A few years ago, when I bought my microRendu and SonicTransporter for disc ripping and storage, I added a linear power supply from HDPlex. I thought the sound was slightly better compared to the iFi power supply. After owning the gear for a year, I saw that I had enough power to also run my Mytek Brooklyn DAC/preamp. Dramatic difference in noise floor compared to plugging into a PS Audio power conditioner directly.
So please show us your data documenting the assertion that a change in power supply produces NO resulting change in audible sound quality?
2 Likes
Even worse, someone who has a company with the word Green in the name uses linear power supplies needlessly wasting electricity not to mention bringing any noise there might be, right smack into the lower audio band where you most certainly will hear it. Itās a marketing tool and one can fill out an amplifier case with huge wasteful transformers and capacitors. Those small circuits on the perimeter are the actual audio circuits. As an Avionics Tech, there is a reason why we use 400hz in Aircraft for decades now, smaller and lighter transformers can be used since weight is everything. Move that up to 500khz switching and itās a no-brainer. One can easily make a power supply that can provide twice the current that will ever be needed in a compact, cool, 90% efficient package that even a bat canāt hear.
1 Like
Hi, Emory. As I said above, I try to keep an open mind on this. And itās certainly true that at one time, switching power supplies were very noisy ā I used to design them in the 80ās, for very high-power industrial applications, before I committed to the one true vocation which is Software 
.
As for data, the indefatigable hobbyist and fellow forum member @Archimago has posted some measurements here which incline me to believe there will be āNO resulting change in audible sound qualityā. So unless I can see some evidence to the contrary from someone whoās not trying to sell me a different power supplyā¦
Oh yes they most certainly do have transformers, the switching just provides a higher frequency to feed the transformer, much smaller ones are required the higher the frequency to do the exact same work. This is why Aircraft have been 400hz since World War II and are light enough to fly. We had linear power supplies in the bad old days because there was no other choice. I have yet to see a switching power supply without a transformer. When you can make a tiny up to 90% (often 70 to 80% in practice) efficient power supply that can supply current far higher if demand requires in a much smaller package, itās compelling. There were more challenges to overcome with an SMPS but the biggest factor is RF shielding. The switching frequency is so much higher than earlier versions that a Bat couldnāt hear it now. Donāt overlook the longer holdup period of a good switching power supply in comparison.
Even as I was forced to use linear power supplies for part of my life, I have no nostalgia at all about them. I wouldnāt be surprised if some governments outlaw the use of linear power supplies, some are only 25% efficient. Itās really bad when one particular company makes a large amp and uses a heat generating linear power supply and then requires one to convert to a 20 amp circuit in the home to power it. I ran as fast as a I can from that product.
I might also mention that all new homes in my state require solar power or no building permit. In this case people should be able to use any type of power supply they want whether I agree with them or not. Iām just providing arguments, not making lifestyle decisions. I will say that at least a linear power supply will not do any harm other than wasted electricity. If it sounds good I donāt think much about how the circuit got itās power. I hate electric bills and more load on my A/C. We could not have the car audio we have today without switching power supplies, in the car it removes a stage. In an AC environment, the AC is first dropped down by a transformer and rectified to DC but it does NOT need to be very big, then that DC is switched to a very high frequency and then that goes through a transformer and rectifier as well but the switching circuit gets some feedback from the output and changes the pulse width if necessary based on demand, in the end you get positive and negative DC to feed to positive and negative DC rails of the audio circuit and the power supply is far smarter than blindly supplying a 60hz sine wave to a huge transformer and then that being rectified and huge Capacitors that can run out of storage in some situations. In the car you already have the DC to switch from.
3 Likes
Iād suggest this post find its way into the Roon KB. Thanks.
2 Likes
Why? 
Cat, pigeons etc.
You could try a rechargeable battery and remove AC from the equation. If that sounds better to you then it might be worth investigating some method of power conditioning.
I do this sometimes but mainly to make my DAC and headphones slightly more portable. I canāt say Iāve noticed much/any difference. Certainly not enough to fork out more money on power supplies.
I use a Paul Hynes SR7-T as a linear power supply for my Roon Nucleus. Easily audible improvement that Iām very happy with. 
1 Like
Thatās not how SMPSs are usually designed.
Basic main function blocks are:
The incoming 50/60Hz AC is rectified to DC, then switched at a high frequency from anywhere around 50 to 500 kHz to be passed through a comparatively tiny transformer to step down the voltage, then finally rectified to DC. Of course there are several filtering stages and regulation loops in between and overall.
1 Like
ā¦ hopefully not using one of these USB powerbanks, since they internally work on a different voltage and typically use high frequency switch mode regulation to achieve their nominal output voltageā¦
1 Like
Iāve updated my post to prefix the word ālargeā in front. I was worried someone would be pedantic about technical details while i was trying to make an easy to understand explanation, but I forgot about that part. I did already prefix ālargeā in the last part of the post.
You clearly know what Iām talking about, so I think ālargeā should satisfy you. If it does not, suggest something better while keeping the easy-to-understand nature of the explanation in tact.
How did this gorilla sneak into the temple of science?
Inclines me to believe otherwise. In fact, the science does show differences, miniscule as they may be when viewed in isolation on a scope. The total frequencies emitted by an orchestra at any one instant might approximate an infinitudeāat least too many to wrangle with home equipment. And they change at speeds that also really do sometimes boggle the mind.
So, thanks for an interesting link.
OK. Well, Iād still love to see evidence of audible differences, if you run across such a thing.
1 Like
Some posters in this thread wanted some measurements on linear power supplies, and maybe how they relate to the RPi.
Here you go. Being new to the Roon/RPi world and coming from a tablet as my USB source I was curious about the impact that different power supplies make on the sound. This is a work in process, much still to do. This is highly technical, I am happy to answer questions you may have. I have yet to do end to end audio performance testing, so this is just about power supply noise. Iām a degreed EE and have some decent test equipment. I donāt work in audio anymore and donāt have any affiliations. Again, please let me know if you have any questions.
https://1drv.ms/w/s!Am4mUJvVR_bzgcUpL_3HSnDwn2VbDw?e=jaIfrp
Tom
7 Likes
I applaud your efforts - adding some real value here (pun intended)!
Looking forward to the analysis of the DAC output measurements.
2 Likes
Tom, thanks for undertaking this. Am I correct in understanding that you are measuring the noise on the power delivery pins of the USB sink? Would be useful to understand that, I think, for USB-powered DACs.
The Shiit Yggdrasil seems a problematic DAC to test with. A lot of electrons have been stripped from their nuclei and sent on careening paths through the Internet debating the noise characteristics, and indeed the audio characteristics, of that DAC. The Benchmark DAC seems to have no such reputation issues, though.
2 Likes
The power supplies used in Naim amplification equipment, DAC etc are linear and make a profound improvement in SQ
Show me the measurements, please, Rob.
1 Like
Bill - Iām measuring both the power supply at the input pins to the RPi, and also Vbus on the USB output of the RPi. I found little difference, as you would expect, with some prominent noise peaks on the Vbus output that werenāt on the power supply input.
There was little to no noise difference between the linear power supply and the switcher when hooked up to the Pi. This would discount using a linear supply on the Pi, as the Pi makes its own noise. It would point to using a supply with very low broadband output impedance, as this had a direct influence on the amount of noise generated.
Then I split Vbus and fed the DAC side from a linear power supply.
Both the Benchmark DAC3 HGC and the Yggdrasil need 5Vdc on Vbus to work, even though neither is powered by Vbus. So while I donāt know their internals, something in them is using Vbus.
I did this to keep the RPi Vbus noise from getting into the DACās internals and coupling into the audio ground.
Nobody listens to Vbus, it would be pretty boring, but maybe better than Primus.
So I need to do some end-to-end testing to see if I can measure a difference.
I can hear a difference for sure, I prefer the low noise Vbus power supply. This is true for both DACās and is repeatable over several weeks of listening. But that improvement could be generated in my auditory cortex for all I know as there is no way for me to do a blind test.
Anyway, I learned a few things:
-
The RPi is noisy (its USB Vbus is noisier than my SurfacePro 3 laptop) and even a low noise linear power supply wonāt help that as the noise is generated by the RPi and modulates the power supply input. A low impedance supply will help here.
-
WiFi is even noisier, with > 100 mV negative going spikes on the supply. Just say no to WiFi and the Pi, at least from a noise perspective.
-
I also measured noise from the power supply (-) output to earth and found that many supplies generate significant noise current. Especially offline switchers like the iFi products. Actually anything with a two wire AC input dumps noise on the (-) output terminal. The easy way to fix this is to simply earth the (-) lead of the DC supply to give this noise current a low impedance path to ground.
-
A Vbus break and low noise Vbus power supply MAY provide a performance improvement and is easy and cheap to implement. Still need to measure it.
Some audio gear might see an improvement using a linear power supply. But there are no guarantees, and in the case of the Pi it may be worse.
7 Likes