[sdiy] DC wallwart +/- on-board V Reg ?

[Sarah Thompson]

Best option is to put a pi filter at your power input followed by a regulator, followed by a reservoir cap. One of the pi filter's caps can serve as a reservoir cap for the input of the regulator. Something like 100uF (with a 100nF in parallel), into a 1uH inductor (ideally shielded, make dure its DC resistance is in the tens of milliohms so it stays cold), into another 100u and 100n in parallel, into a 3 terminal LDO reg, into something like a 10u, then make sure all of your major devices have a 100n between their rails as close to the chip as possibile. Using a ferrite core inductor reduces Q in the filter, which is a good thing because you don't want resonant peaks in the frequency response. Pi filters are simple and really cheap when made from commonly available components rather than bought as modules. Just keep they layout nice and tight and you'll do fine.

This combination is pretty much bulletproof, even in very noisy environments. It also deals well with switch mode wall warts that kick out a lot of RF. The trick is basically that the pi filter's first set of caps give a very low input impedance above a few kilohertz -- choosing low ESR caps really helps here, and putting several in parallel that are orders of magnitude apart in value greatly improves their effectiveness into the VHF/UHF/microwave range. The inductor basically maintains very low resistance at DC, but increases resistance rapidly with frequency, so,the second set of parallel caps get to pretty much nuke anything that got past the first set. This means that any remaining nasties will be low frequency in nature, so the feedback loop(s) internal to the LDO will very effectively reduce them.

If this still doesn't work well enough, you can go to common mode chokes, isolated supplies, etc., but that is rarely needed in practice.

Sarah

Sent from my iPad

> On Mar 5, 2015, at 4:27 PM, Paul Perry <pfperry at melbpc.org.au> wrote:
> 
> When the regularor is in the wart, there could be a problem of
> interaction between circuit sections because of the common resistance in the wart lead.
> 
> paul perry Melbourne Australia
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy