...enhances performance by cancelling USB signal loss and optimizing the USB timing diagram...
Does anyone ever had a USB timing diagram printed out and hanging it in the studio ?

If there is USB signal loss, then it is a big problem not only ground loop noise.

Sorry to say what i think, this is snake-oil marketing language.

This small-body piece shall have a useful transformer inside ?

Don't buy it !

I dont have any problems currently, more interested in the effectiveness of this device and or if its just snake oil.

See this posting from MC.

https://forum.rme-audio.de/viewtopic.ph … 03#p223103

What does ground loop and leakage sound like with regards to usb noise?

well well well .. there is quite a substantial potential to be realized acoustically in using galvanic separation, both on the current and the signal path. of course the rule 'holzohr sei wachsam' will not work here.

the following devices seem promising, i only tested the iFi pulsar device so far:

- iFi pulsar usb (using some TI chip for managing galvanic separation)
- iFi omni usb (ANC, optical galvanic separation, some reclocking all in one device; not sure about this)
- intona 7055-c (use version c only for audio)

in general, i think it is good to understand that one can have 3 outcomes of adding new audio components:
2 types of positive outcomes, and negative outcomes. for the positive, it is important to realize that there is a class of 'pleasant 2nd order phaenomena', like a very energetic, but nervous sound, which can be appealing at times, but is ultimately wrong, at least objective EM standpoint (i know that is a tricky example).

i also tried another iFi device, which is called iPurifier 3 (not that it would make sense to name my products like apple products, but i did not decide it). this is an interesting device operating mainly on some reclocking and 'cleaning' of the usb data signal, but when chaining it before the ucx ii with the iFi pulsar usb cable, this also meant doubled ANC, and that was a bit of a 'transparent audio plus effect' (or other higher series active filtering network NF cable effect), which means, more stability, more substance, more body, but ultimately less dynamics and brightness so to say (ie. too much temporal coherence which can be specifically tricky in the high end frequencies). it remains unclear what exactly is right here, and apparently, one cannot have both: dense, bodily sound, and bright, brilliant and open signature.

and then this is also all less of a binary thing than one may would like it to be. in general, galvanic separation addresses ground related residue classes, and energetic ones like EM feedback, and SMPS inflow, ANC usually does symmetric cancellation (XLR logic) of symmetric residues, which usually are HF and noise overlays. reclocking, most exotic and questionable probably, would adress jitter and frame instability.

so logically, i would try the first 2 (galvanic separation, ANC) and only then reason about reclocking, which indeed is, in my view, the most questionable of the.3 base mechanisms.

galvanic separation, in a high end setup, will definitely make a difference, and ANC usually as well. with reclocking, i do not know since i am not sure what exactly the psychoacoustic references actually are there.

with the other 2, it is easy to derive understanding the class of physical residues adressed. i found the pulsar cable to be most appealing since it is only 1 cable and not 2 cables and a box. the intona is for sure the more 'reference grade' device, with the iFi omni, you get ANC again without being optional.

in addition, i can also recommend thalheimer ERT after the laboratory standard (L) for the energy supply chain. if anyone is interested, i can show a diagram how i made a 'galvanic island' for optimizing the audio quality.

hm. not sure what you are alluding to – the devices have no consciousness, hence they cannot identify anything at all. notions like substance or temporal coherence are more a logical consequence of certain reduction of EM residues. that is all there is, really. and as you mention, bits and bytes are kind of irrelevant here since they 'travel' on the data paths of the usb connection mostly, so as mentioned, i am not sure about the reclocking thing, i do not really understand it. but the value of galvanic separation and ANC is easy to see, given that one would argue that the usual residue classes are existing. it is not about bits and bytes, but about the physical layer they are transmitted on.

the to my understanding most correct picture here is a swinging network (aka the frequencies as EM fields): if a 'swing' (di/dt, dv/dt) is too fast, then things like ringing, overshoot of different sorts are occurring. if a 'swing' is too slow, then the energy will be more dampened. a swing is determined by changes of the signal, which is both the energy supply and the audio signal, in case of usb. in case of energy supply in general, this means the wall plug and its path towards the audio components. but the same physical rules apply along the whole signal and energy supply chain.

this model works very well for me and seems physically 100% correct.

ah sorry. yes, i maybe missed the point where it was specifically on the adi series here, and what galvanic separation would bring specifically to that device.

here you go with my 'boxed conclusion':

galvanic isolation on the usb lane suppresses conducted common-mode currents by breaking the conductive ground path, but residual HF/RF energy can still couple across the isolation barrier via parasitic capacitance. and this effect is not limited to eliminating hum and buzz, but also changes the high-frequency common-mode environment of the system.

ANC does not alter connectivity, but can reduce correlated high-frequency noise on supply and return paths, complementing isolation at frequencies where capacitive coupling dominates (in the range of khz to mhz).

so as one can easily see, both are adressing different vectors that are complementing each other. i do not like active mechanisms in general, but inverting the signal to eliminating equal phase residues sounds like a 'good' mechanism that somehow actively resembles XLR/symmetrical NF cable logic.

what i would like to do is a A/B testing the intona vs. the pulsar, that is indeed interesting since it would somewhat expose clearly the 'auditive function' of the ANC in the pulsar cable (given that the signature effect of the galvanic separation would be more or less equal between the TI chip in the pulsar and the 'harder' separation in the intona).

and to your comment 'sound improvements beyond that noise removal are not to be expected' – that does not mean they do not exist! even one does not hear them. but then you may also do not dare to care.

of course it gets harder to quantify or even describe changes in the signature; i made the experience that it was a rather subtle change to remove the iFi iPurifier 3 from the chain. it did exactly what i was trying to say before: the transparent audio cable effect, or mogami 2534 effect and so on and so forth – noise cancelling networks tend to dampen the signature. this can be good, but it can quickly lead to a somewhat 'dull', 'linear' or 'somewhat boring' sound. dampening in RC or ANC networks is hard to 'quantify', it just alters something, but one cannot really quantify it. that makes it hard to 'buy into' that possiblity of optimization, i get that.

so be aware that when you combine the iFi pulsar and the iFi iPurifier3, you will have 2x chained ANC (in both devices) which in my system results in a more compressed signature.

also note that at a certain level of refinement, you may also want to experiment with the output levels – using a velodyne spl x 10 subwoofer, the correct level is +13dBu in my system, this can alter the signature as well.

and a word to 'works the same with a polluted/clean signal' – this reminds me a bit of a galvanic transformer that can 'filter' DC, since it does not transmit it. but it will still sound better when adding a DC filter BEFOREHAND.

come again?