FET as CCS in power supply?

[Dr.H]

Possibly a daft idea, but what would happen if a JFET was placed in CCS mode in between two caps in the initial ripple smoothing stage of a power supply? Would it lower the noise on the supply or accomplish anything good?

[flg]

Similar in theory to the Pi filter? An inductor between the 2 caps. The inductor is a current source, it resists the change of current.

[sawreyrw]

The Idss of the FET must be larger than the load current, otherwise the output voltge will drop. This means that you can't really do what you want and the FET will be in cutoff. Note that the definitions of cutoff and saturation for a FET can be confusing. When a FET is in cutoff it really means it is "saturated" in BJT terms.

[h_a]

You essentially mean simple linear regulators (be it jfet, bjt or whatever); pretty common for small signal stages and usually prohibitively inefficient for power amps.

Hannes

[analog_sa]

Originally Posted by flg

Similar in theory to the Pi filter?

The big difference between an inductor and a CCS is that the inductor has very low impedance at dc. A direct replacement is impossible.

Dr H's idea has some merit in particular situations and i have done something very similar in low current high voltage supplies. Imagine a tube rectifier feeding a giant capacitor through a CCS. Obviously the supplied circuit has to consume less current than the CCS but in effect you get a very high level of isolation from the mains and really low noise. At the same time you get very low PS impedance thanks to the giant electrolytic. At the time i used an EZ80 feeding a 3300uF and powering a phono pre. You can probably achieve similar technical performance in a more elegant way but i really liked the sound. Took almost an hour for the cap to reach operating voltage.

Not sure if this is useful in the context of power amps.