XMOS DSD 384 kHz / 32bit USB

[analog_sa]

Originally Posted by CheerfulYC

Do you think by introducing a pair of line drivers like the Texas Instruments ISO1540 to the I2S bus will get the job done

I would imagine "low power bidirectional isolators" are the last things you would want to use here.

Whatever buffers you include to split the signal will introduce jitter and subsequent reclocking will, at least in concept, be needed if you aim for a low jitter system.

Otoh, i cannot imagine the need for a low jitter I2S signal at the input of the miniDSP - what you need is a reclock after the DSP.

[analog_sa]

Just realised the miniDSP does not have I2S outputs, so reclocking is obviously not possible.

Maybe Joro can advise on a suitable buffer ic for the MCK. The other signals... i wonder if both inputs cannot be supplied from the same set of outputs with just separate inline resistors...

[joro_s]

Originally Posted by CheerfulYC

..... Do you think by introducing a pair of line drivers like the Texas Instruments ISO1540 to the I2S bus .....

Hi,

ISO1540 is not proper for I2S, it's too slow. If I haven't miss something from the link:

Digital Audio input/outputs via I2S

this device should have I2S input and outputs. I agree to analog_sa, in this case you can connect Buffalo directly and miniDSP through line resistors 100 - 150ohm. Exact resistances depend on input capacitance of I2S line of miniDSP. Buffalo doesn't need reclock, because it's "jitter free" I2S, but miniDSP must has reclock, because it will bring jitter. As for master clock transfer, I would recommend LVDS logic (maybe the line will be long).

Regards,
Joro

[dhbiker]

hi there

are the drivers on XMOS site WDM or not ? i'm mainly a linux user but here and there i reboot to windows and would like to know if it works with windows ?

[WhiteBull]

Originally Posted by joro_s

Hi,

I was playing with the software and I succeeded to bring format for PCM1704 and PCM56. I am wondering, is there any sense to take time (won't be short) to implement it on the main firmware.

Regards,
Joro

Hi Joro,

just discovered this thread ... will put myself on the GB tread for your XMOS board as is, but if you could implement the timings for the PCM56, that would be a clear additional value!

So one more further encouragement from my side for the implementation!

Also, if there would be an option to directly power the board from external 3.3V (A123 LiFoPo), that would be very welcome as well!

Keep up the good work,
Abel

[kvnik]

4. ... or right-justified data
and
I was playing with the software and I succeeded to bring format for PCM1704

What is the difference?

[CheerfulYC]

Originally Posted by joro_s

Hi,

ISO1540 is not proper for I2S, it's too slow. If I haven't miss something from the link:
this device should have I2S input and outputs. I agree to analog_sa, in this case you can connect Buffalo directly and miniDSP through line resistors 100 - 150ohm. Exact resistances depend on input capacitance of I2S line of miniDSP. Buffalo doesn't need reclock, because it's "jitter free" I2S, but miniDSP must has reclock, because it will bring jitter. As for master clock transfer, I would recommend LVDS logic (maybe the line will be long).

Regards,
Joro

Dear Joro,
I will be immensely grateful if you can show me an example with a circuitry to give me a better understanding of what you are talking about! May I thank you in advance!

[joro_s]

Hi,

Originally Posted by dhbiker

...are the drivers on XMOS site WDM or not ?...?

USB AUDIO 2.0 REFERENCE DESIGN | XMOS

https://www.xmos.com/en/download/pub...s-Overview.pdf

Originally Posted by WhiteBull

...Also, if there would be an option to directly power the board from external 3.3V (A123 LiFoPo)...

LP5900 requires 4.3V on input, but can tolerate up to 5.6V. Without changing components, 4 x LiFePo 1.2V will provide enough power for normal operation.

Originally Posted by kvnik

4. ... or right-justified data
and
I was playing with the software and I succeeded to bring format for PCM1704

What is the difference?

The difference is that RJ for one sample form of WS (L / R) is a meander, and the low level data is transferred to the left channel, the high data right channel, i.e. successively. If we want to have stereo PCM1704 without additional treatments, one has to use two separate parallel data channels.

Originally Posted by CheerfulYC

.. if you can show me an example with a circuitry to give me a better understanding of what you are talking about!...

I attached a picture for you.

PS There is thread for PCM1704, PCM63, PCM56 and other DAC and there one can discuss only about DACs which requiring separate data channels:
XMOS direct drive PCM1704, PCM56, PCM63 and others for NOS mode or PC oversampling

Regards,
Joro

[dhbiker]

do you have it on stock now ? the board and the dac module ?

[deanoUK]

Got mine yesterday, thanks.
All working OK.