A demonstration of sample mode, peak detect mode, high-resolution mode, envelope mode and average mode.

Hi and welcome again to our 82nd Test and Measurement Video. The topic before us is sampling modes in the Tektronix MDO3000 Digital Storage Oscilloscope.

Virtually all oscilloscopes and spectrum analyzers manufactured today are digital rather than analog. The advantage is that once digitized, signals can be processed, measured, analyzed, stored and recalled for later examination. The challenge that had to be confronted was that signals in nature and human-made are predominantly analog rather than digital, so conversion is required before processing can take place. At the heart of analog to digital conversion is a process known as sampling, which as we shall see can be accomplished is accordance with several modes. The differences among them have to do with the locations relative to the original waverform where these samples are extracted.

We know intuitively that the accuracy or fidelity of the manufactured digital signal depends upon the number of samples extracted per unit of time. If you took only one sample for each oscillation, you would have no information other than the fact that the wave exists and has a certain minimum amplitude.

The Nyquist-Shannon sampling theorem establishes the minimum sampling rate that permits a discrete sequence of samples to capture all the information from a continuous-time signal of finite bandwidth. This sample rate is twice the highest frequency component of the signal. (The frequency and sample rate are both expressed in the same units of time, such as seconds.)

The key concept here is that the frequency to be used in the calculation is not that of the fundamental, except in the case of a sinusoidal signal, but rather that of the highest-frequency harmonic. Thus, the sampling rate is based not only on the frequency, but also on the complexity of the waveform.

To start, we’ll apply a sine wave from the internal arbitrary function generator to Analog Channel One input. Here we see the sine wave displayed on the screen. The default settings, shown in the AFG bar, are frequency at 100.00 kHz and amplitude at .5 volts peak-to-peak. These values can be adjusted by pressing Waveform Settings and turning Multipurpose Knobs a and b, but they are just fine for what we are doing today.

To change sampling modes, first press Acquire. This frequently-used button, located between the position and scale knobs in the Horizontal section on the front panel, opens up the horizontal acquisition menu below the display.

What we are interested in right now are sampling modes, so we press the soft key associated with Mode, which is the first menu selection on the left.

This opens the vertical Acquisition Mode menu on the right, which permits the user to select one of the sampling modes.

Sample mode is the default, and as such it is the mode that is used most of the time. It is simple and fundamental. It fulfills the sampling function by retaining the first sampled point from each acquisition interval.

Peak Detect sampling mode uses the highest and lowest of all samples contained in two consecutive acquisition intervals. This mode works only with real-time, non-interpolated sampling. It is useful for catching high-frequency glitches.

High-Resolution mode calculates the average of all the samples for each acquisition interval. This mode also works only with real-time, non-interpolated sampling. High-resolution mode provides a higher-resolution waveform at the expense of lower bandwidth.

Envelope mode finds the highest and lowest record points over all acquisitions. Envelope mode uses Peak Detect mode for each individual acquisition.

Average mode calculates the average value for each record point over a user-specified number of acquisitions. Average mode uses Sample mode for each individual acquisition. Use Average mode to reduce random noise.

As you can see, there isn’t much difference in the appearance of the sine wave in each of these sampling modes. That is because of its spectral purity. But look at the difference when we display Noise, which is a random, high bandwidth signal. In Average mode the difference is dramatic, especially when we increase the number of acquisitions that are averaged to the maximum, which is 512. This suggests a commonly used application for signal averaging, which is to eliminate noise from a signal that you want to study. The technique is sometimes preferable to bandwidth limiting, which often works quite well but is sometimes problematic when the signal itself requires the full bandwidth of the instrument.

Thanks for watching. New Videos are added periodically, so check back frequently.