Random err or =
quant err start trg err stop trg err.. .. ..
222
++
The total random errors are thus:
0040606085
22 2
....++»°
(single-shot)
What about random errors caused by internal
amplifier noise? Internal noise contribution is
normally negligible. The phase error caused
by noise on the signal, whether internal or ex
-
ternal, is:
360
2
°
´p
S
N
ratio
For an input signal of 250 mV
rms
and the typi
-
cal internal noise figure of 250 mV
rms
gives us
a S/N-ratio of a minimum of 60 dB (1000
times). This gives us a worst case error of
0.06°. Increasing the input signal to 1.5 V
rms
decreases the error to 0.01°.
Another way to decrease random errors is to
use the statistics features of the instrument and
calculate the mean value from a number of
samples.
n
Systematic Errors in Phase
Measurements
Systematic errors consist of 3 elements:
–
Inter-channel propagation delay difference.
–
Trigger level timing errors (start and stop),
due to trigger level uncertainty.
The inter-channel propagation delay differ
-
ence is typically 500 ps at identical trigger
conditions in both input channels. Therefore,
the corresponding Phase difference is:
<0.5 ns ´360°´ FREQ
See the following table.
Trigger level timing error
The “trigger level timing error” is depending
on two factors:
–
The actual trigger point is not exactly
zero, due to trigger level DAC uncertainty
and comparator offset error.
–
The two signals have different slew rates
at the zero-crossing.
Every counter has input hysteresis. This is
necessary to prevent noise to cause erroneous
input triggering. The width of the hysteresis
band determines the maximum sensitivity of
the counter. It is approximately 30 mV, so
when you set a trigger level of 0 V, the actual
trigger point would normally be +15 mV and
the recovery point –15 mV. This kind of tim
-
ing error is cancelled out by using hysteresis
compensation.
Hysteresis compensation means that the mi
-
crocomputer can offset the trigger level so that
actual triggering (after offset) equals the set
trigger level (before offset). This general hys
-
teresis compensation is active in phase as well
as in time interval and rise/fall time measure
-
ments. There is a certain residual uncertainty
of a few mV and there is also a certain tem
-
perature drift of the trigger point.
Possible Errors 4-19
Measuring Functions
160 MHz
28.8°
100 MHz
18.0°
10 MHz
1.8°
1MHz
0.18°
100 kHz
0.018°
10 kHz and below
0.002°
Table 4-2 Phase difference caused by
inter-channel propagation delay
difference
To Next Page
Loading...
