USB to I2S conversion and reclocking quality doubt

[One and a half]

On 2/21/2018 at 11:26 PM, doraymon said:

As a matter of fact I have the opposite problem, that is my upstream devices have supposedly better clocks than the D/D converter.

This is my chain: SOtM sMS-200ultra > SOtM tX-USBultra >(over USB)> Singxer KTE SU-1 DDR >(over I2S)> Holo Audio KTE Spring DAC L3

So the better clocks are upstream and the worst downstream...

Essentially the SU-1 is there to take care of the USB to I2S conversion (using its Crystek CCHD-575 clocks).

I could indeed skip the SU-1 and feed the the DAC with the USB signal out of the SOtM devices but than the DAC would in any case internally convert this to I2S, using its internal clocks which are much less precise than the 575's.

So the question is: am I wasting my money in placing $2.2k worth of "USB signal cleaning" devices before the SU-1?

I tried to do some tests by adding/removing the tX-USBultra but due to the limited space I have I can't really do a real time A/B comparison and difference was not immediately perceivable.

To add to the confusion I tried to remove the SU-1 and feed the good quality USB signal to the DAC and contrarily to the theory what I am hearing has slightly more detail, fuller bass and the sound is much less fatiguing.

 

I confess I am lost.

Sorry to be a little late in this reply.....despite best efforts, there's common mode noise on USB which unless removed will always be present.  So there's benefit of using USB cleaners on the output of any server system. Later I can find a pdf from TDK which illustrates, especially for USB Class 2.0 audio speeds at 480Mbs the effect of common mode filters.

I2S is usually transmitted over LVDS, whether it has great clocks or not another issue, since the implementation of LVDS varies greatly.

[One and a half]

On 2/24/2018 at 11:08 AM, doraymon said:

Thanks! I'm looking forward to it!

 

...and here I'm lost again...

Here's the paper on common mode noise and filtering. The filters are very small to implement, but not featured enough on USB interfaces. We all wonder why there are differences, here's one!

 

 

TDK USB COMMON MODE NOISE eemc_practice_04.pdf

[One and a half]

On 25/02/2018 at 6:44 PM, doraymon said:

I did a few tests by connecting the Mac mini running Roon Server directly to the SU-1 and than to the DAC against having the Mac connected to the Ethernet and than the sMS-200ultra connected to the USB of the Spring DAC.

 

The solution with the streamer (second) is significantly better, more detail, more air, space and soundstage depth.

 

Now the last question is , do I keep the Mac Mini close to the HFi rack connected to the same mains filter and the same Ethernet switch of the sMS 200or will it be better to move it somewhere else in the room to put its terrible switching power supply far from the rest of the system (but with longer Ethernet path)?

Moving the computer away from the DAC will have benefits.

[One and a half]

13 minutes ago, doraymon said:

For sure, the only drawback is that moving it away from the DAC I won’t be able to connect it directly to the AQ-SWITCH-SE where the network player is also connected.

I thought having both computer and Network Player  connected to the same high quality switch would be an advantage.

I'm not following the connection method and why it would be a problem.

This would work.

 

Network cable infeed -> AQ-Switch-SE

                                        -> AQ-Switch-SE -> SMS-Ultra -> DAC

                                        -> AQ-Switch-SE -> Network Player

                                         -> AQ-Switch-SE -> <= 100m Cable to Mac Mini

                                         -> AQ-Switch-SE -> Spare

                                         -> AQ-Switch-SE -> Spare

                                         -> AQ-Switch-SE -> Spare

                                         -> AQ-Switch-SE -> Spare

 

[One and a half]

10 hours ago, doraymon said:

This direct connection is possible only with the Mac sitting in the same cabinet with the hifi.

if I move it far from the cabinet I can’t have direct connection to the AQVOX due to the architecture of the room. In this case I plug the Mac mini to the main router.

 

just to clarify the AQVOX switch is connected to the Gigabit Port of a network extended placed close to the HiFi.

OK. The mini can be mounted next to the router. The bits will still get through, so will the common mode noise. The level of attenuation depends on the switch's performance, for EUR750, it would want to be exceptional and remove any common mode noise to several MHz.