to obtain the same error rate with the internal or external
coder. For example, if a digital user requires an error
for system operation.
rate equal to or less than 1 x 10-5, the minimum (Eb/No),,
c. The signal power levels and variations expected
ratio needed to support this requirement using only
at the 70 MHz input to the receiver must be determined.
differential coding is +9.8 dB.
If the internal
This information is necessary so that the correct internal
coder/decoder is used, the (Eb/No),, ratio required is
range switch setting may be made during installation.
reduced by 2.7 dB to +7.1 dB. If an external KY-
d. If the fault alarm monitor output is to be used,
801/GSC (Viterbi) coder/decoder is used, the (Eb/No),,
the installer must be provided with sufficient information
ratio required is further reduced an additional 2.6 dB to
to connect and test this function.
e. If the modem is not located in or near the earth
c. The requirement for use of error correcting
terminal, insure that the power losses caused by the
coding. equipment is determined by several factors.
interconnecting 70 MHz input and output cables or
These factors are the signal-to-noise density ratio (C/KT)
networks do not significantly degrade equipment
provided by the communications link, the implementation
operation. Also, insure that interconnecting cables or
loss of the PSK modem, the data rate, the bit error rate
networks provide sufficient bandwidth to support the
required by the digital user, and the link margin required.
expected symbol rates of the 70 MHz input and the 70
To aid in understanding how these factors enter into the
MHz output signals.
determination, consider the case in which a digital
communications system must support a 50 kb/s digitized
In general, a system with a gain/loss
voice link. The link margin required is 1.0 dB, and the
variation less than 1 dB over the
digital user needs a worst-case bit error rate of 1 x 10-
range of 70 MHz 2 RS, is adequate
The signal-to-noise density ratio (C/KT) provided by the
to support the symbol rate (RS).
system, which is determined by the earth terminal figure
f. Interface cables or networks, grounding devices,
of merit, geographical location at the terminal,
cable routes, and patching devices must be carefully
implementation losses of the terminal equipment, and
chosen to minimize cross talk between the 70 MHz
power capability or allocation of the satellite, must first be
determined. Further assuming that analysis of the
system indicates an available C/KT of 57.1 dB-Hz, the
effective C/KT may then be determined by subtracting
a. The transmission of digital data over a satellite
the link margin and the PSK Modem implementation loss
the data sent to the digital user from the PSK modem
receiver. These errors are primarily caused by the noise
Effective C/KT = System C/KT (Link Margin + PSK
inherent in the satellite link. The performance of a digital
modem Implementation Loss)
communication link is generally measured in terms of the
Effective C/KT = 57.1 (1.0 + 1.1) = 55 dB Hz
average bit error rate at the digital output of the link.
Knowing the data rate (50 kb/s), the equivalent signal-to-
Average bit error rate is determined by dividing the
noise ratio (Eb/No,) referenced to a noise bandwidth
number of bit errors occurring in a large number of bits
equal to the data rate (R1))can be determined from the
by the total number of bits in the sample. The resulting
number is the probability of error associated with each
bit. For example, if it is determined that 40 errors have
(Eb/No) = C/KT 10 log R1)
occurred in a total of 10,000 bits, the average bit error
rate (or bit error probability) is 40/10,000, or 4 x 10-3.
This function is plotted for convenience in figure 2-5.
b. The bit error rate produced by a satellite
The resulting (Eb/No) is +8 dB. As shown in the
signal-to-noise ratio present at the PSK modem receiver
operating only with differential coding will provide an
input. If a noise bandwidth equal to the bit rate is always
error rate of 4 x 10- at Eb/No, = 8 dB. Error-correcting
used as a reference, the bit error rate as a function of
coding would therefore not be required to provide a bit
signal-to-noise ratio (Eb/No) for any data rate can be
error rate of less than 1 X 10 .
if only differential coding is used, the E N,, ratio required
to obtain a low bit error rate is higher than that required