Usb Cables

[tony]

Rather than add more posts on the Buildanaudiopc might as well start a new thread.
'Hot topic' on PFM at the moment sure we might as well keep up with our country cousins

Got my Nigecat USB cable in the post this morning and what a surprise. Out of the package I can already see that this cable is going to make a difference. It looks lovely thick copper looking braiding a swarthy look that will no doubt add body and tone to my puny dac.

I have already done a double blind test by putting my current selection of usb cables on the table and each time I picked the NCat.

Have a selection of USB cables here at the moment(virtually all stolen I hasten to add) so will have a go at some point to compare.

On a serious note well done Nigel the cable looks really well made. If the shrink wrap or covering on the usb connectors were labelled it would look like a 'highend' audio product. Very much hoping it sounds as good as it looks


Got m

[nige2000]

it arrived safe then so
good

id prefer it to sound good than to look good,
i do a lot of plugging in and out here so i built it with that in mind
but it does look like something you could tow a car with
i have a shorter cable, similar design that sounds a bit better (its the benchmark)

[jkeny]

What's the construction, Nige or is it commercially sensitive?
I've seen a number of reports now that silver coated copper sounds best.
I believe short lengths act in a number of ways - it probably lessens the effect that impedance mismatches will have (USB is specced at 90 ohms impedance between the two signal carrying wires but I doubt any DIY cables adhere to this?).
Short lengths also represent less of a an antenna for picking up noise & also transmitting sound.

USB signals are high speed square wave signals. So all the considerations for high-speed signalling apply:
- skin effect (with HF most of the signal travels on the outside of the wire & not in the centre - so the silver coated wire might be good)
- reflections from impedance mismatches
- ability of the cable to transmit a clean square wave (the 3rd fundamental is said to contain the most energy for a square wave - so for a USB high-speed signal of 480Mbps, the 3rd fundamental is probably in the GHz region)
- protection of cable from noise intrusion & noise emanation - even though the USB signals D+ D- are said to be differential & therefore any noise on both wires (differential noise) SHOULD be self-cancelling (question - just how good the differential cancelling of USB receivers are & how matched the noise on both wires are)

[nige2000]

What's the construction, Nige or is it commercially sensitive?

if these things were commercially sensitive i must be the worse commercial enterprise going
ive 6 cores of 0.4mm solid core silver wire in teflon/ptfe sleeve made into a braid,
2 cores per usb pin,
no +5v
wrapped in cotton (which i think is pointless)
shielded in a copper braid
gold plated pins (not the plug???)
sealed with adhesive heat shrink

i have a better cable which is the same only half the length no cotton and wrapped in copper foil

I've seen a number of reports now that silver coated copper sounds best.

yes i heard that to,
so i tried a multi core sliver plated teflon coated stuff
wasn't impressed, sure its much better than a printer cable but sounded coppery to me (looser heavier bass, less detail slightly mudded)
maybe i just tried a bad wire or it needed more silver in it or a more correct proportion to balance

I believe short lengths act in a number of ways - it probably lessens the effect that impedance mismatches will have (USB is specced at 90 ohms impedance between the two signal carrying wires but I doubt any DIY cables adhere to this?).
Short lengths also represent less of a an antenna for picking up noise & also transmitting sound.

makes sense

USB signals are high speed square wave signals. So all the considerations for high-speed signalling apply:
- skin effect (with HF most of the signal travels on the outside of the wire & not in the centre - so the silver coated wire might be good)
- reflections from impedance mismatches
- ability of the cable to transmit a clean square wave (the 3rd fundamental is said to contain the most energy for a square wave - so for a USB high-speed signal of 480Mbps, the 3rd fundamental is probably in the GHz region)
- protection of cable from noise intrusion & noise emanation - even though the USB signals D+ D- are said to be differential & therefore any noise on both wires (differential noise) SHOULD be self-cancelling (question - just how good the differential cancelling of USB receivers are & how matched the noise on both wires are)

not sure if using a braid configuration is adding anything?
if i where to shield each wire from each other like start in the center with d+ wrap with foil add d-, wrap and then add gnd
would it be a worthy experiment?

[jkeny]

not sure if using a braid configuration is adding anything?

On the Ciunas - no! With other USB DACs that have the shield connected to ground, I suspect it would have a negative sonic affect

if i where to shield each wire from each other like start in the center with d+ wrap with foil add d-, wrap and then add gnd
would it be a worthy experiment?

I often thought about this i.e stop any noise from affecting the signal wires, rather than trying to remove the noise by differential cancelling at the USB receiver.
It definitely breaks the 90 ohm impedance spec but the cable doesn't have that to start with, anyway :)

So solid silver wire sounds best? Does the Teflon make a difference - tried other dielectrics?

See how much more interesting this thread is already (& we are still only on page one) compared to the same thread on the Borg collective's forum

[nige2000]

the copper braid isnt connected to usb B, which likely doesn't matter it has no +5v to use on other dacs

i meant the braid/weave configuration of the wire same as this early prototype


heres some others the blue ones my favorite


tony has this one atm

[jkeny]

Yes, they look good.

I would also try just twisting the signal wires together & leaving the ground wire separate i.e untwisted.
Two reasons:
- To try to get a more consistent & closer contact between the two signal carrying wires i.e any external noise will more likely appear on both wires equally (differentially) & hopefully be better cancelled by the differential cancellation of the USB receiver chip.
- To try to get closer to 90 ohms impedance but this is just guesswork unless we can measure the characteristic impedance at the frequencies needed for USB. Impedance is determined by the diameter of the wire & the spacing between the signal wires. To be a consistent impedance requires a consistent spacing is maintained for the full length of the cable.

In a sense, Nige, your suggestion about attaching the USB DAC directly to the USB port inside the computer is not such a bad idea if the DAC can be properly shielded. Getting a USB cable right is a difficult balancing act.

[sima66]

I can see that all four cables are quite different, at list from outside.
Which one cables sound better?

Nigel, that's a very nice and neat work. Impressive.
BTW, when the pre-order starts and is there any special discount for the MQn users? ;-)

[jkeny]

I was wrong above - I read it stated on equipment that 720Mhz is greater than the speed needed to pass the 3rd harmonic of hi-speed USB signal & not GHz as I said.

[nige2000]

I was wrong above - I read it stated on equipment that 720Mhz is greater than the speed needed to pass the 3rd harmonic of hi-speed USB signal & not GHz as I said.

do you know the calculation for impedance, (lenght, diameter) i might experiment