Specific question for BADA USB converter users.


What lengths of USB and SDPIF cables do you use? I'm trying to decide between the BADA and the Audiophilleo w/PurePower converters. One of the issues is how much will i have to spend in cables to get optimum sound. BADA advises 1.5 meters for both cables, which contradicts conventional wisdom with regard to having the USB cable as short as possible. If I ignore BADA's advice, I could buy a 0.3m USB cable with a 0.5 meter AES/EBU cable, which would be more comparable in price to the two 0.3m USB cables I could use on the Audiophilleo.

I'd rather have higher quality shorter cables than longer less effective cables. I would also prefer the shorter cables because they would be easier to install in my planned setup. For those who use shorter lengths on the BADA than what they advise, have you tried the 1.5m lengths before deciding on the shorter cables? I'd like to get the impressions of those who have used/are using the BADA USB converter and have used shorter lengths than advised. This is not a general question about cable lengths, about which there are already several threads on various websites.
asindc
hi Asindc, I have used Wireworld Starlight at 1.5m and now using an Atlas cable at 0.5m

I prefer the 0.5m cable but it's not a massive difference.

I've been getting much more significant improvements switching from Jriver to JPLAY
All digital cables that are used in jitter sensitive applications should be around 1.25-1.5m in length. Read this:

http://www.positive-feedback.com/Issue14/spdif.htm

Digital is not like analog. Shorter is not always better, unless it is really short like 4 inches.

Steve N.
Empirical Audio
Thanks Steve for the link. Very informative.

A couple questions come to mind.
1) Does the impedance issue pertain equally to USB transmission as SPDIF transmission?

2) With use of purepower with the audiophillieo. How much of the benefit is due to the ellimination of the problem of 5V transmission along side the audio signal through the USB cable?
"1) Does the impedance issue pertain equally to USB transmission as SPDIF transmission?"

You bet.

"2) With use of purepower with the audiophillieo. How much of the benefit is due to the elimination of the problem of 5V transmission along side the audio signal through the USB cable?"

Its not that the 5V is in the same cable that makes it particularly bad, although the return current in the ground wire can make the common-mode noise worse.

Its that the 5V power is poor to begin with from the computer and then a long cable to transfer it makes it even worse. Replacing this with a better local 5V power is a beneficial thing.

Even without the 5V in the cable, there is still a ground as well as two signal wires. The ground is still a problem because unwanted loop currents flow on this wire causeing common-mode noise at the receiver. The best options are to filter this using a Short-Block or use a USB interface that is galvanically isolated.

Steve N.
Empirical Audio
Steve, I certainly agree with everything you have said, except that I'm curious as to why the optimal cable length for USB would be similar to what is optimal for S/PDIF. The rationale for 1.5 meters that is presented in your paper is based on the presumption that signal risetimes and falltimes are in the vicinity of 25 ns. Aren't risetimes and falltimes far faster than that for USB 2.0, and therefore why would the optimal cable length for the two kinds of interfaces be similar?

Regards,
-- Al
Why not just spend $150 for an ESI Juli@ soundcard? It sounds much better than either USB converter at a fraction of the price. USB is NOT ready for prime time.
Cerrot - you have never heard a good USB interface.

I've had all of the best PCI cards here, including RME and Lynx. Not even close to good USB.

Al - risetimes may be a bit faster with some USB interfaces, but the length minimum still applies.

Steve N.
Empirical Audio
I have heard most every usb interface (including yours) and, regret, none have floated my boat. The usb to me sounds like a bag of cats on masses of strings at nice volume levels. The USB cannot disentaggle complex sounds. its too busy doing too may primitive things it should not be doing. I am actually surprised at the number of people "in the know" who disagree. I guess there are more usb ports on a computer than pepole willing to install a sound card so the market has targeted that and not really looking back. It will. The convenience of the misic server is wearing down so people will now start listening. I think the first few years were an iPod upgrade. Stay tune for the real thing. Other than hires, we will all go back to cd players. The servers are only useful for that (hi res downloads) and the ol background music jukebox.
08-24-13: Audioengr
Al - risetimes may be a bit faster with some USB interfaces, but the length minimum still applies.
Thanks, Steve. Based on your experience I don't doubt that 1.5 meters is a good length for a USB computer-to-DAC cable. The rationale still puzzles me, though. Given that USB 2.0 supports a theoretical maximum bit rate of 480 mbps, using both the positive-going and negative-going clock edges, the risetimes and falltimes of USB 2.0 signals would seem to have to be able to support a clock period of around 4 ns (approximately 1/240 MHz). Which would say that risetimes and falltimes have to be in the vicinity of 1 ns at most. Very different than the 25 ns or thereabouts upon which the rationale for 1.5 meter S/PDIF cables is based.

Regards,
-- Al
Thanks for your answers Steve.
Cerrot - you have never heard an Off-Ramp 5. You must be doing something wrong if all of the USB interfaces you have heard are poor.

Al - at 192 sample-rate, the frequency is only 50MHz.

Steve N.
08-24-13: Audioengr
Al - at 192 sample-rate, the frequency is only 50MHz.
Steve, note that I said that "the risetimes and falltimes of USB 2.0 signals would seem to have to be able to SUPPORT a clock period of around 4 ns (approximately 1/240 MHz)."

Surely the risetimes and falltimes of USB 2.0 signals don't change as a function of data rate. Or so it seems to me, at least in situations where the link is actually being used in USB 2.0 mode, and is not falling back to USB 1.1 operation. And therefore the risetimes and falltimes would have to be compatible with the fastest rate USB 2.0 can support.

Also, isn't 192 kHz/24 bit/2 channel data more like 10 MHz, plus a bit more for protocol overhead, rather than 50?

Regards,
-- Al