DAC Power supply

[elec-tron2]

I post a thread for two reason :

- The first that have maybe been discussed elsewhere is to know what do you think and advise to use as power supply in a DAC, ie just integrated serie regulator like 78xx or LM317 or other, just shunt reg, the two regulators, just a discret serie regulator like Waltz /Jung etc etc ...

- The second is a try to make this thread a book of schematic of power supply and schematic of regulator for DAC designer primarly but other PS schematics are welcome

I invit all that have schematic to add these to their reply.


What do you think of using a discret serie regulator Jung reg with local shunt reg per IC chip ?


Thank you all

[FastEddy]

IMOP cleaning up the source is more important than the actual DAC regulator. The best way to do this is with optical isolation. If the source is just the isolated digital signal, then most of the clean up battle is won.

As for IC regulators, it is very hard to do better than the 78xx / 317 type with plenty of bypass capacitance. If the originating power is from the wall (110 to 250 VAC) then, of course, removing the 50/60 cycle noise and associated harmonics and stray EMF should be done, first and foremost, then straighten out the ripple with a well thoughtout linear regulator. There is nothing wrong with the Waltz / Jung discrete regulation schemes but the differences (after the fact of the diode bridge and main filtration) are insignificant compared to the original noise sources, the power grid.

The best practical scenario so far that I have seen is a pair of motorcycle batteries wired for +/- 24 VDC rails, trickle charged from the wall by a linear regulator, then further regulated down stream with lower power linear regulators for the lower voltage branches, +/- 12 VDC, +/- 9 VDC, +/- 5 VDC, + 5 VDC ... etc. This is a common setup in trade school labs and can also be found at development centers at TI, AMD, Analog ... for use evaluating op-amps without shielding, etc. This is of course based on very common, very reliable, very clean TelCo methodology = POTS long line telephony amps.

Actual measurement of the noise component at the output of the amp compared to no signal input is worth more than all the effort put into the regulators = a set of 100 nF polystyrene snubbing caps right across the power pins of the op-amp covers a world of sins in the PSU ...