SM5847 propagation delay

Status
Not open for further replies.
Hi everyone,

I'm currently designing a new dac based on sm5847 + pcm1704 (can't find ad1862s).

I want to use a master clock at 256fs for sm5847 and use that clock to directly feed the bck of the dacs, since this is the bck that really matters on these.

The question is:
does anyone know the delay between XTI (256 fs masterclock) and the bcko produced by the sm5847 (same freq with x8 os) ??

I measured the sm5842 at around 28ns, but I think the sm5847 is more powerful, maybe more delay ?

thanks to all

I'll post the schematic when it'll be over
 
apolon34 said:
Hi everyone,

I'm currently designing a new dac based on sm5847 + pcm1704 (can't find ad1862s).

I want to use a master clock at 256fs for sm5847 and use that clock to directly feed the bck of the dacs, since this is the bck that really matters on these.


Have you tried this with the SM5842-based dac ?
 
for flshzug : I bought them from vcamerica.com (10)
I plan to resell some after the project is finished.


For the reclocking, I didn't try yet with the sm5842. I plan to do it before the project with the sm5847, as a prototype.

My problem is that I need a bit more logic cause my actual dac is 20bit, so I have to add a 12bit shift register and delay the sm5842 clock to be in sync with the dac one (delay around 28ns for my sm5842).

Temporary schematic:
here
 
I finished wiring the reclocking board for the SM5842.
It is now working !


Here is the schematic:
Reclocker

The 30ns delay is inserted before the flip flops to be in phase with the output clock from sm5842 (bcko). It won't be needed for sm5847 cause the clock is already in phase.

The data are then processing through a 12bit shift register (20 bit dac) or 8 bits (24 bits dac).
The dac clock is the XTI one (the GOOD one).

I had to add some delay to wcko line so the falling edge is synced with the clock one.
 
it does support stop clock operation, that's not the problem.

The digital to analog conversion happens at the second rising edge of bck after wclk goes low. Since the bck is generated by the digital filter and full of jitter, I want to use a clean clock for the dac.

two solutions:
- reclock the existing bck with a good clock
- directly feed the dac from the masterclock in continuous clock mode.

I choose the second solution.
 
Status
Not open for further replies.