[hpsdr] Mercury sampling rate

[Phil Harman]

>
> There is a fundamental difference here.   Unless you dedicate some of
> the ATLAS bus to a parallel interface to OZY, transferring more than
> about 250 ksps serially over the ATLAS will be a challenge.   This
> isn't a problem for use with PowerSDR or most Ham radio applications
> since 250 kbps in enough.  For those wanting to record wide bandwidths
> to disk it will be a problem.   I have not looked to see how many
> lines are left on ATLAS, but if you could get enough for an 8 bit or
> 16 bit parallel bus between Mercury and OZY via ATLAS, the wide
> bandwidths then should be possible.   I think that a lot of the ATLAS
> bus is already dedicated to other functions so this might not be
> possible.  Maybe using LVDS and the 622 mbps serial out of the FPGA
> over ATLAS?
>
> Phil N8VB

There are currently enough I/O lines on Atlas for an 8 or 16bit parallel 
transfer from
Mercury to Ozy  -  wether that holds into the future remains to be seen!
 However, I've been playing with a small daughter board fitted with
an LT2208/9 that plugs into the debug header on Ozy for use with Cyclops.

I also provided a similar header on Mercury so there is potential to  place 
a short
ribbon cable between the two boards and do 16 bit parallel transfer without
needing to use the Atlas bus.

Your suggestion of using high speed LVDS is certainly worth perusing - 
thanks
for the suggestion Phil.  From memory the Atlas bus was characterised in the 
low GHz
region when first developed.

73's Phil...VK6APH 


 1208160614.0