Many questions baconbrain!
It looks to me like you do already understand the basics of this topic as well as anyone, but are perhaps looking for explanations that simply do not exist, certainly in the area of 'clock cascading', an area where I have been looking for answers myself.
Anyway, some thoughts from myself. The sMS200Ultra does not have a 'word clock' input, it has a 10MHz reference clock input (or 'master clock' as SOtM call it). A slightly pedantic comment I know, but these are different things. Of course the Mutec REF10 is a 10MHz reference / Master clock. So a 10MHz reference clock input on the sMS-200Ultra is what you need.
AES3 protocol includes a clock signal embedded in the feed. So, if the REF10 is providing the clock to the Mutec MC3+USB, then the REF10 clock signal should feed to the Trinnov and then on to the Devialet. It is not clear exactly how Devialet use the clock signal in the AES3 feed (AES/EBU), but if it was not using the clock in the feed then SwissBear would not be hearing the improvement that he is. Indeed, if we assume that the implementation of the AES/EBU input is the same in principle as it was on the 'old' Expert range, we have to assume that the Devialet uses this embedded clock, otherwise SQ would not improve with Aurender W20's N10's, Mutec MC3's, dCS Vivaldi's and so on. There have bee too many reports of Devialet SQ improvements via decent AES/EBU sources to make any other assumption.
So, you can attach the REF10 to you sMS-200Ultra, and your Mutec MC3, and in theory the high quality clock signal generated by the REF10 should make it all the way to your Devialet. Which is the point I think, you state that you need at least three devices to run from the REF10 to justify it's cost, I am not sure I would agree with this. What you need is for the best possible clock signal to make it to the Devialet, and for this to offer a sound quality improvement that justifies the REF10's price. The bit I cannot explain is why connecting the REF10 to the sMS-200Ultra will help, because the Mutec MC3+USB has an asynchronous USB input, so the clock in the Mutec MC3 (improved by REF10) will govern here. In theory at least, the 'clock cascaders' on CA say different, even claiming that the clocking the Ethernet switch makes a big difference. (the modified switch is a relatively cheap option with the sMS-200Ultra, so you might want to try this, even though there is no current explanation why it should work!)
So to your question, 'Can anyone explain how exactly does clock cascading work?' If you are referring to the use of a master clock running multiple 'up stream' devices, I do not think anyone has a satisfactory technical answer as to why this works, but many state that it does work. The first half of post #7 of this thread covers this one, as does about 100 pages in a thread on Computer Audiophile.
It looks to me like you do already understand the basics of this topic as well as anyone, but are perhaps looking for explanations that simply do not exist, certainly in the area of 'clock cascading', an area where I have been looking for answers myself.
Anyway, some thoughts from myself. The sMS200Ultra does not have a 'word clock' input, it has a 10MHz reference clock input (or 'master clock' as SOtM call it). A slightly pedantic comment I know, but these are different things. Of course the Mutec REF10 is a 10MHz reference / Master clock. So a 10MHz reference clock input on the sMS-200Ultra is what you need.
AES3 protocol includes a clock signal embedded in the feed. So, if the REF10 is providing the clock to the Mutec MC3+USB, then the REF10 clock signal should feed to the Trinnov and then on to the Devialet. It is not clear exactly how Devialet use the clock signal in the AES3 feed (AES/EBU), but if it was not using the clock in the feed then SwissBear would not be hearing the improvement that he is. Indeed, if we assume that the implementation of the AES/EBU input is the same in principle as it was on the 'old' Expert range, we have to assume that the Devialet uses this embedded clock, otherwise SQ would not improve with Aurender W20's N10's, Mutec MC3's, dCS Vivaldi's and so on. There have bee too many reports of Devialet SQ improvements via decent AES/EBU sources to make any other assumption.
So, you can attach the REF10 to you sMS-200Ultra, and your Mutec MC3, and in theory the high quality clock signal generated by the REF10 should make it all the way to your Devialet. Which is the point I think, you state that you need at least three devices to run from the REF10 to justify it's cost, I am not sure I would agree with this. What you need is for the best possible clock signal to make it to the Devialet, and for this to offer a sound quality improvement that justifies the REF10's price. The bit I cannot explain is why connecting the REF10 to the sMS-200Ultra will help, because the Mutec MC3+USB has an asynchronous USB input, so the clock in the Mutec MC3 (improved by REF10) will govern here. In theory at least, the 'clock cascaders' on CA say different, even claiming that the clocking the Ethernet switch makes a big difference. (the modified switch is a relatively cheap option with the sMS-200Ultra, so you might want to try this, even though there is no current explanation why it should work!)
So to your question, 'Can anyone explain how exactly does clock cascading work?' If you are referring to the use of a master clock running multiple 'up stream' devices, I do not think anyone has a satisfactory technical answer as to why this works, but many state that it does work. The first half of post #7 of this thread covers this one, as does about 100 pages in a thread on Computer Audiophile.
1000 Pro - KEF Blade - iFi Zen Stream - Mutec REF10 - MC3+USB - Pro-Ject Signature 12
