Remember that tho the Corning optical cable passes the data optically, it passes the ground and 5V lines with copper. That does have an advantage over having the data on copper right next to the ground and 5V, but it doesn’t break groundloops, or isolate the 5V.
Oh, really? That is a shame. 
I had been wondering how they made the far end work. Thought perhaps they were using a tiny photovoltaic component to generate the 5V potential. If it’s got copper in parallel with the optical then it’s not what I’m looking for.
Ted … can you please clarify: Does the DS need to see 5V from the USB connection even though it doesn’t use it?
If so, why?
My NAD M52, for example, doesn’t use the 5V and doesn’t need to see it either.
The DS needs the 5v signal if it is running Yale Final. This is more Dennis’s doing than Ted’s. It solved a problem where some people lost the USB connection after rebooting their computers.
alcarp said Ted .. can you please clarify: Does the DS need to see 5V from the USB connection even though it doesn't use it?The VBUS line (often misleadingly called the 5V line) does more than provide 5V for non-self powered devices - it's used by the OS/drivers/HUBs to tell when devices are added or leave and further whether that device is a HUB or not. It's used even more with USB On-The-Go, etc.If so, why?
Whether the VBUS line “needs” to be there can depend on the device, any USB hubs used, the source hardware, the drivers and the PC’s OS. In particular it’s needed to fix the problem of having to reboot the DS at times after you reboot the PC, etc.Also Windows 8.0 (vs 8.1) is very persnickety about it.
The DS is self powered and hence doesn’t any working current over the VBUS line, but the VBUS needs to be connected to guarantee compatibility with compliant USB hubs, PC OSs, drivers, etc. Before the Yale software release some people (but not most) could either never hook up the VBUS or they could boot with the VBUS connected and then disconnect VBUS. But that changed for some of then with new PC OS releases or changes in their hub connections.
In order to allow some handheld devices to work with the DS and to not have to (often) reboot the DS when the PC was rebooted the firmware in the XMOS chip in the DS was upgraded to declare that it DIDN’T need 5V, which ironically meant that everyone has to hook up VBUS.
FWIW, Now with Yale and Windows 10 the Corning 3.Optical cable works fine here. I suspect that it was Yale more than Windows 10, but Windows 10 might have better USB 3.0 drivers than Windows 7 (which I was using previously.)
Right now with a heavy load on my laptop (i.e. the fan is spinning fast) and the room fan running I’m not able to listen to compare it to my normal setup, but it isn’t bad.
Slydog1 and anyone else - any further observations about the iusb power supply, purifier, and cable with the DS? I have all these in the house in my upstairs office system, but haven’t yet tried then in my main rig. Won’t be able to do so yet for another week and half due to travel. But am interested in anyone’s experiences. So far I am running via Bridge II and am very happy with both the stability and sonics, but one must always continue to experiment 
TIA!
Can anyone pls tell me, if Yale will work with dual head cables, now? I read in another thread that it doesn’t (as it is the case in my setup)…Any cure to that or do have to stick to Pikes Peak when listening through USB connection…
The only thing that YALE needs is the 5 volts on the USB cable. I would try it and see if it works.
Dennis
Joe Appierto said Another USB cable that seems to work well with the DirectStream is the Cabledyne Reference Silver. I find it quite neutral across the frequency range and nicely detailed. I can now hear/see why the USB connection is so popular with this DAC.+1 I was originally intending to purchase a dual headed cable to run from from my Antipodes DX but after testing and talking to Ed from Cabledyne, decided to go with the Reference Silver. Awesome SQ and amazing value!
Given that my inexpensive 18" USB cable and standard Macbook Air send the Bitperfect file to the DS successfully, would anyone care to tell me why the file would sound different after investing money in different wire or computer hardware changes? It’s plain to me that analog cables are signal modifiers, but I’ve got a real problem believing that bits aren’t bits in the audio world.
Successful Bitperfect implies that any and all jitter and any highly-unlikely temporary errors were resolved by the DS, so that can’t be a factor in the output sound - is the DS not well-isolated from noise on the USB power and data lines?
Can anyone point me to a credible reference with a believable rationale for the notion that USB cables make a difference when they are demonstrably not introducing uncorrected jitter or hard errors?
@EldRick,
I really, really hope that Ted will help you with this subject, given that he, an MIT-trained software engineer who has worked for Microsoft and Google and designed the DS, is likely the only kind of source a skeptic would trust.
That said, I recommend listening to a free trial. If you hear a positve difference, it matters. If you don’t, it don’t. 
EldRick said Given that my inexpensive 18" USB cable and standard Macbook Air send the Bitperfect file to the DS successfully, would anyone care to tell me why the file would sound different after investing money in different wire or computer hardware changes? It's plain to me that analog cables are signal modifiers, but I've got a real problem believing that bits aren't bits in the audio world.There are lots of assumptions in your question.Successful Bitperfect implies that any and all jitter and any highly-unlikely temporary errors were resolved by the DS, so that can’t be a factor in the output sound - is the DS not well-isolated from noise on the USB power and data lines?
Can anyone point me to a credible reference with a believable rationale for the notion that USB cables make a difference when they are demonstrably not introducing uncorrected jitter or hard errors?
Jitter in audio has nothing to do with whether the data is transferred in a bit perfect manner: we assume that it is and any errors in this assumption would have entirely different effects than the jitter we’re worried about. More later.
Jitter is actually a minor effect in cables compared to the obvious issues of radiating (especially with the frequencies involved in digital as compared to those in audio) and ground loops. Radiation can cause problems anywhere in the system, e.g. it is folded down into the audio band by any non-linearities in the signal path. Ground loops add circular currents in a system that depend directly on the area of the loop and the flux thru the loop. This current is not insignificant and causes haze, etc. even if you aren’t hearing any 60Hz hum…
Still I’ll spend most of this post on jitter since it’s usually the least understood.
Unlike pure digital circuitry where jitter is routinely removed with a buffer (and usually something equivalent to a PLL) in audio it’s exactly where the digital becomes analog that any jitter present becomes an unremovable distortion in the analog. The jitter spectrum is convolved with the analog signal and that means that any jitter affects all frequencies of the analog by “spreading” the frequencies. Let’s say we have a simple uniform jitter spectrum 100Hz wide and a signal that’s a pure sine wave. We end up with a smeared spectrum that’s about 200Hz wide at the bottom at a very low level (near the amplitude of the jitter) and that spectrum is fairly narrow and tall at the sine wave frequency. Jitter is usually much more broadband and can affect the whole audio band even with only a single pure sine wave of audio. Thank goodness the amplitude of jitter is small so that when it is convolved with all of the audio band it doesn’t add a huge amount of musically correlated noise.
Looking at it in an entirely different way:
In the simple case of one bit audio (like the DirectStream) notice that a uniform jitter of, say 10% affects the output very similarly to a uniform noise in the power supply of about 10%. We know that a S/N ratio of 100dB (one part in 10^5) is noticeable and hence in a handwaving sense jitter of -100dB (one part in 10^5) on the system clock would be similarly noticeable - for a 5.6448MHz clock at the digital to analog interface like the DS uses this is about 1.8pSec of jitter. To get the S/N down to -120dB that’s about 0.2pSec.
USB had the additional complication that it runs the power and ground signals very close to the data signals over a long length and hence dirties up the power and ground noticeably. High frequency noise from the high frequency signal on a USB cable is very hard to filter out. Even without a connection of the 5V line in the DS we’d still have a radio signal coming from the USB connector that needs to be carefully dealt with in the power supplies and analog circuitry.
The DS takes great care to filter the jitter better than most DACs out there, but even a perfect DAC that introduces no errors of it’s own, is not affected by any errors (including jitter) on any of it’s inputs and radiates nothing outside of it’s case will add multiple ground loops to the rest of the system with the digital, analog and power connections and further those connections (especially the digital ones) will add radiation that affects the rest of the system. This is especially a problem with USB since there’s almost always a computer on the other end which has an almost infinite number of rude noise sources and (with good reason) is almost always plugged into different outlets than the audio system.
Would you expect something like the Audioquest Jitterbug or an iFi Purifier to have positive effects then?
I don’t know the exact mechanism of the JitterBug, but people I trust hear a difference with it and I’m not surprised. It’s also reasonably cheap as tweaks go and I’d try it before I’d spend gobs on a cable. Anything that affects edge rates of the signal will move you to a different tradeoff between radiation/crosstalk and jitter (sharper edges have more high frequencies perhaps causing more emission problems, but sharper edges lower jitter.)
I’m using a $100 iFi USB cable, which has three RF chokes on the shielded cable already.
A “difference”, of course, is not always in the right direction, but Paul’s comment that Arnie Nudell liked it certainly carries some weight.
AQ claims the Jitterbug contains a high frequency filter on the data lines, so I’d sort of expect that to round off the corners, too, but I’ll order one from Amazon (returnable) to try it out.
I have added a 10% discount code for the PPA Red ii usb cable for forum users, the code is psaudio. I have also reduced PPA’s dual head usb cable for a limited time - both available in 1M only.
The PPA Red ii usb cable is now sold and so is only available to order and the discount code no longer applies, the dual head is available (1 only) in stock at the reduced price until it is sold.
I quail at the thought of getting into the world of usb cables and interfaces – outlaying considerable cash and always wondering if the faraway hills are greener. So, I’ve been playing around with usb, seeing what can be done to clean it up on the cheap – with surprisingly good results.
My source is a standard laptop with stripped down Windows 8 and JPlay in single pc mode, controlled by a second laptop running Windows Remote Desktop. It works well.
To clean up the usb between the PC and the Directstream, I interposed a 7-port usb hub with it’s own power supply (@Ted suggested this). I disabled 4 of the ports by cutting the internal 5V track to the relevant chip.
I also cut the 5V wire in the usb cable to the Directstream (and pulled it up to 5V with a battery/resistor instead - Yale Final needs this to keep working). This stops noise from the PC on the usb cable 5V line leaking across to the usb data lines. Next step is to repeat this exercise on the usb cable between the PC and the hub and to battery power the whole lot rather than use the hub wall wart.
As I say, the results to-date are surprisingly good and bring out the best features of Yale Final – toe tapping timing, and finger nails and melodic sustain on acoustic guitar. It goes to prove that PC noise on the 5V wire is the source of a lot of usb colouration.
Total outlay ca. $40 including the usb cable and the hub, and an hour or two of my time.
The next step is to galvanically isolate the PC usb, re-clock the signal, and convert usb to I2S. This eliminates a couple of stages of usb processing and goes straight to the FPGA. The only device I know that does this is the Audiobyte Hydra Z. The outlay here is not cheap at some $850, but is tempting.
Now the hard question – will it bring worthwhile improvement? Any ideas or experiences?
Lonan
pc, qobuz/foobar/jplay, directstream, arcam, harbeth
lonan said ... I also cut the 5V wire in the usb cable to the Directstream (and pulled it up to 5V with a battery/resistor instead - Yale Final needs this to keep working). This stops noise from the PC on the usb cable 5V line leaking across to the usb data lines. ...Hi Lonan, thank you it's interesting.
How did you produce a stable 5V with the battery/resistor couple? What kind of battery and resistor?
I’m wondering too how the first step of your proposition compares against a Regen device (much more expensive)…
An additional question. Is there any advantage to use a USB hub when you could use a USB cable (reasonable length) with the tweak you mentioned (cut the 5V wire and plug a battery) directly connected to the DAC?