Table of Contents
Advertisement
Quantum and Evolution Series Installation and Operating Handbook
stamping on the 6 timeslots for this destination from the first Rx modem. However Rx
modem 3, still has 18 timeslots to insert, but 12 of the timeslots on the T1 bearer now
contain receive information from Rx modems 1 and 2, and so Rx modem 3 cannot insert all
its data without losing data from another destination. The traditional alternative without
Partial Insert is to use separate bearers for each Rx modem and a DACCS to sort our the
data, but it is lower cost and less equipment to configure and maintain if the DACCS is not
required at all.
Partial Insert becomes mandatory for multi-destinational working when the total receive
data at any one destination exceeds the capacity of the PCM bearers.
Note that the modem can provide multi-destinational operation at aggregate carrier rates
below 1544 kbps using the IBS overhead, but with four backward alarms. This means that
multidestinational working down to 64kbps between nodes become practical (i.e. it does not
require the 96kbps IDR overhead on the links to get multidestinational backward alarm
facilities).
8.8 Choosing Optimum Custom Reed-Solomon Values
Complete variable code rate Reed-Solomon is available if both the Intelsat Reed-Solomon
and `Custom Features` features are both available. (Note that there is nothing special
about the Intelsat values other than that they were approved by Intelsat. They were chosen
because one modem manufacturer could not generate clocks which were not a multiple of
8kHz and these non-optimum values make the satellite clock an 8kHz step).
Where maximum error correction capability is required chose `t` as high as possible (eg
10), and choose `n` as low as possible (eg 60) to maximise the number of bytes than can
be corrected per codeword without adding too much overhead. In this extreme example it
can correct one byte in six (t/n), but the overhead is 50% (overhead n/k, and k = n - 2t, so k
= 40, gives overhead = 60/40). Optimum values may in certain applications mean minimum
delay (particularly at low data rates) and the final paragraph of this section deals with this
aspect.
Due to the synchronisation method used for the interleaver and de-interleaver (Intelsat
specified) on 2 codewords (blocks of `n` data bytes) out of 16 the error correction capability
is slightly reduced by the equivalent of `t`=1. It is therefore best to avoid choosing a very
low value of `t` so this intermittent reduction in error correcting capability is minimised. Note
that in theory error correction codes (100,90,5) and (200,180,10) are equivalent (they can
correct 5 bytes in 100 and 10 in 200 respectively), but the effect of the interleaver
synchronisation words will be more noticeable (equivalent to a loss of `t`=1) when `t`=5
(20% loss) as opposed to `t`=10 (10% loss).
On the negative side the longer the codeword, the proportionally higher the delay through
the encoder / decoder and interleaver / deinterleaver. The delay of the combined encoder
and decoder is approximately 8 x (2 x `n` - `k` + 60) bits, and the interleaver & deinterleaver
a further (8 x `n` x depth) bits (if calculating the delay, use the data rate inc framing and
Reed-Solomon). Intelsat recommend an interleaving depth of 4 for rates less than
1544kbps, and 8 for rates of 1544kbps and above.
8-23
Advertisement
Table of Contents
