Help with linear power supply design for my DAC

I too have a Raspberry Pi (3B+) and a DAC (IQaudIO DAC Pro). Like the OP, I am experiencing interference signals from the switching power supply powering the Raspberry Pi. The sound is a slight high pitched note that becomes audible when the DAC is not playing. Rerouting the audio output wires from the DAC to pass near the SMPS reinforces the high pitched note. This confirms it is the power supply causing it.

The solution is a linear power supply, and the problem with the Raspberry Pi and DAC, is they use a 5V DC supply. Because the Raspbery Pi requires quite some current, 3A, this makes the design of a linear power supply highly energy inefficient.

I would like to have some reply from a guru, there are many ripe ones frequenting these fora 🙂, how I can design a smoothing filter which makes a linear power supply more efficient compared to what I can imagine. The latter is usually a straight forward solution. I am thinking about a CRC filter, but does it do what I want?
 
Last edited:
OK, good. At a civilized 120mA and with a full volt we can waste if we want to, designing the filter shouldn't be too tall of a challenge -- if necessary.😉 Let us know how the test without one goes.

Rerouting the audio output wires from the DAC to pass near the SMPS reinforces the high pitched note. This confirms it is the power supply causing it.

If the supply is 'open frame', what it confirms is that it needs a shielded enclosure. That test also does not address how much of the interference is conducted vs radiated. Since a switching supply has no reason to generate sub-20kHz racket, it is almost certainly a beat product with one of the persistent signals on the RPI or DAC boards.

.. how I can design a smoothing filter which makes a linear power supply more efficient compared to what I can imagine.

A linear supply shouldn't need the smoothing filter. And if efficiency matters to you (and I hope it does since you're managing 3A at 5V), stick with the switching supply and filter THAT. You can't improve the efficiency of a linear supply with more filtering, only reduce it as more power must be wasted.

Cheers
 
Last edited:
Rick PA Stadel said:
If the supply is 'open frame', what it confirms is that it needs a shielded enclosure
The SMPS is housed in a plastic box as supplied by the manufacturer. These days, it is difficult to find a metal box. Furthermore, if the interference is due to frequency mixing it can travel along the power supply lead. This is why I am asking here. Yes, I can simulate an nth order CRC filter but that does NOT substitute the experience ripe gurus have.

I mentioned energy efficiency with the filter design, because it makes sense to write about the minimum energy loss for an optimised filter. ONLY experienced gurus can do that.
 
Well, fairly sure I don't qualify as a *ripe guru*, but did think it was worth considering both options (paths for the spurious signal to take).

Any chance you could find some spray adhesive (used in crafts, car headliners, etc) and line your SMPS box with aluminum foil and ground it? I wouldn't expect total silence from any piece of electronics within eyeshot of an unsheilded SMPS.

Sorry to hear of your metal box shortage.

Regards
 
To contain radiated Electromagnetic Interference. Off line SMPS's can have dV/dt and dI/dt edge rates in the hundreds of volts/amps per micro second. Not only the transformers and inductors generate EMI, even the foil paths do at those edge rates. 😉

Cheers