John,
while you are mainly referring to OS and apps, do you think a stripped down board like the ODROID-HC1 could be a good candidate as a renderer?
I've bought it to make a NAS with it, but it's simplicity made me think it could be very well used as a renderer.
Do you know of any way to get Spotify to work trough minimserver?
Hi swak, thank you for the confirmation regarding the 216j. I knew it was working. The 216se has a lot less processing power, that is why I asked. I'd like to get the SE, but only if it could do 384K.
I am using HiFi Cast on my Android in order to get data from my OMV server to my Pi. It works pretty quick and fine.
Best
Ernst
Ernst, I was speaking of the one bay DS112j, which has an older core 1 GHz CPU, so its processing power should be similar. Transcoding to wav24;348 works on the 112j but it takes a few seconds to begin each song. I am exclusively transcoding to wav24;96 and it works without that delay. Do you need two bays?
Hi swak,
oh, I'd misread the number. Thank you for pointing and clarifying. Hm, I do not really need two bays, but I'd like to have the opportunity of time driven switching on and off. This is not supported by the current single bay product, the 115j. So I'll probably go for the 216se.
Cheers
Ernst
I use Syno 216 J : tried both Audiostation player (ArchLinux) and the Squeezebox server Syno package (which is converting all material in 24 bits!!!!)
It's slow when it comes to change materials in your Library and still has smps powersupply which is bad when it feeds through USB output wire your DAC (no Pi 3 between the Syno and the Dac). Better to swap the Syno for a direct Odroid C1+ (less Noisy than a Pi 3) with direct HDD 2.5 or usb 2.5 HDD. I believe spidf despite ground isolation with Toslink is still a multiplexed so has jitter whatever I2S or not at the end (what is seing Mosaic front end input ?) .
It's slow when it comes to change materials in your Library and still has smps powersupply which is bad when it feeds through USB output wire your DAC (no Pi 3 between the Syno and the Dac). Better to swap the Syno for a direct Odroid C1+ (less Noisy than a Pi 3) with direct HDD 2.5 or usb 2.5 HDD. I believe spidf despite ground isolation with Toslink is still a multiplexed so has jitter whatever I2S or not at the end (what is seing Mosaic front end input ?) .
Hi Henderson,
USB noise is mainly caused by the short data bursts (frames) that occur either 1000 times / second (UAC1) or 8000 times / second (UAC2). This interference can be easily measured with a scope. The more signal changes per second (higher sample rates, higher bit depth) the higher the interference power level. So interference power rises with increased bit depth and sample rate. There is a break even point where further increase in sample rate does not translate to any audible improvement.
Then we have noise on the USB bus power supply, this mainly depends on the power supply and electronic circuits used in the host. We only use the USB bus voltage for sensing using an opto-coupler in the Mosaic UV DAC. Some host noise still manages to get through this opto coupler (stray capacitance).
I tested that we need at least 1 mega Ohm series impedance to arrive at acceptable USB bus voltage noise crosstalk in the DAC. When using bus powered DACs we have the absolute certainty that the DAC will be flooded by USB bus voltage noise regardless of the clean-up circuits that are used. Then we can very clearly hear differences in sound quality when using different digital audio sources, different OS, different applications and different USB interlinks.
USB interlinks only transport data (1 and 0) and 5V power, seems easy and any decent USB interlink would guarantee reliable data transfer and approx. 5V bus voltage at the DAC. When we ignore USB interface noise this would be fine.
But there is considerable noise on the USB bus voltage and these unscreened USB bus voltage lines are stuffed into a cable together with the often unscreened USB data lines. If we are lucky the data signal pairs are twisted decently. With cheap USB interlinks the pairs are not even twisted. So with the conventional USB interlink construction crosstalk between both USB bus voltage and serial data lines is guaranteed, but hey, the construction is cheap and the data gets through so who cares?
Better USB interlink would have separate shielded twisted pairs for data and separate shielded twisted pairs for the USB bus voltage, something like this:
TUBULUS ARGENTUS - USB CABLE V2 - TUBULUS
This way we can prevent USB bus voltage noise to crosstalk to the USB serial data lines inside the USB interlink. So we end up with less USB bus voltage interference on the USB data lines.
The high price of this interlink is mainly related to the pure silver conductors that are moot in these kind of applications. The most important is the construction, and when using shielded twisted pairs from cheap networking cables one can DIY such an interlink for under $10.
USB A and B plugs can be found here:
Custom gold plated USB plugs for DIY (pair of type A / type B)
Connections are here:
http://i.stack.imgur.com/8c96U.png
Use one shielded twisted pair cable for D+ and D- (USB serial data lines).
Use a second shielded twisted pair cable for - and + (USB bus voltage).
One can use two separately shielded twisted pairs form a single S/FTP LAN cable or use two pieces of F/UTP LAN cable (using one twisted pair of each cable)
Design of Screens and Shields - Screened and Shielded Network Cabling Standards Guide, By Siemon
The shield is connected to the shield of the USB plugs.
These separately shielded USB interlinks are most effective with self powered DACs like the Mosaic UV for example where the USB bus voltage is only used for sensing.
The Mosaic DAC USB audio receiver consists of a single 32 bit low power micro controller (300mW under full load) with on-chip USB interface. All Mosaic DAC timing signals for USB, CPU, and DAC are derived from one single masterclock.
We support UAC1 only (low speed, full speed), so the maximum sample rate equals 96 KHz. We did this because we have lowest interference at only 1000 frames / second (UAC2 uses 8000 frames / second) and because the break even point where further increase in sample rate won't result in audible improvement of perceived sound is below 96 KHz.
We use isochronous adaptive in order to minimise CPU load (interference) on both source and DAC, and to obtain most steady USB data stream (and related interference spectrum). The converters are clocked directly from the local masterclock.
The digital audio interface is a native multi-stream interface that consists of 8 serial data lines and two timing signals (total of 10 I/O pins on the chip):
DLL
DLM
DLH
DPOL
DRL
DRM
DRH
RPOL
BCK
LE
All stereo data is clocked into the converters using 8 BCK pulses only, after that the interface is completely shut down until the next sample data load.
This multi-stream interface connects directly to both Mosaic converters. The multi-stream interface offers up to 87.5% interface noise reduction compared to I2S and (isolated) S/PDIF. We experimented with both S/PDIF and or I2S for over a decade and finally gave up on these antique interfaces that were preventing us from making any significant progress.
With I2S we need 64 pulses to clock in this stereo data and the stream doesn't pause between samples, so we have maximum interference during the entire sample period all the time.
With S/PDIF we only have a single serial data stream and we have to add timing information to that same serial data stream. So here we also have continuous interference plus high jitter exposure since we mix both data and timing signals.
Even if we manage to re-generate a low jitter clock with a local masterclock, the jitter spectrum still enters the DAC through the (isolated) serial data signal since we need certain bandwidth for this signal to enter the DAC. So we still end up with an interference spectrum that's related to the jitter spectrum produced by the S/PDIF interface
USB noise is mainly caused by the short data bursts (frames) that occur either 1000 times / second (UAC1) or 8000 times / second (UAC2). This interference can be easily measured with a scope. The more signal changes per second (higher sample rates, higher bit depth) the higher the interference power level. So interference power rises with increased bit depth and sample rate. There is a break even point where further increase in sample rate does not translate to any audible improvement.
Then we have noise on the USB bus power supply, this mainly depends on the power supply and electronic circuits used in the host. We only use the USB bus voltage for sensing using an opto-coupler in the Mosaic UV DAC. Some host noise still manages to get through this opto coupler (stray capacitance).
I tested that we need at least 1 mega Ohm series impedance to arrive at acceptable USB bus voltage noise crosstalk in the DAC. When using bus powered DACs we have the absolute certainty that the DAC will be flooded by USB bus voltage noise regardless of the clean-up circuits that are used. Then we can very clearly hear differences in sound quality when using different digital audio sources, different OS, different applications and different USB interlinks.
USB interlinks only transport data (1 and 0) and 5V power, seems easy and any decent USB interlink would guarantee reliable data transfer and approx. 5V bus voltage at the DAC. When we ignore USB interface noise this would be fine.
But there is considerable noise on the USB bus voltage and these unscreened USB bus voltage lines are stuffed into a cable together with the often unscreened USB data lines. If we are lucky the data signal pairs are twisted decently. With cheap USB interlinks the pairs are not even twisted. So with the conventional USB interlink construction crosstalk between both USB bus voltage and serial data lines is guaranteed, but hey, the construction is cheap and the data gets through so who cares?
Better USB interlink would have separate shielded twisted pairs for data and separate shielded twisted pairs for the USB bus voltage, something like this:
TUBULUS ARGENTUS - USB CABLE V2 - TUBULUS
This way we can prevent USB bus voltage noise to crosstalk to the USB serial data lines inside the USB interlink. So we end up with less USB bus voltage interference on the USB data lines.
The high price of this interlink is mainly related to the pure silver conductors that are moot in these kind of applications. The most important is the construction, and when using shielded twisted pairs from cheap networking cables one can DIY such an interlink for under $10.
USB A and B plugs can be found here:
Custom gold plated USB plugs for DIY (pair of type A / type B)
Connections are here:
http://i.stack.imgur.com/8c96U.png
Use one shielded twisted pair cable for D+ and D- (USB serial data lines).
Use a second shielded twisted pair cable for - and + (USB bus voltage).
One can use two separately shielded twisted pairs form a single S/FTP LAN cable or use two pieces of F/UTP LAN cable (using one twisted pair of each cable)
Design of Screens and Shields - Screened and Shielded Network Cabling Standards Guide, By Siemon
The shield is connected to the shield of the USB plugs.
These separately shielded USB interlinks are most effective with self powered DACs like the Mosaic UV for example where the USB bus voltage is only used for sensing.
The Mosaic DAC USB audio receiver consists of a single 32 bit low power micro controller (300mW under full load) with on-chip USB interface. All Mosaic DAC timing signals for USB, CPU, and DAC are derived from one single masterclock.
We support UAC1 only (low speed, full speed), so the maximum sample rate equals 96 KHz. We did this because we have lowest interference at only 1000 frames / second (UAC2 uses 8000 frames / second) and because the break even point where further increase in sample rate won't result in audible improvement of perceived sound is below 96 KHz.
We use isochronous adaptive in order to minimise CPU load (interference) on both source and DAC, and to obtain most steady USB data stream (and related interference spectrum). The converters are clocked directly from the local masterclock.
The digital audio interface is a native multi-stream interface that consists of 8 serial data lines and two timing signals (total of 10 I/O pins on the chip):
DLL
DLM
DLH
DPOL
DRL
DRM
DRH
RPOL
BCK
LE
All stereo data is clocked into the converters using 8 BCK pulses only, after that the interface is completely shut down until the next sample data load.
This multi-stream interface connects directly to both Mosaic converters. The multi-stream interface offers up to 87.5% interface noise reduction compared to I2S and (isolated) S/PDIF. We experimented with both S/PDIF and or I2S for over a decade and finally gave up on these antique interfaces that were preventing us from making any significant progress.
With I2S we need 64 pulses to clock in this stereo data and the stream doesn't pause between samples, so we have maximum interference during the entire sample period all the time.
With S/PDIF we only have a single serial data stream and we have to add timing information to that same serial data stream. So here we also have continuous interference plus high jitter exposure since we mix both data and timing signals.
Even if we manage to re-generate a low jitter clock with a local masterclock, the jitter spectrum still enters the DAC through the (isolated) serial data signal since we need certain bandwidth for this signal to enter the DAC. So we still end up with an interference spectrum that's related to the jitter spectrum produced by the S/PDIF interface
Thank you ECDESIGNS,
I remember the 5V bus of USB were made also to power devices as a 2.5 HDD.
Does the Mosaic need this 5V to receive the datas (is there common GND between data bus and 5V bus ?) by powering the USB input stage of the Mosaic (usb receiver chip(s) ) ?
Or one can cut himself the 5V line of a cheap USB wire to avoid all the bad things you described ? And keep the GND 5V wire for shielding behavior?
Should we add ferrite beads around the cable (specific one?)
regards
I remember the 5V bus of USB were made also to power devices as a 2.5 HDD.
Does the Mosaic need this 5V to receive the datas (is there common GND between data bus and 5V bus ?) by powering the USB input stage of the Mosaic (usb receiver chip(s) ) ?
Or one can cut himself the 5V line of a cheap USB wire to avoid all the bad things you described ? And keep the GND 5V wire for shielding behavior?
Should we add ferrite beads around the cable (specific one?)
regards
USB wires picking up noise ?
Hello ECDSIGNS,
I tested my USB wires :
Externals plugs (so external cable shield) are not tied to the - V bus (Gnd).
The - V bus (Gnd) seems needed as a return path of the symetrical twisted data lines (+D,-D).
I read than most host devices (our PC, NAS...) have their USB plug grounded to Earth (so the switched power supply and metal cabinet.
My understanding is the USB shield is more to prevent this cable to emit EMI than protect it from external HF fields. But this path should be also a highway to the host noise (SMPS...) towards our delicate DACs !
I assume the - V bus (Gnd return path of the 5V ) on the host side & Dacs is direct to the main grounds of the Pcbs as well.
Is it efficient to cut on the receiver devise side the shield from the USB model B receiver plug to avoid the picked up noise from the source to ravel through the sield ? Or it doesn't matter as the -V bus (USB Gnd) will be not isolated as well ?
I assume the shield has to be grounded on the host side to precent the shielding to be an antenna and pick up also the suroundings noise ? (cellulars ? EMI from the others hifi/pc devices ?).
Saw some are coupling the external wire shielding to the ground (via USB B receiver plug) through C ((1000 pF) or RC...
What are your though about the way to isolate at best the DACs from the host (playback streamer) ? Did you make Something like that (as I understand now from the previous Mosaic DAC does need the V+ bus for sensing) ?
thanks,
best regards
Hello ECDSIGNS,
I tested my USB wires :
Externals plugs (so external cable shield) are not tied to the - V bus (Gnd).
The - V bus (Gnd) seems needed as a return path of the symetrical twisted data lines (+D,-D).
I read than most host devices (our PC, NAS...) have their USB plug grounded to Earth (so the switched power supply and metal cabinet.
My understanding is the USB shield is more to prevent this cable to emit EMI than protect it from external HF fields. But this path should be also a highway to the host noise (SMPS...) towards our delicate DACs !
I assume the - V bus (Gnd return path of the 5V ) on the host side & Dacs is direct to the main grounds of the Pcbs as well.
Is it efficient to cut on the receiver devise side the shield from the USB model B receiver plug to avoid the picked up noise from the source to ravel through the sield ? Or it doesn't matter as the -V bus (USB Gnd) will be not isolated as well ?
I assume the shield has to be grounded on the host side to precent the shielding to be an antenna and pick up also the suroundings noise ? (cellulars ? EMI from the others hifi/pc devices ?).
Saw some are coupling the external wire shielding to the ground (via USB B receiver plug) through C ((1000 pF) or RC...
What are your though about the way to isolate at best the DACs from the host (playback streamer) ? Did you make Something like that (as I understand now from the previous Mosaic DAC does need the V+ bus for sensing) ?
thanks,
best regards
Hi Henderson,
The Mosaic UV needs the 5V USB bus voltage for host detection, 5V USB bus voltage is not used for powering the USB receiver (self powered receiver).
Cutting the 5V wire would prevent the Mosaic UV from connecting with the host (no host detection signal).
Yes, because of wireless technology and SMPS that produce high frequency interference it is mandatory to prevent these high frequencies from interfering with the audio equipment. This can be tested with suitable measuring equipment like these:
YouTube
We can use clip-on ferrites on all power supply interlinks, digital interlinks and all audio interlinks (including the speaker interlinks).
Clip-on ferrite beads are most effective when placed as close as possible at the input, so the exposed part of the interlink between ferrite bead and input is as short as possible.
The Mosaic UV needs the 5V USB bus voltage for host detection, 5V USB bus voltage is not used for powering the USB receiver (self powered receiver).
Cutting the 5V wire would prevent the Mosaic UV from connecting with the host (no host detection signal).
Yes, because of wireless technology and SMPS that produce high frequency interference it is mandatory to prevent these high frequencies from interfering with the audio equipment. This can be tested with suitable measuring equipment like these:
YouTube
We can use clip-on ferrites on all power supply interlinks, digital interlinks and all audio interlinks (including the speaker interlinks).
Clip-on ferrite beads are most effective when placed as close as possible at the input, so the exposed part of the interlink between ferrite bead and input is as short as possible.
Well, wouldn't it make sense to go the extra mile and use a dual head cable like this one:
TUBULUS ARGENTUS – USB CABLE V2 DUAL HEAD
That way the 5V voltage can be provided separately from a clean source, e.g. a battery, eliminating the noise from the USB bus power supply but still providing a signal for host detection. Or am I making a mistake here...?
TUBULUS ARGENTUS – USB CABLE V2 DUAL HEAD
That way the 5V voltage can be provided separately from a clean source, e.g. a battery, eliminating the noise from the USB bus power supply but still providing a signal for host detection. Or am I making a mistake here...?
My understanding is if the host is self powered the V Bus with its 2.5 mA shouldnot have big crosstalk towards the D+;D- bus which has its own noise cancelation- being symetric and twisted-. But the shield is handled to the ground of theSMPS becomes a big Gnd bus as the the internal -V bus made for ground purpose only(return path).
Sothe shield becomes a highway for SMPS of the host towards audio device and alsoa peakup antenna (EMI, ESD) of the external noise (cellulars, TV…)! in the USB datasheetthe shield is not made for audio devices at the origin but for devices as mouses,keyboard, webcams, etc : external wooden shield is only to prevent the USBcable to radiate EMI…
Stranglymost of the manufacterer may protect the -V bus & the +5V bis by serieferite beads before the USB outplug ! Strangly because body of the plug for EMIpurpose is still handled by the main ground of the host ; hence the smps noisepath from the host towards the audio device via the external wooded shield!
Bestbet should be to cut the wooded external shiel of a 2 euros USB cable near thereceiver side (end of the USB cable) : or isolate the receiver of the audiodevice plug from its own ground or at least put also a ferrite bead between themetal receiver plug and the ground of the device : mainly because the -V bus isalso tied to the ground near the plug and is a noise path to the ground of thehost device (PC, NAS) despite the hot manufacterers puts ferrite beads on thegnd bus of the –V line in the USB cable ! (what to waste what the hostmanufacterers do for the best maker on their PC or NAS as we are audioparanoids, humm ?!
Asin the scenario of a splitted cable, my basic understanding is the data busneeds a close ground reference (hence the –V bus gnd cable) because it's a highspeed signal, despite basicly no packets of data seems lost...and noise is alsocanceled due to the symetric diferential signal (+&- Data bus) …. despitethere are some single ended signals in the frame (SOE signal, etc).
Atthe lowest speed a USB cable should not be longer than 3 meter (and 5 metersfor full speed operation), but the longer the cable the longer the antena asthe shield is tied at the two plugs at the ends of the cables (hence are tiedto both host and device grounds...)!
Noticethan if the audio brands isolated the receiver USB plug from the ground youstill had a shield noise path if the Gnd line (-V bus) has no ferrite bead at thebeginning of this line and stays still directly grounded : so maybe in theaudio industry the Gnd wire – V bus should have a ferrite bead as well betweenthe USB plug – V bus pin an the ground : of course if only the B plug bodyis isolated to the ground or also filtered by this same ferrite beads !
Thisis my basic understanding ! But does it matter or is listenable ? Idon’t know !
inthe scenario of a diy usb cable I will putt the -V Bus near the datastwisted lines as a reference for high speed signals and put the +V Bus alonebut farer to avoid crosstalk despite in the case of the Mosaic, the +5V V Bushas only max 2.5 mA current ! I will not handle the external shiel on thereceiver side B plug as well ! But it makes sense imho only if the host is selfpowered as the Mosaic is or the good USB to I2S board 'whom have unluckillytheir B plug directly tied to the ground certainly without ferrite beadsbetween the body and the local ground.... hence the need to not tied theexternal wooden shield near the device... and only if we have the luck than theaudio device manufacter putt a ferite bead between the receiver plug of hisdevice and the close ground - not the case on my Wave I/O for instance !
Thereare also usb isolator chips with galvanic isolation if you isolate the receiverB plug from the ground, but their are made if i understood in the case theground potential of the host is different from the ground of the device... andthis is not cheap must be powered at both side, complicate things !)
Thankyou at ECDSIGN to advert us on the complicate subject of USB audio sources,didn't be aware before reading this thread... just believed asynchronous wasenough, but no, more can be done if I understand (as the need to be the clockof the micrp controller far from the USB plug data lines, etc) what I read andwhat ECEDESIGNS wrote.
somelinks : https://forum.allaboutcircuits.com/threads/usb-device-cable-shield-connection-grounding-it-or-not.58811/
http://www.ti.com/lit/an/spraar7g/spraar7g.pdf
Shielding :
Shieldshould only be connected to Ground at the host. No device should connect Shieldto Ground. : http://www.hardwarebook.info/Universal_Serial_Bus_%28USB%29#Cable
TIseems to doesn't care of shielding to respect EMI norms but they are not audiosided : http://www.ti.com/sc/docs/apps/msp/intrface/usb/emitest.pdf
soas you see : even manufacters disagree between them cause mainly audio devicesneeds more care...maybe !
And the expensive USB cable still have their shield handled to the receiver side so stays no more good than 2 euros ones, imho ! but sure the splitted cable is attractive as not sourcable at 2 euros, lol (even diy should cost you 20 euros for the less with 100 ohms ethernet cables ; believe USB is 90 ohms perhaps but 10 ohms difference for short length certainly doesn't matter : again no packets are lost but sometimes you may here clicks with usb NAS or PC : not present at the same timing if you replayback the song : so clicks are peaked up or came from the ground noise of the host certainly ! the less Noisy source also the better , hence teh Synology NAS despite mine still has a 6A smps powers supply : but certainly quieter than media PC)
regards,
Last edited:
The idea of the external battery on the side of the host is beautifull; but maybe the data bus still need the Gnd retutn path (-V bus) because there is a discontinued ground and high speed signals need ground reference : void are bad for them !)
but dunno, one has to make a poc and listen or ask to high frequency engineer maybe ?
cutting the shield at receiver side should suffice in case of self powered audio device and short USB cable (the shorter the better)
seems our USB to I2S devoces may be improved yet
: as the close micro controller xtal of the host and the one of the audio device whom radiate towards USB data bus; if too close should be shielded as well with a hat and so far ?regards
Last edited:
My work situation has hindered me from trying out my Mosaic DAC, but I have tried to read a little bit about EC's preferred Raspberry set up. One thing I don't understand is the purpose of the Wifi Router in the block diagram. What does it do? Mosaic UV + RaspberryPi 3 Model-B | Support Why isn't the NAS connected directly to the Raspberry?
From my tests a Linux NAS à la Synology is way better by its USB outputt than a Pi ! More of all : detail, soundstage, body & articulation, dynamic... Like to switch from a bad to a good dac ! (even if the low end is still a good dac !)
Despite I feeded the Pi with a battery and I took the I2S outputt of it via the IanCanada GPIO to uf-l plug header ! So I used 4" good uf-l cable !
Despite all this care, the Synology USB outputt of the Synology is way better ! This is more hearable than isolated RJ45 cable from the provider box imo ! What the hell to use wifi to avoid ground loop if the Pi is giving a crap sound
That's why i asked if the USb noise should not be optimised even more with précautions on the earthed wirered shield around the USB cable which are not isolated !
Dead silence to my questions ?
Is it perhaps a bad question or nobody made tests on this subject yet(foccusing just around this splitted +5V/Gnd wires from the twisted Data +/- pair... is is a myth, especially if the device is seld powered as we already do from a time now with async USB to I2S pcbs ?)
No suffisant to cut the shield on the device side if the - V Bus (gnd) is not isolated by an smd feritt bead after the B plug or before in the cable by a round feritt bead ?
It may Worth the try : an USB A to B plug cable is 3 euros...
QUestion to me know is : does a Synology USB outputt is better than the Pi used via its Noisy USB outputt ? I assume if the I2S of the Rpi is bad, the USB should not be better ????
Maybe not the good thread as it's more a Mosaic thread since ECdesigns has given up the TDA1541A chip!
Please ECdesign : make us a good real connoisseur TDA1541A or TDA1543 DIY board please First better than the less good TDA1543 (for ones whom can not afford a Mosaic yet or to wait in between :good advert for your fantastic savoir faire if sound good)
Despite I feeded the Pi with a battery and I took the I2S outputt of it via the IanCanada GPIO to uf-l plug header ! So I used 4" good uf-l cable !
Despite all this care, the Synology USB outputt of the Synology is way better ! This is more hearable than isolated RJ45 cable from the provider box imo ! What the hell to use wifi to avoid ground loop if the Pi is giving a crap sound
That's why i asked if the USb noise should not be optimised even more with précautions on the earthed wirered shield around the USB cable which are not isolated !
Dead silence to my questions ?
Is it perhaps a bad question or nobody made tests on this subject yet(foccusing just around this splitted +5V/Gnd wires from the twisted Data +/- pair... is is a myth, especially if the device is seld powered as we already do from a time now with async USB to I2S pcbs ?)
No suffisant to cut the shield on the device side if the - V Bus (gnd) is not isolated by an smd feritt bead after the B plug or before in the cable by a round feritt bead ?
It may Worth the try : an USB A to B plug cable is 3 euros...
QUestion to me know is : does a Synology USB outputt is better than the Pi used via its Noisy USB outputt ? I assume if the I2S of the Rpi is bad, the USB should not be better ????
Maybe not the good thread as it's more a Mosaic thread since ECdesigns has given up the TDA1541A chip!
Please ECdesign : make us a good real connoisseur TDA1541A or TDA1543 DIY board please
First better than the less good TDA1543 (for ones whom can not afford a Mosaic yet or to wait in between :good advert for your fantastic savoir faire if sound good)The suggested setup is wired to the Pi (note that wifi is disabled in the Pi configuration). A router with wifi is enabled to use a tablet/smartphone as a control point.
Notice an other wifi router can also reduce ground loops and the noise comming from the provider box:
shortly : little 5 v wifi router with a clean powersupply to minimize noise (no Noisy smps) -> short RJ45 ethernet cable to the Rpi -> USB outputt to "Mosaic"
.
the data are sent by the provider box (all have wifi those days) towards this low power wifi router. And the remote control (tablett, smartphones, laptop) is still Wireless...
NAS is linked to the provider box or HDD via usb to it as it has internal DNLA protocol as well.
soundcheck's - audio@vise: networking - my audio data highway
shortly : little 5 v wifi router with a clean powersupply to minimize noise (no Noisy smps) -> short RJ45 ethernet cable to the Rpi -> USB outputt to "Mosaic"
.
the data are sent by the provider box (all have wifi those days) towards this low power wifi router. And the remote control (tablett, smartphones, laptop) is still Wireless...
NAS is linked to the provider box or HDD via usb to it as it has internal DNLA protocol as well.
soundcheck's - audio@vise: networking - my audio data highway
Last edited: