the way that ethernet transports data vs usb transports data made me think that there could be a different SQ wise using a clock.
In this case with he DAC being set to “A” we are talking about synch. That is when the 44.1 or 48 KS/s “cycles” of the DAC and source are identical. That is that both use the correct sample rate frequency and both are also aligned. Hence the need for the source and DAC to be slaved either from the same wordclock or where the source becomes the slave of the DAC. In these latter cases synch is set to or Wordclock* or Master . Further If a USB or ethernet connection is used where there is no wordclock connected to the source with a dCS DAC then the interface will be asynchronous (achieving much the same result in this respect). This would therefore apply to streaming services where a USB or ethernet connection is employed from the source in dCS multi box configurations.
I may be wrong but I understood , however, that using a superior external wordclock not only allows for the synchronous operation of source and DAC but also contributes to its accuracy and stability by acting as reference for the internal DAC clock even if synchronisation of two connected components does not occur. Hence in part the tight accuracy specification of the external clock .Is my understanding incorrect then; it may well be? If so adding external clocks for example with a CD player without wordclcok input and using S/Pdif connection will make no difference ( many users seem to be doing just this with Bartok and third party CD players).
1 Like
This is correct. What the external clock provides is a very low phase noise source to which the DAC’s clock locks to.
If you set the clock for, say, SPDIF1 to wordclock, you will benefit from the same low phase noise as in the asynchronous case. The catch here is it is possible that the incoming SPDIF signal goes a bit too fast compared to the reference clock and eventually fills out the small buffer, or goes too slow and the buffer runs out. In those situations you would get a silence gap and then it will resume.
This is avoided in the dCS transports and network bridge by taking a clock signal from the master clock and clocking the entire chain in sync.
1 Like
Which actually makes me wonder that, if you had a Network Bridge or Vivaldi Upsampler and were running without a master clock, the source of the clock for the chain should be the NB or Upsampler.
Thanks but your answer really deals with synch issues. However what I am trying to establish is whether or not using an accurate external clock connected to the DAC affects accuracy per se rather than the relative timing of two components with a wordclock connection.
As a crude example, If the DAC’s internal clock actually works at 44.11KS/s and an external clock works at exactly 44.1 KS/s does connecting the clock to the DAC if the DAC mean that the DAC now operates at 44.1 or does it continue at 44.11?
The DAC will operate at 44.1KHz. However, the issue whether a DAC is running at 44.100001KHz or 44.1KHz is not really important. A tiny frequency change, which is what you’d get with today’s oscillators, is not critical, you can’t hear that. But drifting around you can. If that drifting in and out of a mean frequency happens 100 times a sec, then you will get +/- 100Hz bands on your output. The size of those sidebands depends on how much the frequency shifts around.
" Which actually makes me wonder that, if you had a Network Bridge or Vivaldi Upsampler and were running without a master clock, the source of the clock for the chain should be the NB or Upsampler."
The clock master should be the DAC in these circumstances as it should be sited as c]ose as possible to the point of conversion. Hence both NWB aand Vivaldi upsampler have clock outputs to allow this.
You mean clock inputs from the DAC?
No, outputs 
I don’t understand… If you connect a NB or a Vivaldi Upsampler to a Vivaldi DAC over AES, your clock will be by default the NB or Upsampler as the input defaults the clock source to “A”, which is recovery of the clock from the source. If you take a clock output out of the NB or US into a Vivaldi DAC, then again the clock master is the NB or US, just that you’d set the master clock in the DAC to be wordclock from the source. If on the other hand you’d be able to get a clock out of the Vivaldi DAC (there is an out but not sure how that works) and go INTO the NB or US, then the DAC would be the master clock.
That’s it. You use the wordclock output on the DAC and set synch to M. See e.g Vivaldi DAC v.2 user manual p.29.
2 Likes
I am pleased to report that both dCS and my dealer have (in my view) excelled themselves…
Kudos to dCS for delivering my Rossini Apex upgrade components to my dealer within 2 months of placing the order…
and also to my dealer for completing the upgrade, and installing the newly ‘Apexed’ Rossini in my system within 2 working days.
Clearly it’s early days, but I’m impressed with what I’m hearing now, so I look forward to hearing how the sound evolves during the (200+ hours) burn-in process.
5 Likes
Congrats Jonathan, and happy listening 
1 Like