TPA - USB Transport

[AndriyOL]

Not on this thread. Definitely off topic.

As for I2S. I have a nice solution for that.

Agree! Will be better to share on BIII thread.

What is your solution, Russ?

 

[NicMac]

As for I2S. I have a nice solution for that.

Differential transmission over a HDMI cable interface is my guess

 

[Russ White]

Not trying to be. Its just more complicated than I want to explain at the moment.

 

[Russ White]

Agree! Will be better to share on BIII thread.

What is your solution, Russ?

No, if you want to discuss such things start a new thread.

 

[Russ White]

Differential transmission over a HDMI cable interface is my guess

Half right. Not over HDMI.

 

[AndriyOL]

So, what a solution? Curious to know..

 

[NicMac]

Half right. Not over HDMI.

Over ethernet, omitting MCLK maybe
A "Twisted P(ai/ea)r Interconnect", or I2S via TPI in short

 

[Russ White]

Yes over cat5, but no you don't have to omit the master clock.

 

[NicMac]

GND connection via shield?

 

[Russ White]

It is differential, no GND connection required. It's a beautiful thing.

 

[mikelm]

Hi Russ,

I know you must be fed up with people asking this but it would really help me to know if you think the release of this USB device will be in days, weeks or months.

 

[Russ White]

My plan is to order boards for this my third board design By October. The rest of the schedule will depend very much on its success.

 

[merlin el mago]

Greast news, I will expect a very good success.

 

[nacder]

Yes over cat5, but no you don't have to omit the master clock.

Will it be able to drive BIII in synchronous clock mode? e.g. MCLK from BIII to USB I2S.

 

[Russ White]

Will it be able to drive BIII in synchronous clock mode? e.g. MCLK from BIII to USB I2S.

More the other way around.

 

[nacder]

More the other way around.

Thanks Russ. So in this case the USB crystal quality can become more important than the one on BIII.

 

[Russ White]

Absolutely, but one does not have to run in this mode, they can still run asynchronously.

 

[Russ White]

The key point is you need two master clocks for the two fundamental frequencies at the USB device.

 

[nacder]

The key point is you need two master clocks for the two fundamental frequencies at the USB device.

Will it be possible (e.g. via some onboard headers) to use some really good clock later (e.g. ultra low phase noise SC-cut OCXO) for these two master clock frequencies?

 

[BrianL]

Russ, I gather that part of this development includes creating drivers to support high-sample-rate transmission. If that is the case, (1) is support (i.e., drivers) limited to one or two O/S's? and (2) what about the future viability of this product as O/S's get upgraded or replaced? It seems like a potentially huge time sink for you or planned obsolescence for the owners.