DAC clocking idea

[preiter]

The whole clock jitter thing fascinates me, because with modern digital electronics being run at at such high speeds, it doesn't seem like getting a good clock to a DAC should even be an issue.

I had an idea, why not feed the output of your DAC receiver into a fifo and then have the DAC read the data out of the fifo on its own clock? This makes jitter on the input irrelevant.

The issue I can see right off is that unless the input clock frequency matches the DAC frequency exactly the fifo will eventually over-run or under-run. But surely there is some way to adjust the frequency of the DAC's clock over the long term without introducing jitter?

I'm sure I haven't invented a new concept here. I only pose the question because I hope that someone will educate me on A) why it won't work or B) it's already being done.

[davecandialex]

In the telecommunications world FIFOs are often used when data jitter is large, in simple terms more than the period of the data clock. For a DAC input usually the jitter is not so big and the local clock can reclock the input data to remove the jitter. The trick of course is in the second part of the question, syncing the clocks of the source and DAC.

My statement above may not be true for data output from USB.

[s3tup]

PLL + FIFO + SPDIF receiver = WM8804/5

Yet it has it's own jitter due to PLL

[wwenze]

Get a DAC with wordclock out and a soundcard that can derive its clock from and sync with SPDIF input (e.g. VIA Envy24).

[abraxalito]

It is already being done, though not by many. See the last page (p11) of this pdf :

On Jitter - Lavry Engineering