Tas5142 Ucd?

[Anthony C Smith]

I have been looking at the TI part TAS5142 its a switching block for a digital amp chain.
Has anyone already tried placing a comparator on the front for a compact UCD, cos it looks like its begging for a try!

http://focus.ti.com/docs/prod/folder...t/tas5142.html

regards

[Ouroboros]

I've not tried this part, but I did try an earlier, higher power part, the TAS5261. That part had internal circuitry that tri-stated the output FETs if there was no switching activity for a certain time. The effect was that if you had a UCD implementation using it that was driven into clipping on a 1kHz sine wave (say), then the output stage would switch off, requiring external circuitry to reset it to start it up again. (as I'm sure you know, a UCD amp will stop oscillating during the clipped part of the waveform).

This part doesn't seem to have that internal function, so it may well work well. No good for commercial use though, as using it would be outside the Philips/NXP constraints regarding licence-free use of the patented UCD technique.

[Anthony C Smith]

There is a note in the text saying about modulation depth over 96% preventing the bootstrap cap charging, causing failure of the upper fet; so this may be a problem
as for the Phillips NXP patent see post

http://www.diyaudio.com/forums/showt...34#post1271334

and given the time it took for me to get a reply from Nick Koper I think they have given up!
regards

[Ouroboros]

The data sheet only shows a 33nF cap for bootstrap, which may well discharge during clipping in a UCD implementation. It looks as if a larger value cap could be fitted, if the internal diodes will stand the pulse current through them.

[Anthony C Smith]

it should be no problem to place a uf4002 between the cap and the GVCC and spare the internal diode its blushes!

[Eva]

The only reason that I can see for such low bootstrap capacitor values is cost reduction.

The usual fluctuations in bootstrap capacitor voltage that arise when the output approaches the rails and duty cycle changes result in fluctuations in switching times and dead time. This translates into increased THD because bootstrap capacitor voltage usually drops and dead time increases.

This can be easily seen on oscilloscope by synchronizing to the input edge and watching the signal dependent jitter on the output edge. Values as high as 47uF are required to reduce this effect to the point that it can no longer be seen.