Schiit Audio Releases its Unison USB Upgrade

[mansr]

1 minute ago, plissken said:

"Forget decrapifiers, regenerators, isolators, and all the USB dongles and boxes you’ve been told you need for USB sound. Our all-new Gen 5 USB input doesn’t require any of that stuff. It’s this simple: Gen 5 is USB, solved."
- Mike Moffat, 2017

Clearly, they realised their business mistake in declaring anything done. The road to fortune is paved with upgrades.

 

On the flip side, there's nothing wrong with admitting that the "solved" statement was premature and making something even better.

[mansr]

2 minutes ago, Samuel T Cogley said:

Is there such a thing as a CD transport that works in USB host mode?

A PC with an optical drive?

[mansr]

9 minutes ago, Miska said:

Or even just RaspberryPi or other small SBC. I do that same with HQPlayer, and get upsampled CD data out of the USB to the DAC.

Yes, I wrote a program capable of doing that nearly 20 years ago, though it wasn't the primary purpose.

 

16 minutes ago, Miska said:

To me, CD is pretty much dead these days...

Yeah, I don't see much point in making a fancy CD player today. It's much simpler to rip the discs and play files.

[mansr]

12 minutes ago, The Computer Audiophile said:

It’s here :~)

 

Could you please take a photo showing the markings on the chips? Or just write down the numbers if that's easier. Some of us are curious.

[mansr]

10 hours ago, audio.bill said:

Here's a direct link to the page, and you can click on the photo to enlarge it: https://www.schiit.com/products/unison-usb

Thanks. So there's a PIC32 microcontroller handling USB communication, hardly surprising. Then there's an SI8640 4-channel isolator, also no surprise. I do wonder what the two Ethernet transformers are for, though.

[mansr]

41 minutes ago, JoeWhip said:

Signal isolation perhaps?

Obviously. What I'm curious about is why they chose transformers for some signals and an IC for others.

[mansr]

25 minutes ago, Miska said:

For correct isolation, clocks should be at the DAC side, not at the interface side (I2S over HDMI/LVDS screws this up completely). So clocks flow to the opposite direction...

Yes, obviously. That alone doesn't mean anything for the choice of isolation technique.

 

25 minutes ago, Miska said:

Transformers are good for high frequency constant signal like master clock. But perform poorly for low frequency things like word clock on I2S. While isolation IC's tend to be slow, so not necessarily good for things like master clock, but good for word clock for example (esp. optical ones). So I suspect they chose transformers for higher frequency time-invariant signals and isolation IC's for varying low frequency signals.

The IC used here is good up to 150 Mbps. Judging by the placement of the parts, you're probably right about a transformer being used for the clocks.

[mansr]

6 hours ago, thotdoc said:

Interesting theoretical discussion: How does it change the sound?

That depends on how it differs from the previous version.