[hpsdr] Ozy/Mercury Protocol

[David McQuate]

One factor that allows current systems to work is the use of 
buffering--both FIFO
and ring buffers are used, so that processing can occur in fits and 
starts, but the
final output stream is uninterrupted, with uniform time between samples.

Dave
wa8ywq

Alberto I2PHD wrote:
> ***** High Performance Software Defined Radio Discussion List *****
>
>   
>
> ------------------------------------------------------------------------
>
> Thanks for the answers about V1.22
>
> While you all are considering changes in the Ozy/Mercury protocol, may 
> I express
> a doubt I have ?
>
> Given the 512 bytes limitations of the FX2 FIFO, samples are sent in 
> blocks of 63
> each sending. Reassembling them in blocks of 512, 1024, 2048, whatever 
> size
> the processing code in the PC has decided to use, is trivial.
>
> But.... if you look at the timing, it is evident that the time 
> interval between the
> instants a block of, e.g., 512 samples becomes ready is not a constant 
> value...
> It has some inherent jitter... Granted, on average that time interval 
> is exactly
> equal to 512 / sampfreq, but only on average...
>
> Doesn't this play havoc with data modes that require phase coherence ?
> I have here a patched version of Winrad that works beautifully with 
> the HPSDR
> hardware through the WinUSB interface, and voice and CW are not 
> minimally affected
> by the jitter explained above, but I am wondering if this is true also 
> for data modes...
>
> IMHO the ideal would be to send 64 samples in each block, moving the 
> control bytes
> elsewhere, maybe using end point 0.
>
> Am I completely wrong ?  Comments ?
>
> 73  Alberto  I2PHD
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> HPSDR Discussion List
> To post msg: hpsdr at openhpsdr.org
> Subscription help: http://lists.openhpsdr.org/listinfo.cgi/hpsdr-openhpsdr.org
> HPSDR web page: http://openhpsdr.org
> Archives: http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/



 1250550016.0