A new kind of Class D amp ?

[Jocko Homo]

But if you have enough $$$$$, in theory, you could buy one of these:

http://www.diyaudio.com/forums/showt...502#post403502

Jocko

[phase_accurate]

I once saw a proposal in a book (by Ben Duncan ??) that was called ampliverter. This was basically an SMPS capable of positive and negative putput voltages, by using some sort of selectable rectifier on the secondary of the transformer. That way it was possible to use an SMPS transformer of the usually small size.
Keep in mind that if you had to transform the output signal right after an ordinary class-d amp (running at a high voltage) you'd have to use an ordinary sized audio transformer.

Another topology is used by Peavey. They use two SMPS that are running a the same frequency, both at a constant duty-cycle of 50%. One of the "carriers" is modulated in phase, proportionally to the input voltage, and the secondaries of the transformers are summed by synchronous rectification. After the rectification you have a PWM signal which can then be low-pass filtered in the usual way.

http://l2.espacenet.com/espacenet/vi...h&LG=de&DB=EPD


Regards

Charles

[Jaka Racman]

A transformerless amp would never be approved in Europe and and in 230 V country.

Looks that B&O has different opinion. Look at this patent application http://appft1.uspto.gov/netacgi/nph-...l&r=0&f=S&l=50.
IMHO it would be posssible provided:
it is mounted in the loudspeaker cabinet
has optical input
non flammable speaker membrane and coil
protectively earthed non detachable metal speaker grill

Best regards,

Jaka Racman

[Jaka Racman]

Another topology is used by Peavey. They use two SMPS that are running a the same frequency, both at a constant duty-cycle of 50%. One of the "carriers" is modulated in phase, proportionally to the input voltage, and the secondaries of the transformers are summed by synchronous rectification. After the rectification you have a PWM signal which can then be low-pass filtered in the usual way.

Hi Charles,

your description is not entirely accurate. There is one primary oscillator runing at 50% duty cycle. Secondary is demodulated by bidirectional switches run at 50% duty cycle and phase shifted towards primary. I have extensively simulated topology in 1999 not knowing it has already been patented . It has some interesting properties like doubling of filter ripple frequency and zero voltage switching for some of the switches. The only difference with the patented circuit was that i used back to back (source and gate tied together) connected mosfets for bidirectional switches instead of mosfet inside diode bridge as Peavey uses.
Do you know if the patented circuit is actually used in a product?


Best regards,

Jaka Racman

[analogspiceman]

Originally posted by Jaka Racman
Hi Charles, your description is not entirely accurate. There is one primary oscillator running at 50% duty cycle. Secondary is demodulated by bidirectional switches run at 50% duty cycle and phase shifted towards primary. I have extensively simulated topology in 1999 not knowing it has already been patented . It has some interesting properties like doubling of filter ripple frequency and zero voltage switching for some of the switches. The only difference with the patented circuit was that I used back to back (source and gate tied together) connected mosfets for bidirectional switches instead of mosfet inside diode bridge as Peavey uses.
Do you know if the patented circuit is actually used in a product?

I worked together with Brian Attwood for a couple of months during the research phase of a class d project some years ago. We had some interesting luch time discussions. He is a very pleasant man of high integrity (although a little bit of a lone wolf). I specifically asked him about that design one day because a friend of mine had done something very similar in order to make a sine wave dc to ac inverter (he used back to back mosfets for the bilateral switches).

My friend could not come up with a simple, passive circuit to recycle the leakage energy dumped when the bilateral switches turned off and ended up using clamping diodes to a couple of caps that were then bled off back to the input by a separate little power converter. After much blood, sweat and tears, my friend got his prototype working reliably, but concluded that it was more expense and trouble than the traditional sine wave inverter approach using an intermediate high voltage dc link.

So, Jaka, how did your design deal with this problem?

By the way, Brian Attwood said that he thought that perhaps the main advantage of that class d amp topology was that a lot of the non idealities (offsets, delays, etc.) in the two phased carriers tended to cancel each other in the final output. (Never looked into it myself - what did you find?)

Regards -- analog(spiceman)

[Steven]

Originally posted by Jaka Racman

...
Do you know if the patented circuit is actually used in a product?

I guess it is this one: Peavey DPC1400X

[ingrast]

I toyed with the idea of a line operated class D over a year ago, for high power and to get ride of a high power PSU. Did not pursue it further for it turned out to be no great improvement in comparison with a conventional one.

The concept can be better understood thinking of it as a conventional switcher supply, only providing a variable output voltage instead of DC.

One approach should require a large DC blocking output capacitor, with the quiescent "supply" output at half maximum output corresponding to a 50% duty cycle driven totem-pole. Note this requires the power switches to be driven independently and balanced from 0 to 50% so as to ramp from 0 to full and at the same time avoiding net DC on the transformer.

A second approach should be dual paralell switching chains, one each for the positive and negative signal excursions. This requires additional controlled switches to short the inactive side to ground so as to provide a return path for the active side.

Not that I think unfeasible either way, but still see no huge benefit.

Comments welcome!!

Rodolfo

[Jaka Racman]

Hi,

at the end I decided that the leakage inductance of the transformer would be a major drawback and never got past the simulation stage. That's why I asked if there is a real circuit behind the patent.

Thanks for sharing a real life experience with the circuit.


Best regards,

Jaka Racman

[phase_accurate]

Yes, it is the DPC series that used this topology. They were favoured amongst bass players some years ago (who are always grateful if at least ONE piece of their equipment ins't too large or heavy....).


Regards

Charles

[Steven]

Some other interesting switching amplifier brands/topologies, that use output transformers witch subsequent switching:

http://www.jam-tech.com/main.shtml
http://www.nphysics.com/whitepaper-classnaudio.htm

And another self oscillating one with feedback:
http://www.powerphysics.com/

I have no experience with these amplifiers. Anyone?

Steven