Devialet Chat

(06-Sep-2018, 13:12)Confused Wrote: [ -> ]The NB will of course work fine without any external clocks, but if you want to use the word clock inputs, it is best to have two feeds using 44.1kHz and 48kHz family rates respectively, in which case I think you would need two MC3+USB's.

Thx for the response. Why would it be best to use one feed per family rate? (pls excuse the Ref10 newbie question)

(06-Sep-2018, 19:10)baconbrain Wrote: [ -> ]

(06-Sep-2018, 13:12)Confused Wrote: [ -> ]The NB will of course work fine without any external clocks, but if you want to use the word clock inputs, it is best to have two feeds using 44.1kHz and 48kHz family rates respectively, in which case I think you would need two MC3+USB's.

Thx for the response. Why would it be best to use one feed per family rate? (pls excuse the Ref10 newbie question)  

Because that is how word clocks are designed to work.  The idea behind word clocks is that they synchronise the clocks through a system.  In addition, a high quality master clock can reduce jitter in addition to providing full synchronisation.  So in a full dCS rig, the master clock might be connected to both the streamer and the DAC.  If you are running 44.1kHz material, then the 44.1kHz clock feed will be used, and exactly the same clock is sychronised between the DAC and the source.  If you switch to 48kHz material (or 96kHz / 192kHz), the 44.1 clock feed will not work, you need one in the 48kHz family.  For this reason, the Network Bridge has two clock input connections, one for each clock rate family.  In practical terms, if you feed the Network Bridge a 48kHz word clock feed, and then play 44.1kHz material, the Network Bridge would ignore the 48kHz feed from the Mutec kit, it cannot use it, and would default to it's internal clock.  This is why upsampling in Roon can mitigate this issue, you could send the Network Bridge a 192kHz feed irrespective of the native rate of the source material, so only the 48kHz family word clock feed would be needed. 

This is a different concept to the REF10 reference clock idea.  The REF10 (and similar products from other manufacturers) provide a 10MHz reference clock signal.  This should improve the clock accuracy of the device to which you feed it, and devices like the SOtM Ultra range and the Mutec MC3+USB can accept the 10MHz signal, and use it for whatever sample rate music you might want to listen to.  It is in this way that the MC3+USB can used with the REF10 to produce a very accurate word clock feed, this can be sent to the Network Bridge, the MC3+USB can accept the 10MHz feed and has a clock board that can generate the 44.1 or 48kHz frequencies.  The sCLK-EX board in the SOtMultra devices works in a similar way, it can accept a 10MHz reference feed, and can distribute an accurate clock feed at a variety of frequencies, to suit a streamer, a network switch, or a whole range of other things.

When used with a Devialet, you cannot synchronise the clock in the DAC as you would in a dCS rig, but if the word clock is accurate enough, it should in theory reduce the jitter in the Network Bridge and improve the performance.  I say in theory, the end result would depend on losses in the clock cables and how good the built-in clock is, the better it is, the harder it will be to improve upon.  The built in clock has the advantage of being directly connected to whatever is using the clock, with no cable losses.  10MHz clock cables can suffer losses and interference, which goes some way to explain why the REF10's performance seems to be dependant on the quality of cables, far more than you would expect compared to S/PID or USB cables.

Page 25 of the Network Bridge manual provides some clarity with respect to the word clock inputs:

https://dcsltd.a.cdnify.io/wp-content/up...-v1_0x.pdf

Or you could just use AIR, not worry about this nonsense and enjoy a tune or two.

Quote:Thinking about all this makes me realise that my current "spaghetti" solution is actually quite simple in use. The MC3+USB and SOtM kit all accept a 10MHz reference clock feed, you can feed the sMS-200Ultra any sample rate you want, and it just works. The SOtM kit also works with HQPlayer, which the NB does not.

This is perhaps the biggest issue to me that it becomes lots of boxes and cables. I really do like my SOtM kit and the Paul Hynes PSU but I just cant stop thinking about making it easier.

So I am thinking of replacing my current dac streamer solution with a dCS Rossini dac and replace five boxes with one and keeping it all in one box and much less cable to fiddle with. Moving the PSU to power the switch and NUC or just sell them with the SOtM kit.

(07-Sep-2018, 07:58)Confused Wrote: [ -> ]...
Or you could just use AIR, not worry about this nonsense and enjoy a tune or two.

I must say that was what was going through my head as I read this...  

(07-Sep-2018, 09:34)thumb5 Wrote: [ -> ]

(07-Sep-2018, 07:58)Confused Wrote: [ -> ]...
Or you could just use AIR, not worry about this nonsense and enjoy a tune or two.

I must say that was what was going through my head as I read this...  

I was thinking exactly the same. But, it wouldn't come to my mind to use words "AIR" and "enjoy" in the same sentence... My experience with AIR hasn't been particularly enjoyful  

The SOtM kit and NB both look great alternatives to the AIR. Thank you for sharing your experiences with SOtM kit and Mutec @Confused

(07-Sep-2018, 08:31)octaviars Wrote: [ -> ]

Quote:Thinking about all this makes me realise that my current "spaghetti" solution is actually quite simple in use. The MC3+USB and SOtM kit all accept a 10MHz reference clock feed, you can feed the sMS-200Ultra any sample rate you want, and it just works. The SOtM kit also works with HQPlayer, which the NB does not.

This is perhaps the biggest issue to me that it becomes lots of boxes and cables. I really do like my SOtM kit and the Paul Hynes PSU but I just cant stop thinking about making it easier.

So I am thinking of replacing my current dac streamer solution with a dCS Rossini dac and replace five boxes with one and keeping it all in one box and much less cable to fiddle with. Moving the PSU to power the switch and NUC or just sell them with the SOtM kit.

That was my motivation when my Trinnov became Roon Ready. I sold my mR, power supply, cables, mutec mc-3+ and just streamed from my Roon Server via ethernet. Initially I was quite happy with the decision but now over time I have the feeling that there is something missing .... go figure

(07-Sep-2018, 07:58)Confused Wrote: [ -> ]Because that is how word clocks are designed to work.  The idea behind word clocks is that they synchronise the clocks through a system.  In addition, a high quality master clock can reduce jitter in addition to providing full synchronisation.  So in a full dCS rig, the master clock might be connected to both the streamer and the DAC.  If you are running 44.1kHz material, then the 44.1kHz clock feed will be used, and exactly the same clock is sychronised between the DAC and the source.  If you switch to 48kHz material (or 96kHz / 192kHz), the 44.1 clock feed will not work, you need one in the 48kHz family.  For this reason, the Network Bridge has two clock input connections, one for each clock rate family.  In practical terms, if you feed the Network Bridge a 48kHz word clock feed, and then play 44.1kHz material, the Network Bridge would ignore the 48kHz feed from the Mutec kit, it cannot use it, and would default to it's internal clock.  This is why upsampling in Roon can mitigate this issue, you could send the Network Bridge a 192kHz feed irrespective of the native rate of the source material, so only the 48kHz family word clock feed would be needed. 

This is a different concept to the REF10 reference clock idea.  The REF10 (and similar products from other manufacturers) provide a 10MHz reference clock signal.  This should improve the clock accuracy of the device to which you feed it, and devices like the SOtM Ultra range and the Mutec MC3+USB can accept the 10MHz signal, and use it for whatever sample rate music you might want to listen to.  It is in this way that the MC3+USB can used with the REF10 to produce a very accurate word clock feed, this can be sent to the Network Bridge, the MC3+USB can accept the 10MHz feed and has a clock board that can generate the 44.1 or 48kHz frequencies.  The sCLK-EX board in the SOtMultra devices works in a similar way, it can accept a 10MHz reference feed, and can distribute an accurate clock feed at a variety of frequencies, to suit a streamer, a network switch, or a whole range of other things.

When used with a Devialet, you cannot synchronise the clock in the DAC as you would in a dCS rig, but if the word clock is accurate enough, it should in theory reduce the jitter in the Network Bridge and improve the performance.  I say in theory, the end result would depend on losses in the clock cables and how good the built-in clock is, the better it is, the harder it will be to improve upon.  The built in clock has the advantage of being directly connected to whatever is using the clock, with no cable losses.  10MHz clock cables can suffer losses and interference, which goes some way to explain why the REF10's performance seems to be dependant on the quality of cables, far more than you would expect compared to S/PID or USB cables.

Page 25 of the Network Bridge manual provides some clarity with respect to the word clock inputs:

https://dcsltd.a.cdnify.io/wp-content/up...-v1_0x.pdf

Or you could just use AIR, not worry about this nonsense and enjoy a tune or two.

Thanks for the excellent response!

I have had the Neo version of my sMS-200Ultra in my system for about a month now. As I posted earlier, my initial impressions of the Neo were very positive, I definitely prefer it to the original version. With the system back into what I consider a good balance, my plan was to stop tinkering and simply enjoy the music. This is exactly what I have been doing through September, but as often happens, it is when you stop A/B testing and fiddling with the kit, that you start to notice things. Overall, the system is sounding great, the SOtM / Mutec kit is providing a far more balanced presentation, details, dynamics, the lot. Things are really coming together now. Just one little niggle. I had one weekend in early September where I had some decent listening sessions, I listened to a variety of music, and was just plain happy that things were sounding good, the Neo was definitely improving things. A couple of weeks later, another weekend with some decent listening time, and something was bothering me. With some tracks I was again finding the treble a little harsh, maybe even distorted. It seamed to depend on what I was listening to, so I did wonder if it was just a case of issues the source material, but it seemed a little strange that sometimes all sounded good, other times not so good. Last week something occurred to me, I was happy with the system when back with the microRendu, this was during August when the SOtM kit was back in Korea for the Neo upgrade and clock cable mods. During this time I tried the mR with the SOtM sPS-500 power supply, I found that I defiantly preferred the mR powered by the Uptone LPS-1, but the point is that it also occurred to me that the reason I did not like the sPS-500 providing the power was a sense of some slight distortion and HF harshness, exactly the issue I am now finding with the Neo.

So on Sunday I tried an experiment. In fact, this is re-trying something I did back in the Spring. I took the sPS-500 out of the system I powered the sMS-200Ultra and tX-USBultra from a 12 volt rechargeable battery pack that I have. Now this is not "high end" kit, it is a £20 battery pack with a cheap and spindly Y cable, the DC cable alone for the sPS-500 is many times more expensive. When I tried this before, in an attempt to tame the overly bright pre-Neo sMS-200ultra, it seemed to make little difference. With the more balanced sounding Neo, the difference seemed very clear to me, the HF distortion and harshness that was troubling me seemed very much diminished. Later on Sunday I managed to persuade a friend to have a listen. We played three tracks, I switched the power supply to the battery and played the same tracks again. My friend had no idea what I was changing so this was a blind test. Their verdict was something along the lines of "they sound pretty much the same but I thought that 2 was a bit clearer". OK - this is hardly a statistically significant double blind test result, but it was significant to me that a third party, who had no prompting about my HF concerns, preferred a £20 battery pack to the £500 sPS-500. I left the SOtM kit on battery power for a while on Sunday, and I much enjoyed this as a set-up. The thing is, battery power is not considered the best for digital audio, these devices need a very low output impedance that battery power does not provide. Yet despite this, both myself and my clueless buddy preferred the SOtM kit with battery power. It is a case of something in the HF being a little niggling, just that little bit annoying, which then detracts from enjoyment of the rest.

I had a plan that I would simply enjoy the system for a while, then maybe try the dCS Network Bridge, depending on how that comparison worked out, I would then maybe get some Paul Hynes SR-4's if I had decided to stick with the SOtM kit. I have now changed my mind, if I do the comparison with the dCS NB now, I am sure that I will inevitably focusing on the HF distortion issue, so may prefer the dCS for this reason alone, whereas this is an issue of the power supply, not the sMS & tX. So I have decided to take the plunge, order the SR-4's now, and hope for the best! Although I am pretty confident this one will work out well. Interestingly, @octaviars proposed that I should go with SR-4's many months ago, now I am sure as I can be that he is right, soon I will find out for sure!

Nice to see that you are geting some SR4 @Confused they are really good.

I did find some tunes that got some distorted highs even with the Neo and SR4 but I changed my usb cables between sMS-200ultra to tX-USBultra to DAC and this removed it completely. 

From the tX-USBultra I use both usb outputs one for signal and one for 5v (the cable only have two conductors in each cable) so they are seperated to my dac. 

The cable are from Pangea and I bought them from Audiophonics in France.

[attachment=3105]

[attachment=3106]

@octaviars - a very interesting post, thanks. Do you know how USB input to the tX-USBultra works? Does it actually take it's power from the feeding device, or does it just need a 5V "handshake"?

I am definitely interested in trying this, quite intriguing. (and the prices for the Pangea look reasonable also)

As for the SR4's, I have been in touch with Paul Hynes, apparently the they are made in batches of 10 which are released at the end of each month. They are now taking orders for the end November batch, October is sold out, so at best the SR4's will be an early Christmas gift to myself.