Explanation of qpsk Demodulator
Note : the info below is not for the totally non-technical.
To understand it, you really ought to 1st read my Digital Satellite Faq,
available on several websites as "Digfaq.zip" & complete witha block
diagram of a typical digital satellite receiver.
(Try , amongst others , http://www.awihg-sat.de
QPSK Demodulators -what does S/R & 3/4 mean ??? etc.
Most digital satellite receivers convert from L-band (900-2100 MHz)
down to a fixed intermediate frequency of 480 MHz.
It is here that some of the "magic" new technology takes place using
qpsk demodulator chip-sets from companies like LSI Logic, VLSI Technology
& Broadcom. In fact the latest versions are a single chip rather than 2 or 3
[The first so-called "single-chip" solutions cheated in that they still
a further chip for the analog-to-digital conversion (ADC) - *latest*
chips tend to include the ADC within the chip]
The following is my attempt to describe the conversion of the modulated
signal from the DBS tuner to error-corrected digital data -which is then
to the MPEG transport demultiplexer.
In other words, I am trying to shed some light on the error correction
The main figure of merit for a digital demodulator is the minimum Eb/No
that the receiver can tolerate whilst delivering data at a specified BER
(bit error rate) to the mpeg2 sections.
Eb/No is the ratio of Energy per bit to Noise available at the demodulator.
Since satellite channels tend to be noisy , a low order modulation scheme
scheme is used with lots of Forward Error Correction (FEC).
The DVB scheme uses a CONCATENATED FEC ,meaning that 2 types of
coding are used - CONVOLUTIONAL and BLOCK coding.
VITERBI coding is a form of convolutional coding & the "code rate" refers to
the number of output bits for a given number of input bits.
A rate of 1/2 means that only 1 bit is outputted for every 2 bits going in -
of the bits are "wasted" in the decoding/error correction.1/2 code rate would
used for channels with lots of noise (low Eb/No). The error correction comes
from the redundant coding data that is transmitted. You may also hear a term
called "constraint length" , 7 for DVB , which is the number of bits over
the code is transmitted.
The process of deriving higher convolutional code rates from a lower rate is
called PUNCTURING , thus in the DVB 2/3 , 3/4 , 5/6 , 6/7 , and 7/8 rates are
progessively cleaner (less noisy) channels. Thus the code rate is often
as the "Puncture rate" .
The Operator decides which rate to run and must either tell the qpsk
which rate is to be decoded or else the qpsk chips must test for the code
The above described convolutional coding is used because of its noise
The Block Coding used by the DVB is called Reed Solomon (R/S) , with
coding by interleaving blocks of bytes.Block coding gives redundancy over
periods so ,as well as improving the final BER , it is resistant to impulse
The DVB uses 204/188 which means 188 bytes come out for 204 in - the rest
The DVB sets a minimum BER & minimum Eb/No (for each code rate).
qpsk demodulator chips must meet or exceed these specs.
The NEXT generation of qpsk demodulators is likely to bypass the 480 MHz IF
stages & demodulate directly form L-band ( or at least mix the L-band signal
down within the chipset) ; this would allow the Receiver Manufacturers to
away with having a separate metal-cased tuner ...instead this function
be integrated into the qpsk chipset & would save the cost of a separate
Chris Muriel , Manchester , 12th August 1997.
[Other mailing lists]