Re: [Discuss-gnuradio] synchronizing sound cards in a cluster (fwd)

[John E. Perry]

Chris Albertson wrote:

> This is to long, skip to last paragraph.
Actually, I'm glad I didn't.  Reading the intervening stuff piqued my 
memory...

> One more idea:  This is harder and it costs you bandwidth but not
> a whole channel like above.  Add a reference signal to your adalog
> signal.  You will later filter this out  Like above the reference
> signal must be delivered in phase to each sound card.
> ...
> My opinion is the sooner a stable clock gets mixed with your data
> the better.  More of their paths will be the same.  but also the harder
> to do. The best an hardest is to mix a time code signal with the analog
> data before it is sampled.  
This is actually an excellent way to synchronize any number of channels 
at practically zero cost!

A simple resistive combiner at the input of each channel of the input 
cards will allow you to
apply a low-level pseudorandom sequence, say, one or two bits amplitude, 
to every channel
in your system.  Straighforward postprocessing tells you the timing to 
within a sample  interval
or so of every channel at intervals of the sequence length.  Removal of 
the probe sequence from
your input data is also straighforward. Fancy processing of the probe 
signal should be able to
give sub-sample timing, though I'm not sure what you'd do with it (I 
know what I did
with it in a very different application).

Quick, easy, noninterfering, accurate.  You have to make sure you don't 
saturate your
input channel, or you lose the sequence information in that period 
(although you could infer
it from whatever you did get by comparing it with timing from other 
channels).

John Perry