DIYHiFi.org

- Have a local power supply, which powers your DAC.
- Have the async USB circuitry (XMOS chip, ARM, whatever you're using) running from the USB power supply, which is galvanically isolated from the local supply.
- Put the playback clock on the local power supply, and feed it over to the USB power domain using digital isolation.
- Feed the I2S/whatever coming from the USB domain into the local clock domain using digital isolation. Then retime them on the "local power supply" side with the local clock.

Crosstalk between power/data doesn't matter, as it's only affecting the power supply on the USB side. Since the USB and local side are galvanically isolated, power supply trash doesn't make it to the local side. And if power supply noise affects the timing of data on the USB side, the retiming on the local side cleans that up.

I don't think you have to specialize in EMI/RFI to design a good system, you just have to understand everything that's going on and deal with it accordingly.

And yet different cables sound different. I have no idea why.

It would need to measure around -90 dBc, @10 Hz, to get those numbers. Don't have one of those fancy machines, and the only person I know who could afford one, decided against buying it.

Having said all of that.....................the phase noise rig I built has noise floor, of -90 dBc, so I can not see below that. I have built clocks that are in the noise, at a few Hz. So, it is possible. Whether it does it or not.............no idea.

Hi Gary,

What you described is *exactly* how the Ayre USB DACs are made. And yet different cables sound different. I have no idea why.

Thanks,
Charlie

If you keep clicking you will see an example phase noise plot. It is about -30 dBC @ 0.1 Hz, -70 dBC @ 1 Hz, -110 dBC @ 10 Hz. And it bottoms out at around -140 dBC. About average for a fixed XO. Nothing too special there.

I also have tried exactly that, even using some of Charlie's excellent component recommendations. The clock and DAC chip were even battery powered, no less.

The cables still mattered.
More memory mattered, as did an SSD.
The type and extent of AC mains filtering needed for the computer mattered.

I have no idea why, either. I have no allegiance to any of the exotica; I'd be much happier if none of this made one bit of difference and the simple way was better.

Oh, changing the subject, eh?
-------------------------------------------------------------------------------------
No, you brought it up.
-------------------------------------------------------------------------------------

No one said anything about 192 kHz. One could make the case that 192 kHz doesn't work, period. Has nothing to do with USB, just a process that probably stresses your poor computer to the point of gagging.
-------------------------------------------------------------------------------------
It is well known that I listen to hires stuff. If someone makes a car specified to do 150 km/h, and 30% misfires at 100, would you then say it it is ok since you only drive at 50?

Interface weakness within spec is very much the point. If something doesn't work relaibly, don't spec it and don't accuse others of being dim witted.
-------------------------------------------------------------------------------------

But, since SPDIF is pretty much just 44 kHz, discussion of your high-res stuff is really way, way, way off-topic.
-------------------------------------------------------------------------------------
Who says, the rest of the audio world uses it and ity is reliable at 192, even 384k
-------------------------------------------------------------------------------------

Jocko

I have a XMOS evaluation board on my desk, 3 feet from me, which disagrees with you. I've had 24/192 asynchronous audio running reliably under both Windows 7 and Linux, and I've never heard a dropout.

Hell, I'll make the board flag a buffer underflow, put some 24/192 FLAC on loop and start up Prime95 to drive the computer to 100% CPU usage for a few hours, just to prove you wrong.

I admit that there may be issues beyond 96 kHz with:

a) Computers that used the first generation of USB 2.0 controllers. They were buggy. But this was a small group of computers manufactured in 2005 or 2006. This is easily fixed by changing or adding a new controller.

b) Home-built computers. All bets are off.

c) If the USB cable is longer than 3 meter and/or if it is built by some stupid audiophile company that has no clue as to what they are doing.

d) The user needs to use the appropriate application to make sure that the DAC is not on the same controller as another high-speed device (hard drive or iPod).

e) You have to use the proper operating system. For Windows, it is XP, Service Pack 3 newer and for Mac it is OS X 10.6.4 or newer.

That's it. It's not rocket science.
--------------------------------------------------------------------------------------
Add
f) Your computer should be used for playing music only, otherwise you may get breakthrough, noise or interruptions.

Comment: since when are there so many restrictions in spfid, firewire, sdif3, or I2S? And how can the intereface be claimed as 'universal' as the name implies

--------------------------------------------------------------------------------------
Complaining about older machines that won't work at 192 kHz is like complaining about an old computer without USB ports that won't support a USB DAC. We have sold nearly 2,000 QB-9's. If the kind of problems you describe existed, we would have to hire ten people to man the support telephones and would be giving refunds left and right.
--------------------------------------------------------------------------------------This has nothing to do with what I said
--------------------------------------------------------------------------------------

But that isn't the case. These things work really, really well. Gordon and XMOS and Thesycon did a lot of work, and that work has paid off.
--------------------------------------------------------------------------------------
People with issues don't seem to get them fixed quickly; read the forums and no I am not a library or a database
-------------------------------------------------------------------------------------

I can't speak for other companies and how robust their solutions are. I know that you have a bunch of low-cost products that you like to complain about, but have you ever used an Ayre or Wavelength or the new Grace Design? These things just plain work.

Why does everybody automatically assume that the variability brought on by different cables or different computer set-ups is due to jitter?

What I said was that variability in sound and voicing with async usb is as much, if not more, in usb connections than good spdif ones.

This, to my ears, clouds any advantages of the claim of superior performance.

Now don't say that I don't know what jitter can and does sound like.