3 Level PWM

[dcmdcm]

Does anyone has tried a three level PWM amp?
I'm currently working on one,
Look at the pics atached
Does someone speak spanish?

 

[phase_accurate]

I once made simulations with a tri-state delta-sigma topology that looked very promising. Never tried it in practice though. Yours seems to be a bit similar.

Properly implemented it would have advantages like lower switching frequency, lower RFI levels, improved loop stability, reduced supply pumping and decreased body-diode conduction problems - compared to two-level delta-sigma amps.

Regards

Charles

 

[dcmdcm]

sch

Here's a simple sch of the amp, is based on 2 carriers, they are fed into x5-1, and x5-3.
It's done this way because i'll run 3 ch with the same carriers to sync them.

 

[kASD]

In trilevel PWM either you have 2 out of phase carrier wave to feed the comparators or you have 2 out of phase audio signal to fed.......

In your case it seems your audio signal is common and dual carriers are present but they need to be out of phase with each other.....

 

[dcmdcm]

exactly, i use 2 carriers, out of phase 180º, triangular shape, to lower the frec, but also they could be in phase with a dc offset so they don't overlap, like the AC inverters.

 

[kASD]

Any updates on your project dcmdcm??

 

[dcmdcm]

i'm working in the pcbs, and i need to change the inductor because i'm using normal ferrite cores, but the magnetic fields interacts and cause a small distorsion in the high frecs. But here is quite dificult to find toroidal cores.

Any sugestion for the pcb layout

 

[dcmdcm]

a little pic of the working proyect

 

[Tekko]

T106-2 from micrometals is a good choice.

 

[fernando_g]

My estimado dcmdcm:
Cuando copies un esquematico en forma de un archivo de imagen, utiliza mejor el formato de imagen GIF o aun mejor el TIF, ya que el formato de imagen JPG distorsiona las lineas y caracteres y las hace dificiles de leer.
Aun mas, si en el formato TIF lo conviertes a 1 bit (monocramatico) y le aplicas compresion tipo CCITT grupo 4, consigues una imagen de alta resolucion, con un tamaño muy compacto.
Todas estas opciones estan disponibles en tu editor de imagenes.

English Translation
Dear dcmdcm:
Whenever you copy a schematic as an image file, it is best to employ the GIF or better the TIF image format, as the JPG format distorts lines and characters, and makes them difficult to read.
Even more, if the TIF format is converted to 1 bit (monochromatic), and you apply CCITT group 4 compression, you get a high resolution image of very compact size.
These options are available from your image editor.

 

[kASD]

I once made simulations with a tri-state delta-sigma topology that looked very promising. Never tried it in practice though. Yours seems to be a bit similar.

Properly implemented it would have advantages like lower switching frequency, lower RFI levels, improved loop stability, reduced supply pumping and decreased body-diode conduction problems - compared to two-level delta-sigma amps.

Regards

Charles

Charles,

How can you say that body diode conduction will be reduced in 3-level modulation as compared to 2-level modulation?

regards,
Kanwar

 

[Pafi]

How can you say that body diode conduction will be reduced in 3-level modulation as compared to 2-level modulation?

Agree! Diode conduction problem depends on output topology, not modulation method.

 

[phase_accurate]

The whole was depending on a single-ended three level output stage and the preferred way I would have implemented it would give some relief regarding reverse-recovery.

Regards

Charles

 

[kASD]

Charles,
What would be that preferred way, just like the autoformer [referring to your cross-conduction reduction technique thread]at switching node?


regards,
Kanwar

 

[phase_accurate]

No it would have been a "real" three-level output stage.

Regards

Charles

 

[kASD]

No it would have been a "real" three-level output stage.

Regards

Charles

Charles,

You mean using series schottky or i am missing something?

regards,
Kanwar