[Discuss-gnuradio] Re: TAPR Digital Communications Conference

[Eric Blossom]

Matt and I attended the TAPR Digital Communications Conference at the
end of last week.  We worked like maniacs putting together a couple of
demos.  There are pictures around somewhere of our hotel room strewn
with minicircuits parts, random cables, USRPs, test equipment and
laptops.  It was a sight to behold.

In the end we demo'ed a 4 simultaneous channel FM transmitter, where
we were running 4 narrow band FM transmitters spaced about 50 kHz
apart (could have been further apart without a problem).  In addition
we demo'ed a parameterizable FSK transmitter and receiver.  We were
hoping for 1Mbit/sec, but in the few hours we had to work on it, got
it up to 100kbit/sec.  The system used a simple packet framer that
added a barker code to the beginning for synchronization, converted
bytes to bits to +/-1 symbols, interpolated them (with our new fast
interpolating FIR filter code) and then ran them into the same
frequency modulator block used for the FM modulator.  The demodulator
was the inverse, feeding an 8x oversampled proto-symbol stream into a
correlator that watched for the barker code.  The output of the
correlator was a properly bit and packet aligned stream of bytes.

We'll get all this code into CVS over the next few days.

Expect new tarballs for everything sometime next week.

I met a lot of interesting folks and saw a few good papers presented.
Females were seriousy underrepresented.  Can you say zero?

We also got to spend some time with a few folks working on "Eagle" --
one of the next amateur satellites.  We brainstormed possible ways to
build a relatively wide-bandwidth digital repeater on the satellite as
opposed to the basic "bent pipe".

The discussion centered around using up and down links in the 5 GHz
range with, say 50 channels of FDMA uplink using BPSK and a single
1Mbit downlink using QPSK.  Both links would incorporate contemporary
FEC which should allow the use of simple antennas.  With any luck, a
patch antenna would work for the ground segment -- no dishes required.
TDMA was ruled out because of the high peak power requirement for the
ground station.  The satellite would receive the 50 channels of uplink
with a single RF front end and use SDR techniques to demod the 50
separate channels.  Actually, we thought about splitting the 50
channels into two banks, using two RF front ends and SDRs for
redundancy.  The digital repeater would allow the ground based
stations to run less power and lower antenna gain.  The downlink would
contain all of the repeated uplinks plus additional information such
as usage history on all the uplinks and received bit error rates.
You'd monitor the downlink, then pick an uplink that hadn't been used
for a while.  The up and downlinks would be 5 MHz apart, so you'd be
able to talk and listen at the same time.  Uplinks could contain
digital data, digital voice, or a combination of the two.

Definitely worth the trip!

Eric