USB Audio 2.0: Problems with I2S and various DACs

[misterten]

Hi,

I made a layout near to reference design. The development board L1 contains a Crystal sound chip which converts I2S data. Anyone out there who uses other chips?

I am using a Burr Brown which causes trouble. When changing sample rate the I2S clocks stop (LRCK and BCLK). The Burr Brown DAC reacts on this with an offset on the ouput, regardless wether MUTE has been set or not!!! :o

Can anyone tell me which thread generates the clocks in which I could see and prevent the turning off of these? I'd like to generate a dummy clock while XMOS is calculating and configuring the real clock.

Any help appreciated because I am new in multitasking cores.

[lilltroll]

Is the BB CODEC running in master I2S mode ?
(And XMOS is the slave)

[misterten]

No, XMOS as master and CODEC as slave. Like in reference design.

[lilltroll]

Search for "#ifdef CODEC_SLAVE" in audio.xc (module_usb_aud_shared)

You will find the generation of lrclk och bclk there from the masterclock.

[Ross]

Have you thought about using an external mute circuit?

[misterten]

@Ross:

regardless wether MUTE has been set or not!!

I even pulled MUTE on my DAC hard to low. No chance. Seems to be a bug on the BurrBrown DAC I am using. :(

I've taken out PLL_CLOCK_DELAY and DAC reset which causes something like a clock stretching for just one period. The clock looks like strechted about 4 times. Even this causes a crackle on the DAC.

It gets even worse when selecting an audio file of other frequency and bit rate, which causes a pausing period of clocking of about 500us.

So, BurrBrown doesn't like clock switching. :evil:
I guess I won't be able to avoid these cracklings... :(

[Ross]

@Ross:

regardless wether MUTE has been set or not!!

I even pulled MUTE on my DAC hard to low. No chance. Seems to be a bug on the BurrBrown DAC I am using. :(

I've taken out PLL_CLOCK_DELAY and DAC reset which causes something like a clock stretching for just one period. The clock looks like strechted about 4 times. Even this causes a crackle on the DAC.

It gets even worse when selecting an audio file of other frequency and bit rate, which causes a pausing period of clocking of about 500us.

So, BurrBrown doesn't like clock switching. :evil:
I guess I won't be able to avoid these cracklings... :(

No, I mean an *external* mute circuit to clamp the output of the DAC.