Raspberry : power supply with LT3045 regulator (low noise)

[th2j]

Hello everyone,

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

Thanks
Thomas

 

[antcollinet]

Hello everyone,

I read this article where someone explain that he tweaked his raspberry to power it with 3 LT3045 regulator (1,8v and 3,3v and 5v), witch is a very low noise regulator

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

Thanks
Thomas

Pretty unlikely - and absolutely zero difference if a digital output to an external DAC is used.

For example - look at the review for the Allo Boss here. A raspberry pi based streamer with DAC hat installed directly to the Pi.

It basically measures as audibly perfect. No special power mods needed. (A dedicated external PSU is used in the test - but I'm not sure how much difference this would make)

Allo Boss2 Review (Stereo Streamer)

This is a review and detailed measurements of the Allo Boss2 Ethernet (and wifi) streamer/player. It was sent to me by the company for testing and costs US $149 by itself. My tested configuration included the Nirvana power supply which brings the cost up to US $209. I must say, I am very...

www.audiosciencereview.com

 

[DVDdoug]

Do you have a noise problem?

I don't know anything about the Pi circuitry but I assume the existing voltage regulators are linear and linear regulators are generally quiet. Switching regulators can be noisy but they usually operate above the audio range so they don't usually generate audio noise.

Switching power supplies are usually quieter than linear supplies because they operate above the audio range and the higher frequencies are easier to filter-out than 50/60Hz hum (or the rectified 100/120Hz).

The digital electronics (most of the Pi) is immune to (normal) noise.

If the DAC's power supply is noisy, that noise can leak-into the analog output. For the Pi's built-in DAC, that's probably the 5V supply.

The normal data-switching of a processor can generate noise on the power supply, so the noise may not come from the power supply itself and changing the voltage regulator won't help.

 

[th2j]

no i don't have noise problem, i just want to improve my system
i use a SMSL sanskrit mk2 power by the usb of the rpi

the regulators witch create 3.3v and 1.8v in the pi are switching (XR77004 chip on rpi3 for example)

 

[elberoth]

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:

 

[LTig]

Hello everyone,

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

Thanks
Thomas

Did he show any measurements to proof his claim?
Is his measurement equipment good enough (let's say an RME ADI-2 PRO fs or the Cosmos ADC)?

If one of those answers is NO I would stay away from fiddling with the power supply.

 

[antcollinet]

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:

For the reasons I gave above (stock Pi audibly perfect) all those Ian Canada "upgrades" - for me - amount to snake oil.

 

[MCH]

no i don't have noise problem, i just want to improve my system
i use a SMSL sanskrit mk2 power by the usb of the rpi

the regulators witch create 3.3v and 1.8v in the pi are switching (XR77004 chip on rpi3 for example)

i have the same question that the forum you linked. How do you plan to disable the onboard power managment chip? (maybe keeping global_en shorted to ground??)

here is the schematic of the 4b. Seems that the chip is controlled via i2c. The test points are at the outputs of the regulators. I don't know if just soldering an external regulator there does anything to noise. Maybe others here can tell.

see also at least the 3.3V output is labelled as "LDO"

I have no idea but it seems to me that it might be a very complex project with probably no benefit.... my 2c.

check this out, this guy managed to replace a broken chip and reprogram the new one, maybe you can get some useful info there:

RaspberryPi 4 MXL7704-R4 PMIC replacement and repair - Raspberry Pi Forums

forums.raspberrypi.com

 

[th2j]

i have the same question that the forum you linked. How do you plan to disable the onboard power managment chip? (maybe keeping global_en shorted to ground??)

here is the schematic of the 4b. Seems that the chip is controlled via i2c. The test points are at the outputs of the regulators. I don't know if just soldering an external regulator there does anything to noise. Maybe others here can tell.

see also at least the 3.3V output is labelled as "LDO"

I have no idea but it seems to me that it might be a very complex project with probably no benefit.... my 2c.

View attachment 309632

check this out, this guy managed to replace a broken chip and reprogram the new one, maybe you can get some useful info there:

RaspberryPi 4 MXL7704-R4 PMIC replacement and repair - Raspberry Pi Forums

forums.raspberrypi.com

in fact my plan was to do it with the rpi 2
the process in much simpler

 

[radix]

no i don't have noise problem, i just want to improve my system
i use a SMSL sanskrit mk2 power by the usb of the rpi

the regulators witch create 3.3v and 1.8v in the pi are switching (XR77004 chip on rpi3 for example)

If you want to improve your system, take a look at your speakers, speaker position, subs, EQ, digital room correction, and room treatments. Those things can make huge differences (i.e. 10dB or more). Have you used a measurement mic and REQ to test your system?

Tweeking the power supply of a DAC will make minuscule differences, and it's arguable if they are even audible. Most anything on the solid-state side will make minor differences, assuming you have at least passing equipment (e.g. anything "recommended" here).

 

[th2j]

ok thanks for your answers
you have convinced me to give up this project
maybe i will change my dac in order to have a dac witch is not powered by usb

 

[SuicideSquid]

ok thanks for your answers
you have convinced me to give up this project
maybe i will change my dac in order to have a dac witch is not powered by usb

What is the specific problem you are trying to solve?

You said you don't have a noise problem. The SMSL Sanskrit Mk 2 is a transparent DAC with no audible jitter (on USB, coax is a different story), noise or distortion.

So what is the problem?

 

[direwolf08]

Not the same thing, but once I bought an RPi4 power supply that had audible hum itself, coming from the wall wart. I could hear it during quiet parts of the music. I swapped it out.

 

[somebodyelse]

maybe i will change my dac in order to have a dac witch is not powered by usb

Why? A competently designed USB-powered DAC will have all the internal regulation it needs to deal with the expected noise on the USB power line. Archimago has a few articles with measurements of the noise on the USB power line, and the total lack of difference it makes at the output of the DAC.

 

[charlielaub]

The "noise problem" with the Pi series has nothing to do with how noisy the voltage regulators are. The noise source is the board itself and the fact that is contains lots of transistors that are rapidly switching on and off.

My solution to computer noise was to switch to external USB audio interfaces which have their own power supply. Noise coming in via the USB connection can be filtered out, and the PS is independent of the computer. I use good quality pro audio recording interfaces, which are well designed, offer excellent performance, and give you both ADC inputs and DAC outputs. This allows you to use everything like a preamplifier in your audio system, plus you can do DSP processing if you would like.

 

[magicwolf]

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.

 

[HarmonicTHD]

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.

Proof it please. See above. Your claim is highly unlikely and therefore only facts can tell.

 

[magicwolf]

Proof it please.

Unfortunately, I know only one way to prove it - joint listening. Alas, this is impossible. You can also be curious and repeat my experience yourself. Perhaps you will succeed.

 

[HarmonicTHD]

Unfortunately, I know only one way to prove it - joint listening. Alas, this is impossible. You can also be curious and repeat my experience yourself. Perhaps you will succeed.

Na sighted listening, joint or not won’t do it. The results are always influenced by human expectation bias. Look around the forum for ABX testing.

Electronically there is no reason why there would be an audible difference (see above). Therefore personally I don’t waste my time chasing ghosts.

 

[magicwolf]

Electronically there is no reason why there would be an audible difference (see above).

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.