Mocomtech CDM-570L Manuel D’Utilisation
CDM-570/570L Satellite Modem with Optional IP Module
Revision 4
Forward Error Correction Options
MN/CDM570L.IOM
improvements in error performance without significant bandwidth expansion. The coding
overhead added by the RS outer Codec is typically around 10%, which translates to a 0.4
dB power penalty for a given link. Reed-Solomon codes are block codes (as opposed to
Viterbi which is convolutional), and in order to be processed correctly the data must be
framed and de-framed. Additionally, Reed-Solomon codes are limited in how well they
can correct errors that occur in bursts. This, unfortunately, is the nature of the uncorrected
errors from a Viterbi decoder, which produce clusters of errors that are multiples of half
the constraint length. For this reason, the data must be interleaved following RS
encoding, and is then de-interleaved prior to decoding. This ensures that a single burst of
errors leaving the Viterbi decoder is spread out over a number of interleaving frames, so
errors entering the RS decoder do not exceed its capacity to correct those errors. In the
case of the CDM-570/570L, two different RS code rates are used, according to the mode
of operation.
A 220,200 code is used in transparent closed network modes, and a 200,180 code is used
in framed (EDMAC) modes. (220,200 means that data is put into blocks of 220 bytes, of
which 200 bytes are data, and 20 bytes are FEC overhead.) These two codes were chosen
because they fit well into Comtech EF Data’s clock generation scheme, and they have
almost identical coding gain. When Viterbi decoding is used as the primary FEC, an
interleaver depth of four is used. The increase in coding gain is at the expense of delay.
The interleaving/de-interleaving delay and the delay through the decoder itself can be as
high as 25 kbits. At very low data rates, this equates to several seconds, making it highly
unsuitable for voice applications. Additionally, the de-interleaver frame synchronization
method can add significantly to the time taken for the demodulator to declare acquisition.
A characteristic of concatenated RS coding is the very pronounced threshold effect. For
any given modem design, there will be a threshold value of Eb/No below which the
demodulator cannot stay synchronized. This may be due to the carrier-recovery circuits,
or the synchronization threshold of the primary FEC device, or both. In the
CDM-570/570L, and Rate 1/2 operation, this threshold is around 4 dB Eb/No. Below this
value, operation is not possible, but above this value, the error performance of the
concatenated RS system produces exceptionally low error rates for a very small increase
in Eb/No.
overhead added by the RS outer Codec is typically around 10%, which translates to a 0.4
dB power penalty for a given link. Reed-Solomon codes are block codes (as opposed to
Viterbi which is convolutional), and in order to be processed correctly the data must be
framed and de-framed. Additionally, Reed-Solomon codes are limited in how well they
can correct errors that occur in bursts. This, unfortunately, is the nature of the uncorrected
errors from a Viterbi decoder, which produce clusters of errors that are multiples of half
the constraint length. For this reason, the data must be interleaved following RS
encoding, and is then de-interleaved prior to decoding. This ensures that a single burst of
errors leaving the Viterbi decoder is spread out over a number of interleaving frames, so
errors entering the RS decoder do not exceed its capacity to correct those errors. In the
case of the CDM-570/570L, two different RS code rates are used, according to the mode
of operation.
A 220,200 code is used in transparent closed network modes, and a 200,180 code is used
in framed (EDMAC) modes. (220,200 means that data is put into blocks of 220 bytes, of
which 200 bytes are data, and 20 bytes are FEC overhead.) These two codes were chosen
because they fit well into Comtech EF Data’s clock generation scheme, and they have
almost identical coding gain. When Viterbi decoding is used as the primary FEC, an
interleaver depth of four is used. The increase in coding gain is at the expense of delay.
The interleaving/de-interleaving delay and the delay through the decoder itself can be as
high as 25 kbits. At very low data rates, this equates to several seconds, making it highly
unsuitable for voice applications. Additionally, the de-interleaver frame synchronization
method can add significantly to the time taken for the demodulator to declare acquisition.
A characteristic of concatenated RS coding is the very pronounced threshold effect. For
any given modem design, there will be a threshold value of Eb/No below which the
demodulator cannot stay synchronized. This may be due to the carrier-recovery circuits,
or the synchronization threshold of the primary FEC device, or both. In the
CDM-570/570L, and Rate 1/2 operation, this threshold is around 4 dB Eb/No. Below this
value, operation is not possible, but above this value, the error performance of the
concatenated RS system produces exceptionally low error rates for a very small increase
in Eb/No.
CAUTION
Care should be taken not to operate the demodulator near its sync threshold. Small
fluctuations in Eb/No may cause total loss of the link, with the subsequent need for
the demodulator to re-acquire the signal.
fluctuations in Eb/No may cause total loss of the link, with the subsequent need for
the demodulator to re-acquire the signal.
Table 7-2. Concatenated RS Coding Summary
FOR
AGAINST
Exceptionally good BER
performance - several orders of
magnitude improvement in link
BER under given link conditions.
performance - several orders of
magnitude improvement in link
BER under given link conditions.
Very pronounced threshold effect - does not fail
gracefully in poor Eb/No conditions. Additional
coding overhead actually degrades sync threshold,
and reduces link fade margin.
gracefully in poor Eb/No conditions. Additional
coding overhead actually degrades sync threshold,
and reduces link fade margin.
Very small additional bandwidth
expansion
expansion
Significant processing delay (~25 kbits) - not good
for voice, or IP applications
for voice, or IP applications
Adds to demod acquisition time.
7–3