Raspberry : power supply with LT3045 regulator (low noise)

[HarmonicTHD]

From the point of view of electronics, the operation of clock oscillators depends on the quality of power supply, as well as on the oscillators themselves. The better the power supply of the oscillators, the more accurately they generate a given frequency. So the phase noise in the sound region that the oscillator creates is very important. Unfortunately, this is the case and I personally have seen it in practice.

Yet you still lack the proof of audibility. (Not everything what can be measured can be heard, but everything what can heard (controlled tests eg ABX) can be measured).

 

[magicwolf]

Yet you still lack the proof of audibility.

Alas, you demand the impossible from me. I think you understand what devices are needed (or rather, there are no such measurement methods!) to show you this difference. I'm just an amateur. But nevertheless I hear this difference. Last week I listened to an amanero powered by rpi and a separate power supply - the difference was very noticeable. And the problem is not that the accuracy of reproduction of sound frequencies has improved, but in the positioning and visualization of sound images, as well as in the quality of reproduction of low frequencies (since they are longer in time). I do not know of such instruments or measurement techniques that would evaluate, at least qualitatively, the positioning and visualization of sound images.

 

[antcollinet]

Alas, you demand the impossible from me. I think you understand what devices are needed (or rather, there are no such measurement methods!) to show you this difference. I'm just an amateur. But nevertheless I hear this difference. Last week I listened to an amanero powered by rpi and a separate power supply - the difference was very noticeable. And the problem is not that the accuracy of reproduction of sound frequencies has improved, but in the positioning and visualization of sound images, as well as in the quality of reproduction of low frequencies (since they are longer in time). I do not know of such instruments or measurement techniques that would evaluate, at least qualitatively, the positioning and visualization of sound images.

More than likely you are hearing things that are not in the musical signal, but created by your subconious as result of expectation bias.

You could check by controlled blind testing. You don't have to of course - just enjoy what you hear.

But don't expect others to be convinced that your PSU is actually making any real world improvement. Whether you can hear it clearly or not.

 

[fpitas]

Alas, you demand the impossible from me.

No. No, they really don't. If you claim that there are important improvements to be made, the onus is on you to demonstrate them.

 

[magicwolf]

More than likely you are hearing things that are not in the musical signal, but created by your subconious as result of expectation bias.

I understand that music depends on the psychological and emotional state of the listener. But the subconscious has nothing to do with it.

You could check by controlled blind testing. You don't have to of course - just enjoy what you hear.

I don't like blind testing. Firstly, if you conduct blind testing with several people, then they should be able to listen, and as I have seen, there are not very many such people. Secondly, if you conduct blind testing only with yourself, then someone has to switch sources, but in this case it was impossible. Because it required some technical manipulations, and there was no one to carry them out except me. In addition, as I noticed with blind listening, you don't always evaluate correctly, because you start to get nervous. I do not conduct blind listening, but listening to selected fragments of music, but mostly focus on a relaxed perception of music for a long time.

But don't expect others to be convinced that your PSU is actually making any real world improvement. Whether you can hear it clearly or not.

Yes I agree, I can't speak for others, but for example, I hear in the song Dire Straits "Water of love" аt the very beginning of the song there is a knock and an echo. If others don't hear it (echo), it's not my problem.

No. No, they really don't. If you claim that there are important improvements to be made, the onus is on you to demonstrate them.

I am very sorry that you misunderstood me. I didn't claim anything, I shared my own experience. Besides, in this case, I can't demonstrate measurements because I don't know how to do them.

 

[antcollinet]

I don't like blind testing. Firstly, if you conduct blind testing with several people, then they should be able to listen, and as I have seen, there are not very many such people. Secondly, if you conduct blind testing only with yourself, then someone has to switch sources, but in this case it was impossible. Because it required some technical manipulations, and there was no one to carry them out except me. In addition, as I noticed with blind listening, you don't always evaluate correctly, because you start to get nervous. I do not conduct blind listening, but listening to selected fragments of music, but mostly focus on a relaxed perception of music for a long time.

And yet blind testing is the gold standard of all scientific evaluations of human perception. You don't like it? fine, no-one is forcing you. But your statement that you can hear the difference you say you can with sighted listening carries no more weight than my statement that I can run at 30mph. Prove I can't.

 

[ldovr.com]

I read this article where someone explains that he tweaked his raspberry to power it with 3 LT3045 regulators (1,8v and 3,3v and 5v), witch are very low noise regulators
3x powersupply for rpi : Raspberry Pi - Page 6 • RuneAudio Forum
Is there really an improvement by doing this ?
I would like to try but LT3045 are expensive

Hello Thomas,

This modification is compatible with Raspberry Pi 2 and 3B, but not with 3B+ or 4 due to the differences in their onboard DC/DC converters. And yes, it will enhance the audio performance of the Raspberry Pi. For the Raspberry Pi 3B+/4, you can consider using a single 5V board rated for a minimum of 1A (like dual lt3045 board)

 

[ldovr.com]

I use a power supply unit with lt3042 to power rpi 3:

https://aliexpress.ru/item/1005003348091186.html

It really improves the sound quality. But to power the amanero via usb, I use another power supply based on the tps7a4700, as this helps to provide good power to the сlock оscillators. The best option is to use a usb i2s converter with galvanic isolation.

Using a variable resistor within the reference circuit is not advisable. It's recommended to replace it with a fixed resistor for better stability and reliability.

 

[radix]

Alas, you demand the impossible from me. I think you understand what devices are needed (or rather, there are no such measurement methods!) to show you this difference. I'm just an amateur. But nevertheless I hear this difference. Last week I listened to an amanero powered by rpi and a separate power supply - the difference was very noticeable. And the problem is not that the accuracy of reproduction of sound frequencies has improved, but in the positioning and visualization of sound images, as well as in the quality of reproduction of low frequencies (since they are longer in time). I do not know of such instruments or measurement techniques that would evaluate, at least qualitatively, the positioning and visualization of sound images.

What was the original power supply you used that the fancy PS does better than?

 

[TheWalkman]

Perfectly possible. Not sure I would go that far tho - I would probably limit myself to tweeking the main 5V power input with Ian Canada bits:

I‘m applying for a trademark for that streamer: I’m calling it the FrankenPi!

 

[antcollinet]

Can I refer everyone again to my post (2nd in this thread)

If a Pi with stock on board PSU components measures audibly perfect, then no amount of fancy electronics is going to improve on that.

If you think different - and want to recommend any particular device, you really need to put in the effort to prove that it works as you think it does - else you are just making suggestions that will result in wasted money for other people if they take what you say at face value.

 

[radix]

Can I refer everyone again to my post (2nd in this thread)

If a Pi with stock on board PSU components measures audibly perfect, then no amount of fancy electronics is going to improve on that.

If you think different - and want to recommend any particular device, you really need to put in the effort to prove that it works as you think it does - else you are just making suggestions that will result in wasted money for other people if they take what you say at face value.

I was wondering what PS he originally used. I have seen under-powered supplies causing problems, especially if you have a bus-powered DAC or on-board DAC.

 

[TheWalkman]

Can I refer everyone again to my post (2nd in this thread)

If a Pi with stock on board PSU components measures audibly perfect, then no amount of fancy electronics is going to improve on that.

If you think different - and want to recommend any particular device, you really need to put in the effort to prove that it works as you think it does - else you are just making suggestions that will result in wasted money for other people if they take what you say at face value.

There’s an old American aphorism that applies here and in many other audio scenarios:

“If it ain’t broke, don’t fix it.”

QED

 

[magicwolf]

And yet blind testing is the gold standard of all scientific evaluations of human perception.

IMHO this is a method of evaluation, and not a generally recognized international standard.

But your statement that you can hear the difference you say you can with sighted listening carries

It's strange that most people listen like that.

Using a variable resistor within the reference circuit is not advisable. It's recommended to replace it with a fixed resistor for better stability and reliability.

Thank you, I'll remember your remark. But at the moment I sometimes change the voltage, so a variable resistor is used.
I also note that I had to replace the radiator, since the transistor was very hot. Other than that, this power supply is not suitable for USB-hub power supply. ;(((
But this power supply was better than my makeshift power supply made on the basis of tps7a4700. This power supply has less noise, while I was able to connect the dsi display, and the sound is the same as when using a tps7a4700-based power supply.

What was the original power supply you used that the fancy PS does better than?

In the first message, I talked about two power sources. The sound is better when I use both power sources. If the amanero as rpi3 is powered by a single power source (lt3045) - the sound is worse. Unfortunately, powering the usb-hub from its own power source turned out to be quite strange. I use a usb-hub ORICO-HS4U-U3-WH-BP in order to power it, it is necessary that the supply voltage of this hub is greater than the voltage coming from rpi3.

 

[popej]

IMHO this is a method of evaluation, and not a generally recognized international standard.

There are standards, like ITU-R BS.1116.

Anyway, nobody expect you to pass controlled listening tests. Just show measurements.

 

[Greenman]

but mostly focus on a relaxed perception of music for a long time.

I’m with you here.
I agree with many here that expectation bias can dictate results of listening tests to a large degree. I do not like swapping from one source to another using small fragments, you can convince yourself of anything this way IME. What I find is that listening over an extended period means you forget those expectations and just enjoy the music (or not) this way I think you don’t so much focus of minute differences but do just naturally absorb the changes. Swap back a few times and you will know if you prefer one or the other or if they are indistinguishable.

Another forum convinced me that my DAC fed by the inbuilt streamer via LAN was night and day better than DAC fed by toslink via WIIM, I did the tests as above - indistinguishable. I got given an expensive fuse…night and day difference they say …did the tests - indistinguishable. Same with power cords. Yet I hear huge differences changing DACs, many here just won’t believe you..

As for power supplies, it’s well understood that these can have huge (and measurable) impacts on Analogue stages. As for purely digital, in theory a well designed digital circuit should be immune to minor power supply issues (assuming not so bad that it causes glitches), yet other forums seem to have hundreds (or thousands) of people that insist power makes a huge difference even on digital products. I have tried a few different ones on my WIiM but none seemed to have any effect one way or another, but I haven’t tried a really high quality one as yet. If you have done your tests as you say I think it’s perfectly reasonable to give your impressions, you dont have to prove anything. People will evaluate your findings as they see fit, many will dismiss them as fake news, some who may be doing similar experiments might find them useful.

 

[magicwolf]

There are standards, like ITU-R BS.1116.

Thank you very much! Very interesting recommendations. I've never heard of the “double-blind triple-stimulus with hidden reference” method. IMHO three stimuli are better than two (I used two stimuli the last time I tried to arrange a blind audition).

Anyway, nobody expect you to pass controlled listening tests.

Yes, alas, most likely it is. And I can't find twenty experienced listeners. However, I won't do anything grandiose.

Just show measurements.

I love measurements. They make it clear that I'm listening to a technically sound device. For example, I have a weiliang ak4497 board, I have repeatedly made measurements in RMAA. Tell me what measurements you want me to make, if I can, I will definitely do it. I have an e-mu 1616m.

that expectation bias can dictate results of listening tests to a large degree.

I agree with you, I definitely won't pass controlled listening tests, so sometimes you have to listen for quite a long time to how changes in audio equipment affect the sound. There is practically no choice, I don't have golden ears, alas. Of course, as you gain listening experience, you can immediately notice some features in the sound, but these features are usually either immediately audible or not immediately audible.

 

[popej]

Tell me what measurements you want me to make, if I can, I will definitely do it. I have an e-mu 1616m.

Emu was great
I still use 0202 USB: https://gmaptool.eu/audio/emu0202/emu0202-24-96.htm

Would be nice to compare RMAA of audio output with unmodified and modified supply.

 

[BDWoody]

Yet I hear huge differences changing DACs, many here just won’t believe you..

Most will absolutely believe you hear those differences. Most (here at least) will also understand that without controls people hear all kinds of things.

 

[fpitas]

Most (here at least) will also understand that without controls people hear all kinds of things

But, we're prepared.