TAS5706

Scanty details on this "digital" amp so far regarding the output stage correction circuitry that reduces power supply noise. Anyone know more?

For all intents and purposes, it's a closed loop amplifier. Normally, most class D amps are open loop, meaning any noise on the power supply will get coupled into the output. (as class D essentially switches the power supply into the speakers multiple times a second).

By having integrated feedback, the noise of the power supply gets rejected.

Open loop amplifiers require power supplies that are superior to those required for closed loop. (superior supplies = $$)

/R

I'd still like to know how the loop is closed---evidently, according the the blurbs, it is confined to the output stage, as opposed to some other schemes. [ EDIT: That is to say, the question is not why, but how.] The PSR with frequency didn't seem to be shown, although there was a remark about 120Hz rejection improvement over open loop, and the important distinction about measuring with a signal and load compared to what most do at zero signal.

I may have remarked in here that I was negotiating a few years back with TI to do some interim hybrid design while the European team worked on a next-gen higher-power digital modulator system. We deadlocked on price, but before that got obvious I had hinted at some approaches to doing a form of PS compensation. Not that I am claiming those ideas were lifted. What may be greater issues are possible conflicts with some of Gerald Stanley's recent patents.

But then the switchamp arena is a minefield for IP.

I have seen class D where the loop is closed around the output stage.



some discussion here.

http://recforums.prosoundweb.com/index.php/m/326030/22399/0///361/#msg_326030

This is not your old school PWM with triangle wave and comparator. The lag from the output filter determines the amp switching rate.

JR

Yeah, that's basically Bruno Putzeys' uCd, and it works very well indeed. I was told NXP/Philips will let anyone use it if you agree to populate with at least some NXP silicon.

Don't use the H* word with the Hypex folks though---they seem to be unduly sensitized.

But the TI beast is clearly a digital modulator-based amp. They don't seem to want to talk about the details of how the output stage correction is done. Perhaps it is still in patent purgatory.


*heterodyne---the toplogy is inherently variable frequency, thus can be prone to artifacts in multiamp systems.

Here's a nice little story about Bruno I just found:

http://spectrum.ieee.org/feb08/5924

I have to smile about the significant other in...Rwanda :roll: