23
APPENDIX II
UNIQUE CHARACTERISTICS OF DIGITAL STORAGE OSCILLSCOPES
Digital Storage Oscilloscopes (DSO’s) use a digital sampling
technique to convert analog input signals to a series of digital
words that can be stored in memory. Since digital sampling has
disadvantages as well as advantages, it is important to be aware
of these unique characteristics of DSO’s.
ALIASING
This DSO uses Real Time Sampling when sweep TIME/DIV
settings of 50 sec/div to 20 µs/div are selected. Real Time
Sampling simply means that samples of the input signals are
taken at equal spaces (e.g., every 0.25 ms when the 50 ms/div
range is selected). With Real Time Sampling, a phenomena
called “Aliasing” can occur when the input signal is not
sampled often enough. This causes the digitized signal to appear
to be of a lower frequency than that of the input signal. Unless
you have an idea what the input signal is supposed to look like,
you will usually be unaware that Aliasing is occurring.
A l i a s i n g E x a m p l e
To see what actually occurs when a Digital Storage Oscil-
loscope is Aliasing, perform the following example.
1.
Apply a 10 kHz signal to the input jack and set the sweep
TIME/DIV control to 50 µs/div. You should see about 5
cycles of the waveform on the display. Since the DSO
samples the input waveform 200 times per division, each
cycle is sampled 400 times.
2.
Now change the sweep TIME/DIV control to 2 ms/div.
The display should look crowded. Because the DSO takes
200 samples per division, the sample points are 10 µs
apart. Since the input signal is at a frequency of 10 kHz, it
is being sampled 10 times per cycle. The resulting display
is too crowded to be useful, however, it is not incorrect (it
is very similar to what you would see on a conventional
analog oscilloscope).
3.
Change the setting of the sweep TIME/DIV control to 20
ms/div. Vary the frequency a slight amount (until the
display is readable) to obtain as few cycles as possible on
the CRT. If you were to calculate the frequency of the
signal from the display, you would come up with a much
lower frequency than that of the actual frequency of the
signal at the input jack. As an example, if three cycles are
displayed, the calculated frequency would be
approximately 15 Hz. This is
obviously incorrect. This occurs because the DSO is
taking one sample every 0.1 ms and a 10 kHz signal has
one cycle every 0.1 ms. What is actually happening is that
the frequency is off (not perfectly 10 kHz) by just enough
to cause the DSO to take one sample at a slightly
different place on each cycle of the waveform.
A v o i d i n g A l i a s i n g
Aliasing is not limited to the above example. This phe-
nomenon can occur anytime that at least two samples per cycle
are not taken (whenever the sweep TIME/DIV setting is much
too slow for the waveform being applied to the input).
Whenever the frequency of the signal is unknown, always
begin with the fastest real-time sweep speed (20 µs/div) or by
viewing the waveform in the analog mode of operation first.
N O T E
Viewing one-time events or glitches is not
possible when sweep TIME/DIV settings
higher than 20 s/div are selected. Viewing
one-time events poses no problem with Ali-
asing because Aliasing can occur only with
repetitive waveforms.
EQUIVALENT TIME SAMPLING
This oscilloscope uses a sampling method called Equivalent
Time Sampling when sweep speeds higher than 20 µs/div are
selected. This method permits viewing of repetitive waveforms
to 20 MHz, although the maximum digital sampling rate is 10
Msamples/sec. When the Equivalent Time Sampling mode is
activated, one sample is taken during each cycle. Of course if
one sample is taken during each cycle at the trigger point (the
same point on each cycle), only a flat trace would be produced.
Therefore, it is necessary to take each sample further (in time)
from the trigger point than the last sample. This incremental
delay is deter-mined by the sweep TIME/DIV control setting.
Because 1024 (1 k) samples are needed to fill the display, the
oscilloscope must sample 1024 cycles of the waveform.
Therefore, only repetitive waveforms should be observed in
this mode. Irregularities that are present on an otherwise
repetitive waveform are not likely to show up when the
Equivalent Time Sampling method is used. With only one
sample being taken during each cycle, it is very likely that
glitches and other irregularities will be skipped over.
