[Date Prev][Date Next][Thread Prev][Thread Next] - [Date Index][Thread Index][Author Index]

Re: Geosynchronous Satellites

>the same stuff used to implement Internet service on CATV plants.  These
>typically used QAM (16- and 64-QAM if I recall on vendor in particular).

64-QAM is popular in cable modems for the forward (downstream) link.
That carries 6 bits/symbol. At 5 megasymbols/sec (which can be
filtered to fit in a 6 MHz TV channel), that gives 30
megabits/sec. Toss in some error correction, and you get the 27
megabits/sec advertised for the Motorola CyberSURFR.

>Interestingly, they'd typically use QPSK in the upstream direction
>because of the training time required for the equalizer used in
>the QAM modem wouldn't be nearly fast enough for the TDMA uplink..

Right, the first generation CyberSURFR also used QPSK. The newer
versions also support 16QAM, which doubles the data rate in the same
600 KHz upstream channel.

>had to bolt-on to the modem design because, well, you don't get
>very much multipath on a CATV plant, but free-space MMDS was a different
>matter completely :-)

Right. Multipath is one of the main reasons spread spectrum is so
effective on terrestrial radio links, and why higher order modulation
(e.g., 64QAM) is such a bad idea. Multipath is like self-interference,
but unlike ordinary noise and interference you can't overcome it by
just turning up the power.

For amateur satellite applications I'd recommend the old stalwarts:
QPSK (or maybe offset-QPSK) at high C/No ratios and BPSK at low C/No

QPSK requires more accurate carrier phase tracking than BPSK at the
receiver, and there's also a greater SNR penalty in recovering a
reference carrier from QPSK than from BPSK, especially when the data
is strongly FEC-coded. This means QPSK should be used only at high
C/No ratios that can support a high data rate relative to the
uncertainty in the carrier phase due to unmodeled doppler, LO
miscalibration, etc, and where the 2:1 bandwidth savings are
important. DirecTV uses QPSK.

BPSK is better at lower C/No ratios because it can tolerate a sloppier
carrier phase reference than QPSK, and also because there's less of a
SNR penalty involved in deriving this reference.

At really low C/No ratios, or when there is a lot of carrier phase
uncertainty, BPSK with a residual carrier is the traditional
choice. The residual carrier does consume transmitter power that can't
be spent on the data, but it allows the receiver to track it with a
conventional PLL without any squaring losses.

I've long felt that a residual carrier should have been part of the P3
telemetry format. It would have greatly simplified the hardware
demodulator designs and alleviated the false-lock problems that plague
so many of them. More importantly, it would have allowed these
demodulators to maintain lock at very low SNRs, keeping open the
option of adding error correction that could permit operation at these
low SNRs.

Nowadays, the case for a residual carrier is lessened with the
availability of DSP platforms that can recover a suppressed carrier
essentially by brute force, e.g., by trying a bunch of frequencies and
phases in parallel and seeing what decodes. But without error coding
on the downlink, there's not much more that can be done.

Via the amsat-bb mailing list at AMSAT.ORG courtesy of AMSAT-NA.
To unsubscribe, send "unsubscribe amsat-bb" to Majordomo@amsat.org