[hpsdr] single board rigs *YES* please...

[Philip Covington]

On 10/3/07, Tony Langdon <vk3jed at gmail.com> wrote:
> At 03:31 PM 10/3/2007, Naylor Jonathan wrote:
>
> >I disagree fundamentally with Bill about A/J/O versus a single board. I
> >see the hardware merely as a means to do interesting things in
> >software, after all it is *Software* Defined Radio, as long as the
> >hardware is good enough then it can be left and people can start doing
> >the truly interesting and revolutionary stuff in the backend. That is
> >where the differentiation between us and Kenwood/Yaesu/Icom will
> >happen. PowerSDR as it stands is not the be all and end all of SDR
> >development.
>
> I can see where you're coming from, but there is more than one way to
> skin the proverbial cat here.  Look at the difference between Janus
> and Mercury/Penelope for starters.  It's clear the DSP starts at a
> different stage in the overall signal processing.  Having the
> building block approach means we can try different hardware/software
> combinations easily.  Of course, the most useful ones could become
> single board units in themselves, once the experimentation bears fruit.
>
>
> >What's really needed is people with DSP knowledge, or willing to learn.
> >The number of people working on the DSP portion of these projects is
> >small, not including people writing yet another SDR GUI. There's a huge
> >amount of fun to be had on the software side, but relatively few people
> >are having it. I admit it's not trivial to get into, but it's fun.
>
> Wish I had the time and energy for studying DSP.  It's an area that's
> fascinated me, but as I'm in a situation where I have to constantly
> self learn, there's not a lot left over Ironically I did my best
> learning when I was unemployed and could devote a lot more energy to
> hobby pursuits.  Much of that knowledge gained in those years formed
> the foundation of the last decade of my career. :)  As much as DSP
> interests me, I can't afford the unemployment that enabled me to
> learn and experiment last time around. :-/
>
> Now if I won the lottery, expect me to want to learn and tinker.  :-)
>
>
> >So count me in for the minimalist single board solution.
>
> I think there is room for single board solutions in the future, but I
> suspect I'll find the A/O + (J/M/P/whatever) platform will be more
> interesting for me for the time being.
>
> 73 de VK3JED
> http://vkradio.com
>
>

I can understand both sides of the argument.  Currently the plug in
board model is great for experimenters.  The building blocks allow you
to assemble a system in different configurations without committing
completely to a final design like you would end up with if the same
modules were condensed down to a single board.

The downside to the plug in model is that there is a lot of redundancy
and extra capability that most end users (those not necessarily
interested in experimenting with building blocks) will not use.  For
example, OZY has a FPGA, so does Penelope, and so will Mercury.  The
Cyclone III picked out for Mercury easily is large enough to handle
all functions that three different FPGAs are doing on the three boards
if condensed down to one board.  So the plug in model ends up being
more expensive.

I think some will be surprised at the cost of the Mercury board.  The
single most expensive component on the Mercury will be the LTC2208 at
~$90.  The Cyclone III FPGA is about $40 and the configuration device
for the Cyclone III will be around $15.   Being on the cutting end is
not cheap!

I think the largest need right now in HPSDR is for PCB designers.
There are plenty of people with ideas and schematics, but too few guys
to do PCBs.  For the released boards so far it has been two guys.  For
me it is not so much the work involved in getting to gerber files to
make the PCB, but in the singnificant amount of time required in
ordering/kitting/ and distribution of parts for the alpha boards.
Maybe those tasks could be split up if someone is interested.

73 Phil

 1191412204.0