
17
STEL-2060C
Input
Initial State
Alternate States
1
2
G1
G1
n
G2
n
* G2
n+1
G2
n
* G2
n+1
G1
n+2
G2
n+1
G1
n+2
G2
n+2
*
(No delay)
(One symbol
delay)
(Two symbol
delay)
PARL
Input
Initial State
Alternate State
0
G1
G1
n
N.A.
G2
n
G2
n
N.A.
G1
n+1
G2
1
G1
G1
n
G2
n
G2
n
G1
n
G2
Input
Initial State
Alternate States
3
1
2
4
5
G1
G1
n
G2
n+1
G2
n+2
*
G2
n
* G1
n+2
G2
n+3
(No Invert/Swap (Invert/Swap
No delay)
G2
n
* G1
n+2
G2
n+3
G1
n
G2
n+1
G2
n+2
* G2
n+1
G2
n+2
* G1
n+4
G2
n
* G1
n+2
G2
n+3
G2
n+1
G2
n+2
* G1
n+4
G2
n
* G1
n+2
G2
n+3
(Invert/Swap
One symb. delay)
*G2
n+1
G2
n+2
*G1
n+4
G1
n+2
G2
n+3
G2
n+4
(No Invert/Swap
Two symb. delay)
G1
n+2
G2
n+3
G2
n+4
*
G2
n+1
G2
n+2
*G1
n+4
(Invert/Swap
Two symb. delay)
G2
(No Invert/Swap
One symb. delay)
No delay)
3. RATE
2
/
3
, PARL = 1 (QPSK MODE)
NODE SYNC SEQUENCES
1. RATE
1
/
2
2. RATE
2
/
3
, PARL = 0 (BPSK MODE)
BER PERFORMANCE
The coding gain obtained by the use of Convolutional coding
and Viterbi decoding is extremely dependent on many
parameters. Not surprisingly, the code rate is a primary
factor, but so are the bit error rate (BER) and amplitude of the
input signal. The BER affects the coding gain because the
error correction capability of the Viterbi decoder is
dependent on the statistics of the errors, specifically the
clustering of errors. As the BER of the input signal increases,
so does the clustering, causing a reduction in the error
correcting capability of the device, along with the coding
gain. The signal amplitude is important because of the
weighting given to the signal amplitude as an indication of
the likelihood of an error in a given symbol pair.
E
b
/N
o
BER
10
–2
10
–3
10
–4
10
–5
10
–6
10
–7
R =
1
/
2
R =
2
/
3
R =
3
/
4
R =
7
/
8
10
9
8
7
6
5
4
3
2
Uncoded
10
–1
11
Consequently it is important to maintain the signal
amplitude at an optimum level in order to maximize the
performance. The performance curves shown above were
measured using a digital link simulator with the signal level
set at one half of full scale; i.e., the signal amplitude without
noise ranged from 101 to 001 in signed magnitude format, or
101 to 010 in offset binary format. The coding gain under
these conditions is about 0.2 dB less than that under
optimum signal level conditions. The performance of the
STEL-2060C is shown here for unpunctured operation
(Rate
1
/
2
) as well as punctured operation at the rates for
which internal depuncturing is supported (Rates
2
/
3
,
3
/
4
and
7
/
8
). The error rate for uncoded data is shown for
comparison.
Powered by ICminer.com Electronic-Library Service CopyRight 2003