Possible Errors
Phase can be measured on input signal fre
-
quencies up to 160 MHz. However, at these
very high frequencies the phase resolution is
reduced to:
100 360ps FREQ´´
o
Inaccuracies
The inaccuracy of Phase A-B measurements
depends on several external parameters:
–
Input signal frequency
–
Peak amplitude and slew rate for input sig
-
nals A and B
–
Input signal S/N-ratio
Some internal parameters are also important:
–
Internal time delay between channel A and
B signal paths
– Variations in the hysteresis window be-
tween channel A and B
Let us look deeper into the restrictions and
possibilities of using phase measurements.
Inaccuracy: The measurement errors are of
two kinds:
–
Random errors
–
Systematic errors
The random errors consist of resolution
(quantization) and noise trigger error.
Systematic errors consist of “inter-channel de
-
lay difference” and “trigger level timing” er
-
rors. Systematic errors are constant for a given
set of input signals, and in general, you can
compensate for them in the controller
(GPIB-systems) or locally via the MATH/LIM
menu (manual operation) after making cali
-
bration measurements. See Methods of Com
-
pensation on page 4-20.
n
Random Errors
The phase quantization error algorithm is:
100 360ps FREQ´´°
For example, the quantization error for a
1 MHz input signal is thus:
100 1 10 360 0 04
6
ps ´´ ´ °» °.
The trigger noise error consists of start and
stop trigger errors that should be added. For
sinusoidal input signals each error is:
360
2
°
´p
S
N
ratio
Let’s use the example above and add some
noise so that the S/N ratio will be 40 dB. This
corresponds to an amplitude ratio of 100 times
(and power ratio of 10000 times). Then the
trigger noise will contribute to the random er-
ror with:
360
2 100
06
°
´
»°
p
.
The sum of random errors should not be added
linearly, but in an “RMS way”, because of
their random nature. Let’s do so for our exam
-
ples above.
Measuring Functions
4-18 Possible Errors
To Next Page
Loading...
