WLAN TX Measurements (Option K91) R&S FSL
1300.2519.12 2.132 E-11
with
being the relative clock deviation of the reference oscillator. Normally a symbol–wise timing jitter
is negligible and thus not modeled in equation (12). There may be situations where the timing drift has
to be taken into account. This is illustrated by an example: In accordance to [6] the allowed clock
deviation of the DUT is up to
=
max
20 ppm. Furthermore a long packet with 400_ =symbolsnof symbols
is assumed. From equations (10) and (12), it results that the phase drift of the highest sub–carrier
26
k in the last symbol symbolsnofl _= is 93 degrees. Even in the noise–free case, this would lead to
symbol errors. The example shows that it is actually necessary to estimate and compensate the clock
deviation, which is accomplished in the next block.
Referring to the IEEE 802.11a measurement standard [6], the timing drift
)timing(
,kl
phase is not part of the
requirements. Therefore the "time tracking'' (Tracking Time) is not activated as the default setting of the
R&S FSL–K91/K91n.
The time tracking option should rather be seen as a powerful analyzing option.
In addition the tracking of the gain
l
g in equation (10) is supported for each symbol in relation to the
reference gain
1
g
at the time instant of the long symbol (LS). At this time the coarse channel transfer
function
)L(
ˆ
S
k
H is calculated. This makes sense since the sequence
kl
r
,
' is compensated by the coarse
channel transfer function
)L(
ˆ
S
k
H before estimating the symbols. Consequently a potential change of the
gain at the symbol
l (caused, for example, by the increase of the DUT amplifier temperature) may lead
to symbol errors especially for a large symbol alphabet
of the MQAM transmission. In this case the
estimation and the subsequent compensation of the gain are useful.
Referring to the IEEE 802.11a measurement standard [6], the compensation of the gain
l
g is not part
of the requirements. Therefore the "gain tracking'' (Tracking Gain) is not activated as the default setting
of the R&S FSL–K91/K91n.
How can the parameters above be calculated? In this application the optimum maximum likelihood
algorithm is used. In the first estimation step the symbol–independent parameters
rest
f and
are
estimated. The symbol dependent parameters can be neglected in this step i.e. the parameters are set
to
1=
l
g and 0=
l
d
#
. Referring to equation (10) the log likelihood function
2
lkNNhasep
lTfNNhasep
with
eHarfL
s
gti
l
rests
common
l
symbolsnof
lk
hasephasepj
LS
k
klklrest
gti
kl
common
l
××××=
×××=
××=
%%
= =
+
~
/2
~
~
/2
~
)
~
,
~
(
)min(
)(
_
121,7,7,21
2
~~
(
)(
,,1
)min(
,
)(
)
(13)
must be calculated as a function of the trial parameters
rest
f
and
. The trial parameters leading to
the minimum of the log likelihood function are used as estimates
rest
f
ˆ
and
ˆ
. In equation (13)(13) the
known pilot symbols
kl
a
,
are read from a table.
In the second step for every symbol
l the log likelihood function
lkNNhasep
dlTfNNhasep
with
eHgardgL
s
gti
l
lrests
common
l
k
hasephasepj
LS
k
lklklll
gti
kl
common
l
××××=
+×××=
×××=
%
=
+
#
#
)
)
)
)
/2
~
/2
~
~
)
~
,
~
(
)min(
)(
21,7,7,21
2
~~
(
)(
,,2
)min(
,
)(
2 The tilde generally describes an estimate. Example: x
is the trial parameter of x.
To Next Page
Loading...
