[hpsdr] New project idea - self-contained Hermes/Apollo transceiver

[Steven B. Dick]

Now that Hermes (and presumably its Apollo companion) are getting close to
production, I would like to bring up a point that has been bothering me for
awhile. It's a tremendous accomplishment by the design team to have achieved
such a high performance transceiver board set. However, packaging is
currently lacking. The Apollo discussion says "The idea is to build a self
contained HPSDR Transceiver into a box similar to the one used for the two
<http://openhpsdr.org/wiki/index.php?title=ALEXIARES> Alex boards. 

The box, made by Hammond has a sliding cover on one of the sides (Series
1455, PN 1455N1601). This cover can be used as Front Panel and behind this a
display up to 4 inches may be installed. There will also be space for a
controller (Beagleboard or something else). One of the goals is a small
platform to use as controllers for a self contained transceiver." 

I believe there would be a lot of interest in this. A self-contained
processing capability coupled with an excellent mechanical package and
excellent operator interface (controls/displays) without being tied to an
external P.C. would provide a professional level transceiver, not only in
performance but also in looks and usefulness. The board set will not be
inexpensive, and potential buyers will be looking at other packaged
alternatives such as the upcoming Elecraft KX3. With regard to operator
interface, take a look at the Juma TRX2 or the KX3. Elegant simplicity.

A list of requirements for the operator interface needs to be developed. A
fundamental decision would be whether to use a simple 2 row LCD display or a
more capable graphics display that could display a relatively wideband
spectrum in real time, depending on available processing capability. The
size of the case should not necessarily make it as compact as possible but
rather be determined by what controls/displays we want to have.

The extra resources in the FPGA, which has a lot of growth capability,
could be used to do much of the "heavy lifting" such as FFT/IFFT based
filters and even dedicated filters for a spectrum display in addition to
other transceiver filters.  The actual controller board "guts" in my opinion
is secondary to the look and feel of the transceiver. Many more people would
want to buy this as a self-contained, professional looking unit than just
the board set. One could still use the audio or Ethernet capabilities with
an external p.c. for all the bells and whistles or use the internal
self-contained capability.  Very low cost but high performance
microcontrollers are coming on the market now, some of which have very good
DSP performance.  The selection of the processor should be dictated by the
availability of a good software development environment that is available al
low or no cost besides its hardware performance.  I would look at the Arm
Cortex M4 type processors as low cost/high performance candidates.  Take a
look, for example at the STMicroelectronics STM32F4 Discovery board. (See
link below)

 
http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/USER
_MANUAL/DM00039084.pdf

It has a 168 MHz/210 DMIPS Cortex-M4 with single cycle DSP MAC and floating
point unit while providing outstanding power efficiency, Up to 1 Mbyte of
on-chip Flash memory, 192 Kbytes of SRAM

What do you all think? Any ideas about operator interface requirements?

Regards, "Digital Steve", K1RF

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/attachments/20111210/a15421a7/attachment-0002.htm>