Low Noise 5V Reg

[FOXYE]

What do you think about the following circuit:

http://img859.imageshack.us/i/5vreg.jpg/

[Ken Newton]

Well, it looks a lot like the feedforward 'clean-up' shunt regulator design of Wenzel Associates. Except that it has an second (incorrect) phase inversion due to having cascaded two grounded-emitter stages.

[AndrewT]

Why are R1 and R2 both in there?
Is something missing?
Why has you used a remote server to store the pic?
We will lose it when image shack decides to break the link.

Attach a compressed, appropriate resolution pic to your post.

[ywara]

I can't say I'm certain what you're trying to achieve here. You'll get a maximum 2dB rejection across the audio spectrum with full 1A load, and effectively zero rejection at reasonable loads.

[ywara]

I found the original source of this topology (well, ONE original source) here: Finesse Voltage Regulator Noise! (Fig. 2)

Basically, the OP's circuit needs a PNP as the high-current shunt transistor.

The suggested PNP darlington pair wasn't too much fun to bias with only 0-5V to play with, so I gave the 2n4403 a shot. With proper biasing you can get around 18dB rejection at 120Hz. It also seems incredibly sensitive to the surrounding components, especially R1/R2. It doesn't seem terribly sensitive to differing loads.

It's a pretty useful topology, but with temperature drift and component tolerances potentially wreaking havoc on its ability to reject high frequencies, it might just be a waste of power.