volt regulator and the 10 volt zener on the upper driver = smoked a few mosfets/speakers due to the unequal gate drive, that was a bandaid job, I should have torn out the 12V regulator while I was at it. Think I'll do that tonight.

Aside from that, it's bare bones as it can get, you'll get it.

[Konrad]

Talking abt upper driver

Just an idea maybee?