Its a very specific device, its converting a fiber optic signal to ethernet. I didn't even know that an aftermarket option existed.
I assume it would replace the existing digital converter? I don't know enough about this type of equipment.
Yeah its basically a FMC with a better/good clock and PSU, than the standard FMC, I think.
I should say although a little off topic from master clocks, I feel there is some relevance, given how much better my OCK-1 slaved DAC performed in my system when fed what I guess must be ‘cleaner’ (see below) or less jittery (ditto) ethernet data by the LHY SW-8 switch. The delta between external clock on & off is greater with the switch than without. So in theory at least this sort of device that brings galvanic isolation could bring similar and in my case, hopefully additive, improvements. And then there’s the fact the Teradak uses seemingly better oscillators than generic fibre media converters (FMCs).
There’s some other very expensive aftermarket devices that do the same thing, often with the fibre section internal to the box. Then there’s the generic networking FMCs a lot of folk have been using with good results in several threads I’ve been on - R26 & U18 in particular. This sits in between.
In terms of the mechanism of action, as to why FMCs, or for that matter, audiophile switches with no FMC element, work to improve the sound, from what I can tell from my limited reading to date the jury is still out. There’s certainly theories and robust debate. Removal of ‘noise’ riding the ethernet signal and the disruption it causes to the receiving device most likely as increased jitter, the noise assumed electrical or EMF induced, is most commonly often offered up as an explanation. Fibre based galvanic isolation and use of lower noise toroidal power supplies on the downstream/DAC side play into that explanation.
But then there’s a lot of anecdotal evidence that upstream network devices with more precise clocks (eg. EtherRegen switch or my experience with the LHY switch) are better than the same devices with less precise timing. Eg. EtherRegen internal OCXO vs using a higher quality external clock with it. The suggestion being there is less jitter in the signal chain as a result. How to explain this given we are talking about packets of data, with error correction which are buffered and reclocked at the receiving end.
I don’t know. Still educating myself. Not sure there is clear evidence either way but then as the saying goes absence of evidence isn’t evidence of absence.
What I do know is that for whatever the reason my DAC (+ ext clock) sounds far better with the SW-8 upstream. So that’s a provisional personal experiential/correlation association tick for the role of both linear PS and OCXO in a network device, which coupled with the experience of others with FMCs and switches, make it more likely to me they may play a role in improving the sound of the Teradak FMC too. All depends on implementation of course, but let’s assume competent wiring, circuit design for now, despite the low price.
So that’s how I’m trying to make sense of things, approaching this FWIW.
and 
I know. And/or toggling ext clock on/off.
