<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">On 5/28/2017 1:33 PM, Kjell Karlsen
wrote:<br>
</div>
<blockquote cite="mid:op.y0yq9hg4nknqug@kjellkarlsen-pc" type="cite">*****
High Performance Software Defined Radio Discussion List *****
<br>
<br>
<br>
6 New FPGA. Phil, VK6PH is looking for a more modern FPGA to
replace the
<br>
Cyclone III that is used today<br>
</blockquote>
<br>
....<br>
<blockquote cite="mid:op.y0yq9hg4nknqug@kjellkarlsen-pc" type="cite">
<br>
When the construction is tested and found OK, we can make a list
and
<br>
produce fully populated and tested boards. Also bare boards will
be
<br>
offered for the brave ones.
<br>
<br>
</blockquote>
1) Could we drop the requirement that the FPGA not be a BGA version?
I know that means that people who want to assemble their own boards
would have to be really really brave :), but it opens so many more
options, and the number of people who want to build their own are
significantly in the minority. <br>
<br>
I'd like to see the board laid out for a BGA FPGA. Pick a family of
FPGA's that are pin compatible but have a large range of
capacities. Then when people reserve a board they can pick which
size FPGA they want to have, and as long as there are enough people
at each size, we can have runs done for multiple FPGA sizes, but
everything else being the same. The people testing/writing the FPGA
code can choose one particular size to test / develop on, and they
can make that size known in advance, for those who want to make sure
they get the "supported" size. People who have the other sizes can
get the base firmware, change a few parameters and build a version
that supports their larger capacity board without having to know
much of anything about FPGA programming, i.e. just enough to run the
tools (of course they could run into timing closure issues). My FPGA
skills are fairly weak, but I would certainly be willing to create
new Hermes FPGA's for at least whatever size I buy and make it
available to others, given a release of the RTL code for the
"supported" size.<br>
<br>
For example, we could choose the Cyclone V E FPGA family in the 484
pin BGA package. Using the 484 pin package allows the full range of
Cyclone V E FPGA capacity to be available. The cheapest FPGA's
within that package are the one's with no hard PCIe controller or
hard memory controller, with the slowest fabric speed grade of "8".
That allows for 5 pin compatible FPGA's:<br>
<br>
<br>
<tt> <big><big>Product Code LE's Price</big></big></tt><big><big><tt><br>
</tt><tt> 5CEBA2F23C8N 25 $37.76</tt><tt><br>
</tt><tt> 5CEBA4F23C8N 49 $53.30</tt><tt><br>
</tt><tt> 5CEBA5F23C8N 77 $88.04</tt><tt><br>
</tt><tt> 5CEBA7F23C8N 150 $159.36</tt><tt><br>
</tt><tt> 5CEBA9F23C8N 301 $209.76<br>
</tt><br>
<small><small>The A4 variant is about equivalent (slightly
better) to the size of the Cyclone 3 FPGA used on the
current Hermes board, so perhaps we shouldn't offer the A2
variant at all. The prices are single quantity Digi-Key
prices, are are presented purely to give an idea of relative
pricing vs. size. Perhaps to keep it simple some options
might be eliminated, e.g. perhaps only the A5 and A9
variants, i.e. "better than Hermes", or "max". Anyway, this
is just an example of what I am suggesting. </small></small><br>
</big></big><br>
I for one would like a newer generation Hermes with a much larger
capacity (at least enough capacity to run one receiver slice on
every ham band at the same time), but I don't think that should be
done by forcing everyone to pay an additional $100-$200 for the
largest capacity FPGA.<br>
<br>
2) Either make sure the new power regulators can handle powering the
largest possible FPGA, or at least continue to make it possible to
bypass the SMPSU(s) if necessary (i.e. I don't mind still needing to
provide external regulated power for 5V and 3.3V if it is required
for those who want the largest possible FPGA capacity).<br>
<br>
3) Although I think it's great that other boards may be done that
will be compatible with the new Hermes, but please make sure that
the new Hermes remains completely independent like the old one, in
order to maximize flexibility. My Hermes is the only OpenHPSDR board
in my current setup.<br>
<br>
4) Consider taking the ACC port connector off of the edge and
placing it on the top or bottom of the board. Many people don't use
the features on this board, or if they do, they could bring it out
with an internal connector wired to an external connector on their
custom chassis as they see fit. This would leave more space
available for other feature / external connector alternatives.<br>
<br>
5) If there is any thought of adding a second ADC, please make that
an alternate fill option (possibly requiring those who want it to
exercise their SMD soldering skills). My guess is that a majority of
people don't care for the extra expense for a feature they have no
plans to use.<br>
<br>
6) Try to keep the budget for the final production board relatively
close to the previous Hermes price, at least with the smallest
supported FPGA (still that should be >= current Hermes capacity).
<br>
<br>
73,<br>
<br>
John<br>
AC0ZG<br>
</body>
</html>