Audiophile power supply for Raspberry Pi

Try the cheapest 2A 5V PSU you can find and change it if you have problems.

1 Like

The issue with power supplies is PSRR, over frequency response. All high-end digital systems seek out the best PSRR. In digital circuits it improves clock functions and helps reduce Jitter. There is also a lot to be said, for the way op amps reject noise too, for they will also reject audio frequencies in doing so. I go for the best power supplies and regulators, if you build your own, as I do, here is a useful URL to choose the best low noise regulators currently available. Build your rails and watch out for ground loops. Chose good low ESR caps, I like to use lots of large microfarads on my rails. My DAC has over 30k Microfarad worth of caps within it. This isnā€™t rocket science it is standard design concepts for power supplies, on high end system.

hifiduino.wordpress.com/2011/03/ ā€¦ egulators/

I like the ADM7151 and have replace some LODs in my DAC with them, after I reclocked it. Made a huge difference on detail.

Differences in Toslink cables can be heard, cheap ones promote high Jitter, and will sound a lot like wow and flutter problems on a turntable with the cheapest. I like to only use Glass, not plastic. Then again if the jitter is lower the you can hear, it will not show up as distortion, but will be reflected as missing detail in the sound itself.
Have you ever heard music that sounded like it was coming from behind a blanket. It isnā€™t distorted, but not sharp, not clear. The frequency response is hindered, or just that something is missing. Bad cables as well as PSRR and Jitter issues will present itself like this.
The only way to measure this is comparing the frequency response from source to playback or by listening, I have been listening for over 30 years, and have well trained ears. I trust them over any measurement or debate. I donā€™t need someone to keep telling me it is some placebo effect, because that is bull. I am not alone.
Those of us who have Tube systems have a much higher level of distortion then Solid State, does that make Solid State better, because it has better numbers, when it comes to distortion? Absolutely not!

@bearington: subjective 1000% agreement!

High End Digital audio is a very sensible thing to deal with.
It took decades to discover the origins of the bugs in digital audio technology.
But nowadays we are able to face the facts. So letā€™s make the best out of it.
Why should someone make such big efforts in creating an audiophile music player like
Volumio when no one is willing to bring out the best of itā€™s potential?

Greetingz, Robert

And why should someone spend so much money if the person cannot hear it?

So when someone asks (and therefore probably has no particular requirements) I say: take the cheapest and change if if necessary. Spending like 5-10 dollars more if it wonā€™t be good enough is a small problem, spending from the beginning tens of dollars without needing it is silly.

In my case everybody who was present at a comparison of different quality PSUā€™s did recognize the difference clearly. Also the guys were able to hear the difference between a cheap toslink cable and a lifatech glass fibre for example. Even in a blind test.

I think, that some people are so terribly convinced of their individual knowledge, that they just donā€™t want to hear differences in some cases. So this seems to be a psychological problem.

But- people should decide on their own if the hearable improvements are worth investing some more money and efforts or not. There are definitely listeners out there that arenā€™t sensible enough. Some few tests will clear the facts for everyone and all time. But I donā€™t think that never trying it out is the way to go.

1 Like

In order not to get me wrong:
If someone has no motivation to get the last percentage in sound quality, itā€™s free for him to decide to stuck with a 5$ psu and just to enjoy the comfort of a player like volumio.

Different point of views and experiences are the well of an inspiring forum :wink:

Do any of you guys, who obviously know a heck of a lot more about this than me have any opinions on using an ATX power supply for the Pi? Either using: 1) the stand by 5v; 2) main 5v rail; or 3) the 12v rail, with a variable regulator reducing voltage to 5v. I am using 3) and this seems fine to me - plenty of power on tap. Using 18awg twisted pair to get power from the external ATX supply to the regulator.

Apologies if this seems like a dumb question. :unamused:

Hi Pete!

There are no stupid questions at all!

Iā€™ve never tried an ATX power supply in this case so Iā€™ve no experience concerning that results. Your chosen way to use it seems to be the best way.
Quality of ATX power supplies vary a lot. Fact is that most of them are switching PSUā€™s and usually noisy.
Even the regulator will not eliminate all of the noise but regulating down from 12 to 5 V will logically bring down noise also.
To make it more stable and clean you can add about 8000 to 10000ĀµF of capacity on the input side and additional 1500 to 5000ĀµFā€™s on the output side of the regulator.
Try to get them as close as possible to the pins. Also you should add on both sides about 10nF MKPā€™s and 1Kiloohm (1 Watts) resistors in parallel- this will reduce high frequency noise of the ATX- PSU and the resistors will bring Voltage down after power off. Take high quality MKPā€™s ones with high V/Āµs rate (and 630VDC max). For the large capacities I recommend Panasonic FM 35V/1500ĀµF (you can put several of them together in parallel- this makes the whole package much faster).
Most benefit is achieved if separating 5V rails of raspi (or any other computing platform) and DAC. Also you could use separate PSU to HDD if used. With an external 3.5 inch HDD this is solved automatically.

Try to avoid heating up the regulator! 12V-5V means that there are 7V left. In case of drawing 1A current you have 7 Watts on that part (Measure temperature). You should cool the regulator- best when isolated and screwed to a housing. Depending on the regulator youā€™re using be careful not to contact grounded housing with the back side of the regulator. Most of them have Vout on the back side of the design and this would short the circuit (look for an isolating kit for your used regulator in this case or use ground or mass isolated heat sink- must be effective enough!). Check the regulatorā€™s datasheet!

Please make changings only if you know what youā€™re doing! :mrgreen:
No guarantees at all for all the above info! :mrgreen:
Generally pay attention to mains supply!!!

Greetingz, Robert

The problem with smpsu are they are a switching device and can produce noise. Also the standard 3 pole voltage regulator such as 705 series also produce HF noise. I would advise where possible to use a toroidal transformer followed by full wave rectifier and good volts stabiliser . if you do use the 705 range do not be tempted to omit the low value .47 Mfd capacitors on the input and output rails. Whilst you may not experience this noise if using an external DAC remember that noise can still be coupled to the DAC unless using an optical coupling

A regulator can be enough, provided you smooth out spikes.

youtube.com/watch?v=WxhjLIu-vPg
cds.linear.com/docs/en/applicati ā€¦ an101f.pdf

It is very interesting and teaches that you can use a switching followed by a linear one, so that you have a more compact and mode efficient (less heat) setup than a normal toroidal+linear.