Table of Contents
Advertisement
OPERATION
Once the sampling rate and duration
values are chose",
the next thing
we must
consider
is the triggering
parameters. Since we are in effect catching a single pulse,
these parameters must be chosen carefully
to place the
sampling
window
on the appropriate
segment of the
pulse--in
this case, on the falling edge.
The first aspect we must determine is whether
to trigger
on the negative or positive slope of the input waveform.
Since we are attempting
to measure the falling edge, we
would obviously opt for negative slope triggering. The next
aspect to consider would be the trigger level. In the exam-
ple of Figure 3-29, we have chosen a trigger level of 9.5V.
Thus, the instrument
measurement sequence will be trig-
gered when the pulse amplitude
drops to 9.5V when go-
ing in the negative direction.
One final triggering
aspect to be considered is whether to
place the instrument
in the single or continuous
trigger
mode. If a one-shot pulse is to be measured, naturally
we
would
use the single trigger mode. However,
the con-
tinuous
trigger mode could be used if a train of identical
pulses is to be measured.
Once the pulse has been captured, and the resulting data
is stored in the measurement
buffer, we can then use the
recall mode to determine
the 10% and 90% amplitude
points. In the case of the pulse in Figure 3-29, these are
simply 1V and 9V amplitude values. The fall time can the"
be determined
from the relative buffer locations and the
programmed
sampling
interval
as follows:
ti = (LlO% - L9w)
x t,
Where:
tf= fall time
L1"% = buffer location number at 10% amplitude
Lqoyi = buffer location number at 90% amplitude
ts = sampling
interval.
For example assume that the 90% and 10% buffer location
points are 150 and 900 respectively, and that the sampling
interval is 1Opsec (100kHz sampling
frequency).
The fall
time under these conditions
is:
tf = (900.150) x 10 x lo-6
tf = Z5msec
3.22.6 Reducing
Noise in the Measured
Signal
Very often 50 or 6OHz noise can creep into a DC input
signal, resulting in erratic or erroneous readings. Such un-
wanted signals can be induced as normal mode noise (ap-
pearing between input high and input low), or common
mode noise (appearing
between input low and chassis
ground).
While the Model 194 has more than adequate
noise rejection for most situations, additional
noise reduc-
tion may be required
in more difficult
cases.
Figure 3-30 shows a sinusoidal noise signal riding on a DC
level. If we assume that the noise signal waveform is sym-
metrical about the DC level, its average value will be zero;
thus, such noise can be effectively cancelled by taking a
number
of samples and the" taking the average of the
samples.
For optimum noise rejection when using this method, the
sampling sequence duration should be exactly equal to (or
exact multiples
of) the period of the noise waveform. The
period of a 60Hz noise signal is 16.667msec. Thus, we might
choose a sampling interval of IO~sec, and program the in-
strument
for 1667 samples, resulting
in a duration
of
16.667msec per sampling sequence. The period of a 50Hz
waveform is 50msec, so a total of 2000 samples would be
programmed
with a 10~s~ interval to obtain the required
20msec sampling
sequence duration.
Once the signal is connected for measurement,
use the
average function
to display the average of the measure-
ment. The degree of noise reduction
will depend on the
symmetry
of the noise signal, as stated earlier. If the
superimposed
noise signal is not perfectly symmetrical,
its DC or average value will not be zero, resulting in a DC
level shift in the final reading. The amount of shift will
depend
on the noise amplitude
and the degree of
non-symmetry
_--
-
-
v w
DC LEVEL
t
Figure 3-30. Noise Superimposed
on DC Signal
3-50
Advertisement
Table of Contents
