Sub-Sampling Fundamentals - Agilent Technologies 3458A User Manual

composite waveform with a period equal to that of the input signal.
The advantage of sub-sampling is that samples can be effectively spaced at
a minimum interval of 10ns versus 10µs for DCV digitizing and 20µs for
direct-sampling. This means that sub-sampling can be used to digitize
signals with frequency components up to 12 MHz (the upper bandwidth of
the signal path for sub-sampling). Sub-sampled measurements use the
track-and-hold circuit, which has a 2 nanosecond aperture. Sub-sampling
(and direct-sampling) have less trigger jitter than DCV digitizing (see the
Specifications in Appendix A). The disadvantages of sub-sampling are that
the input signal must be periodic (repetitive) and sub-sampling is not a
real-time measurement.
You specify sub-sampling using the SSAC or SSDC command. The SSAC
command selects AC-coupled sub-sampling which digitizes only the AC
component of the input signal. The SSDC command selects DC-coupled
sub-sampling, which digitizes the combined AC and DC components of the
signal.
Sub-Sampling
In sub-sampling, the samples in the composite waveform can be spaced very
closely together. This means that the interval between samples in the
Fundamentals
composite waveform (effective_interval) can be much smaller (and the
effective sampling rate much greater) than in the DCV or direct-sampling
methods. For example, assume you need to digitize a repetitive 10kHz input
signal with a 5µs effective_interval between samples. This is a sampling rate
of 1/5e-6 or 200,000 samples per second, (This application would be
impossible using DCV or direct-sampling since their maximum sampling
rates are 100,000 and 50,0000 samples per second, respectively.) Figure 31
illustrates how this can be done using sub-sampling. The effective_interval
is specified as 5µs and specified number of samples is 20. The
effective_interval and the total number of samples are specified by the
SWEEP command. After specifying the effective_interval and the number
of samples, the multimeter calculates how many bursts (a burst is a group
of samples) it needs to make and how many samples will be in each burst.
For this example, on the first period of the input signal, the multimeter takes
a burst of 5 samples. On the second period, the multimeter delays the trigger
point by 5µs and takes another burst of 5 samples. On each of the remaining
two periods, the multimeter delays the trigger point by another 5µs and takes
a burst of 5 samples. As shown in Figure 32, when all the samples are
arranged in the proper sequence, the result is one period of the input signal
consisting of 20 samples spaced at 5µs intervals. In this example then, the
effective sampling rate is 200,000 samples per second.
140 Chapter 5 Digitizing