Agilent Technologies J-BERT N4903B, Руководство пользователя, Страница: 281 / 572

The Q High Level is the mean, calculated from the linear regression curve for
the high level data:
m
=
-A
B
•
Q High Level Std.Dev
The Q High Level Standard Deviation is the σ (Sigma), calculated from the linear
regression curve for the high level data:
s
= -
1
B
•
Q High Level Nr. Points
This is the number of data points used for the calculation of the Q High Level
value. It depends on the setting of the
BER Threshold
and also on the setting
of the
Min BER for Q
parameter.
The minimum for calculating Q-factor values is two points. It is recommended
to use more than 5 points.
•
Q High Level R^2
The R
2
parameter is an indicator that shows how well the converted data points
fit to the straight line. It is calculated as:
R 2 =
S
XY
-
(
S
X
)(
S
Y
)
n
2
S
X
2
-
(
S
X
)
2
n
S
Y
2
-
(
S
Y
)
2
n
The R 2 parameter should be examined before trusting the Q-values. Its
maximum value is 1.0. It must be seen in conjunction with the number of data
points.
For example: Two data points always fit perfectly well, but the resulting Q-factor
calculations are not reliable. On the other hand, 50 data points may reveal a
poor R
2
value. This tells you that the linearization is prone to errors.
If the R 2 value falls below 0.75, the Q-factor calculations are not applicable.
•
Q Low Level
The Q Low Level is the mean, calculated from the linear regression curve for
the low level data:
m
=
-A
B
Advanced Analysis
6
Agilent J-BERT N4903B High-Performance Serial BERT 281