Meteor LRPT requires a QPSK receiver. Since the data rate is low enough to do all the digital processing in the
PC, we simply need a way to dump the I and Q output symbols directly to a file on disk. The obvious solution is
the, now common, SDR - software defined radio, but hardware solutions are also available, especially using
dedicated DSP chips etc. I currently have a cumbersome hardware setup that has several shortcomings, namely I.F.
bandwidth, complexity of use and cost. One day, I would like to come up with a simpler, low-cost, dedicated
receiver that could be used for both LRPT and LRIT... The data rate for Meteor LRPT is 72kBits/sec in its current
format. For a short time the spacecraft transmitted the LRPT stream in the Metop format that raises the bitrate to
80kBits/sec.

The I.F. bandwidth needs to be in the order of 150kHz. Technically, if the Metop spec is to be believed to apply
to the Meteor signal, the I.F should be a matched, root raised cosine filter with and alpha of 0.6 for best
performance, but near enough is good enough to produce usable imagery. Please feel free to Google matched
filtering! www.filter-solutions.com/raised.html is a good introduction.
A good antenna that gives you plenty of margin will cover most sins. I can only get my filter width down to around
300kHz, giving up 3dB or so of S/N margin, and I still get data down to the horizon. The old rules still apply to
these new signal sources. SDR, FEC, DSP and KFC are no substitute for a properly constructed aerial, correct
matching, interference rejection filters and a good pre-amp.

Two further considerations concerning the receiver data output impact on the following step of viterbi decoding.
Firstly, the viterbi decoding stage performs up to ~2.5dB better (in terms of S/N performance/capability) if it is
fed with 'soft decision samples'. This simply means that instead of saying that the voltages or whatever from the
output arms of the receiver demodulator are either a hard-and-fast 1 or 0, they are given a range of values on a
sliding scale. Say, from a strong zero to a weak zero to a weak one through to a strong one. An example would
be to 'slice' the outputs into 8 levels, represented by 3 bits, giving you 25%, 50%, 75% and 100% 'confidence'
levels to send to the soft decision viterbi decoder. Read more about this on the viterbi pages.
Secondly, the nature of these 'suppressed carrier' transmission methods such as BPSK and QPSK mean that,
what is called a phase ambiguity, is introduced by the demodulator. What this means is that the '0's and '1's
coming from the receiver can be translated from the original modulating data source. This QPSK mapping needs
to be dealt with before being sent to the viterbi decoder. Read more about this on my coherent demodulation
page and viterbi node synchronisation page.

My software BPSK/QPSK demodulator

BACK