I was just basing on a basic Pioneer DVD players coax vs toslink output. I have a standard pioneer and an elite player which I've also tried and both of them sound more bloated and incoherent using coax.

So I see that USB type A plug has 4 pins:

1 - GND (Ground)
2 - D-
3 - D+
4 - Vbus (+5V)

What happens if either pin 1 or 4 are covered with electrical tape?
Maybe it would be possible to prevent a ground loop by just covering pin 1, but would this have any negative effect?
Anyone tried this?
Is it the equivalent to this what devices like the ifi iusbpower basically do?
If the USB device is powered by the Vbus line then I still would need the ifi iusb or something similar with a power supply, correct?
So I guess that this simple tape mod alone is not effective?
Obviously I would still have Jitter to deal with, but since the GND and Vbus lines are cut right at the source and don't contaminate the D+/- lines throughout the cable lenght, maybe it would be conceivable that this could have some jitter reduction potential... no? (just talking out of my behind)

IF you tape (disconnect) pin1 USB doesn't work anymore as the GND is a reference that is actually used and not just acting as a 'screen'.
USB is not a truly symmetrical (balanced) signal, if it were it could do without the ground or have it implemented in another way.
Disconnecting pin-4 can only be done IF the circuit it is connected to (a DAC or converter) is fed via a separate power supply or something like a USB power injector.

disconnecting pin 4 and feeding the receiving end with an external power supply is likely to improve that supply voltage 'quality' for the USB receiver chip(s) as the power supply voltage is cleaner (USB power voltages from a a PC can be very noisy).
Note that the incoming 5V from the USB connection is, or should be, filtered on the receiving side anyway.
From a technical viewpoint using an extra power supply feeding the USB device (DAC for instance) does NOT 'disconnect' the connection with the PC but only the 5V supply voltage and does ZILCH against common mode RFI as that passes via the ground AND datalines.
In fact it may even worsen the situation as MORE common mode (depending on the kind of extra power supply) is introduced.
An option to get rid of this possible extra problem is feeding the receiving end from a battery BUT the common mode RFI will STILL be passed on unattenuated though through the ground and data-lines.

In my (rather limited technical opinion on this specific matter anyway) when a DAC is designed with internal jitter elimination (most newer ones are) the jitter on the USB bus doesn't (shouldn't) matter BUT the RFI can still f#k things up by becoming audible in a number of ways.
Those using their ears as analysers may have different views though.

About SPDIF via coax:
Although the signal is/could be galvanically separated by the in- and out-put transformers used (when the shield/ground/common is not connected to the shield/ground/common of the input and/or output device) nasties like and DC and audio-frequency AC ground-loops are/could/should be broken, but common mode RFI will still be 'conducted' almost completely unattenuated and can still cause troubles.

No, it's the worst. I2S is ideal, but is only really usable in short lengths. 75Ohm S/Pdif via BNC is generally the best, most practical solution overall. AES/EBU is fine, but can run into impedance mismatch problems, same with S/Pdif over RCA. Toslink is the worst. While it does break the electrical connection, the timing errors that it introduces are too large to ignore. ST glass fiber is better, but nobody really uses it anymore outside of Bel Canto.

While I agree that I2S is the best (optical I2S anyone?), the timing errors introduced by modern Toslink aren't large enough to worry about. Remember that it's digital signal and that the jitter introduced by reasonable converters is below -110 dB in J-test, which is worst case signal. With real signals, it's most likely even less audible.

(Have a test: http://nwavguy.blogspot.com/2011/02/jitter-does-it-matter.html - and that is a relatively crummy source!)

Of course, there do exist broken converters. It's trivial to find them, but you need great measuring hardware for this.

Also remember that USB, both power supply and general signal conversion to I2S in the interface chip, can cause jitter - whether it will be better or worse than optical SPDIF depends on the device.

Quote from: DaveBSC on December 22, 2012, 10:00:27 AM

No, it's the worst. I2S is ideal, but is only really usable in short lengths. 75Ohm S/Pdif via BNC is generally the best, most practical solution overall. AES/EBU is fine, but can run into impedance mismatch problems, same with S/Pdif over RCA. Toslink is the worst. While it does break the electrical connection, the timing errors that it introduces are too large to ignore. ST glass fiber is better, but nobody really uses it anymore outside of Bel Canto.

While I agree that I2S is the best (optical I2S anyone?), the timing errors introduced by modern Toslink aren't large enough to worry about. Remember that it's digital signal and that the jitter introduced by reasonable converters is below -110 dB in J-test, which is worst case signal. With real signals, it's most likely even less audible.

(Have a test: http://nwavguy.blogspot.com/2011/02/jitter-does-it-matter.html - and that is a relatively crummy source!)

Of course, there do exist broken converters. It's trivial to find them, but you need great measuring hardware for this.

Also remember that USB, both power supply and general signal conversion to I2S in the interface chip, can cause jitter - whether it will be better or worse than optical SPDIF depends on the device.

Sorry, but I don't buy it. I've heard the "Modern DACs wipe out all incoming jitter rendering the connection meaningless" argument many times, and it just isn't borne out in reality. Optical still stinks, especially out of a high jitter source like a PC soundcard.

Less than 1000ps of jitter is "good enough"? For a clock!? You can't be serious. I've tried both optical and coaxial from PC motherboards and soundcards, and the results are very sub par. The Lynx with an AES breakout cable is average at best.

The best results from a PC, bar none, is asynchronous, with 44.1 and 48 based clocks that are sub 1ps.