Is this really working?

[lumanauw]

Patent #6844777 or #6937092, are these patents really working in reality (for high quality classD amp)?

Isn't that transformer coupled signal has a certain minimum pulse width to work well?

If it works, we can make high quality, high power classD with only IRFZ44

[classdphile]

For a single advantage such simplicity is not without cost. Why not wrap a quality control loop around it, if it's estable enough pyre can manufacture it and sell it back to us. Or is this better seen as one of those open loop things.

[lumanauw]

If the output always 50%-50% there will be no trouble. But for small/large duty cycle, what will happen? Considering the set-reset of Volt-Second area.

[darkfenriz]

LOL

Since the Switch-mode transformer cannot transform low frequency audio signals, then one should use heavy tube-style transformer.
Then why class-D ??

Authors don't seem familiar with the existance of DC-DC converters (in 2003).

[luka]

Originally posted by darkfenriz
LOL

LOL indeed

[phase_accurate]

It is theoretically possible to build such an amplifier but the technical requirements for the transformers wouldn't give any advantages over ANY other amp topology.

Apart from that this has been discussed in web forums etc so many times that it shouldn't be patentable in the first place.

Regards

Charles

[lumanauw]

In patent #6844777, the key is fig.1 to fig.2. Full bridge classD is being isolated by a transformer. If fig.1 is working at 400khz, then the transformer in fig2 should be built by -2 material (maybe T130-2).
Then the carrier is eliminated by L-C like usual.
Then we can move to fig.3, the push-pull one.
At first glance, it seems to be working.
But when I experiment with transformer working with very narrow (or very wide) duty cycle, the weakness of transformer appear. I cannot fool the Volt-Second product of a quadrant.

[lumanauw]

If this really works, it would be very nice. I can make +/-150V swing amp only with IRFZ44N (don't have to look for quality high voltage mosfet), in push-pull mode (like primary of car SMPS) without the need for level shifter etc Just adjust Np:Ns.

[thespeakerguy]

I tried similar in the lab once. Used a MAX4295 open-loop Class D amp, driving N FETs. It dis work, doing aabout 4 watts into 8ohms at 1kHz or so. Anywhere below that it died, because this becomes close to DC flowing through the core.

With a large enough core it could be improved. it should have some feedback to correct for FET, voltage and xformr non linearities.

An alternative is to reverse the current in the transformer by having active rectifiers. This way no DC in the core.

[lumanauw]

Hi, thespeakerguy,

Anywhere below that it died, because this becomes close to DC flowing through the core.

That's an important restriction

An alternative is to reverse the current in the transformer by having active rectifiers. This way no DC in the core.

Could you tell more about this?