ClassD phono preamp

[lumanauw]

Is it good or not good to make classD phono preamp? Given a very weak input signal?

[zilog]

Why do you want to add more distortion than necessary here given the low power requirements? IMHO class D is a compromise for efficiency and size, you do not have these limitations in this application so why have the all the drawbacks for no apparent reason?

[lumanauw]

Just because it sounds "different"

[BWRX]

As long as it has a high switching frequency and high gain, sure why not. It is much simpler to use a non switching topology though.

[lumanauw]

Will a selfoscilation UCD type fit in? This topology has residual sinusoidal to keep the amp switching, and it is quite big, maybe bigger than the phono signal itself.

In this case, the idea won't work because the audio signal is smaller than the residual signal needed to keep the amp working?

[BWRX]

That's a good point but the self oscillating class d amps should still be able to amplify those low level signals. I also believe the UcD works well with higher than normal gain settings. A higher gain will reduce the level of the output signal seen back at the input.

To be honest I'm still trying trying to fully understand the relationships between hysteresis, propagation delay, feedback, etc. in the self oscillating topologies.

[Pano]

A unique idea!

My 1st thought would be that noise would swamp the tiny phono signal.

I disagree somewhat with Zilog. Small size and efficency might have been the original goals of Class-D, but for me, the great advantage is the switching. In a switching amp we are not asking the transistors to work in a linear way - simply as an on/off switch.

How well that would work at phono or microphone signal levels I can't imagine.

[JohnG]

I have given this some thought over the last year or so as well, although I was thinking of a gain stage for a regular preamp.

With small signal levels, this could be implemented with a high speed comparator and some chip inductors and capacitors. Switching frequency could conceivably be very high, in the tens of MHz. This would allow for a fair amount of post-filtering if desired, possibly with air core inductors of low cost and reasonable size. Layout would be tricky, but should be doable.

Just my $0.02.

John

[BWRX]

Originally posted by JohnG
With small signal levels, this could be implemented with a high speed comparator and some chip inductors and capacitors. Switching frequency could conceivably be very high, in the tens of MHz. This would allow for a fair amount of post-filtering if desired, possibly with air core inductors of low cost and reasonable size. Layout would be tricky, but should be doable.

I would agree with you John. The switching frequency has to be high in order to recover as much low level signal information as possible, but tens of MHz might be tricky. Layout and parts selection would put a cap the upper limit.

[raintalk]

had this in my links:
http://www.oliveaudio.com/index.php?page=20