Oscilloscope special acquisition modes – EDN


Digital oscilloscopes are normally operated in real-time acquisition mode, where the analog input is sampled and digitized at a user-selected sampling rate and written continuously into the acquisition memory.

There are, however, other acquisition modes available in most oscilloscopes: random interleaved sampling (RIS), sequence mode, and roll mode (Figure 1).

Figure 1 The sampling mode selections in a typical oscilloscope include sequence, RIS, and roll in addition to the normally used real-time mode. Source: Arthur Pini

 These special modes are useful in applications for measuring specific types of signals. RIS mode increases the effective sampling of the oscilloscope rate for periodic signals. Roll mode is useful in displaying low-frequency signals having very long durations. Sequence mode reduces the dead time between acquisitions and is also applied to signals that have low-duty cycles with long dead times between significant events.

Why Nisshinbo Micro Devices is Expanding ASIC/ASSP Business in Europe

06.10.2024

Shift Left for More Efficient Block Design and Chip Integration 

06.10.2024

Open-Source Development Comes to Edge AI/ML Applications 

06.10.2024

RIS acquisition mode

Let’s look at these acquisition modes,starting with RIS mode. RIS mode is a form of equivalent-time sampling that allows the oscilloscope to acquire repetitive signals at very high sampling rates. Multiple acquisitions are combined in RIS to create a composite waveform with a higher effective sampling rate. The trigger event and the sampling clock are not synchronous. The time between the trigger and the first sample in an acquisition is randomly distributed. Oscilloscopes use a time-to-digital converter (TDC) to measure the time delay between the trigger and the nearest sample of each sweep. This delay is called the horizontal offset. The acquisitions are grouped by delay to provide samples spaced by as little as 5 ps, an effective sampling rate of 200 giga-samples per second (GS/s). Selected waveforms are added together, creating a composite waveform, as shown in Figure 2.

Figure 2 Random interleaved sampling creates a composite waveform based on the measured delays between the trigger and the nearest sample point. Source: Arthur Pini

RIS requires a periodic input waveform with a stable trigger. The maximum effective RIS sampling rate is achieved by making multiple acquisitions and selecting those with horizontal offsets, yielding the desired sample spacing. The random timing between digitizer sampling times and the event trigger provides the time variation. The instrument uses multiple triggers to complete an acquisition. The number of acquisitions required depends on the effective sample rate. The higher the effective sample rate, the more triggers are required. Figure 3 compares real-time and RIS acquisitions of a high-frequency sine wave.

Figure 3 Comparing the real-time acquisition of a 5 GHz sinewave at 40 GS/s (lower trace) with a RIS acquisition at 200 GS/s (upper trace). Source: Arthur Pini

The real-time acquisition of the 5 GHz sine is sampled at 40 GS/s and produces a waveform with 8 samples per cycle. Viewing that waveform using linear display interpolation produces a ‘boxy’ display. The RIS acquisition has an effective sample rate of 200 GS/s, yielding 40 samples per cycle and a smoother display. Sin x/x interpolation could be used to smooth the waveform, but as the waveform bandwidth approaches the Nyquist frequency, especially for pulse-like signals, the potential for interpolator errors such as Gibbs ears increases and the RIS acquisition mode produces a more accurate representation of the acquired waveform. In the oscilloscope used, RIS is only available for timebase settings of 10 ns/division or less; similar restrictions will apply to all oscilloscopes.

Roll mode

Roll mode operates at slow sweep speeds, where the sampled data is acquired at a sufficiently low sampling rate and displays the samples in real time as they are acquired. This is important at slow sweep speeds because it eliminates “pre-trigger” acquisition delays. Usually, an oscilloscope holds off display until the acquisition is complete. If you are using a one-second per division horizontal scale setting, you have to wait at least ten seconds before you see the acquired waveform. At slow sample rates used with long acquisitions, roll mode writes to the display as the sample becomes available, eliminating that delay. In roll mode, the trace moves slowly to the left in the manner of a strip chart recorder display (Figure 4).

Figure 4 The simulated progression of a waveform acquired in roll mode. Source: Arthur Pini

This figure shows the progression of an electrocardiogram signal acquired in roll mode over time. The waveform is written starting on the right and moving to the left over time. Each grid in the figure, starting at the upper left, shows the waveform at a later time until it fills the display in the lower right when a trigger event is detected and the acquisition is complete. Roll mode can be entered depending on the sample rate. This oscilloscope enables roll mode at sweep speeds greater than 100 ms/division (higher if more acquisition memory is used).

Roll mode is useful with low-frequency signals like this electrocardiogram waveform.

Sequence mode

Sequence mode is ideal for capturing many fast signal events in quick succession or for capturing a few events separated by long time periods. This mode breaks the acquisition memory into smaller segments and allows multiple acquisitions within the acquisition memory. Sequence-mode acquisitions minimize dead time between acquisitions (typically <1 µs) by holding off display until all segments are captured. Each acquired segment is time-stamped at its trigger time with the real-time clock, the delay between segments, and the elapsed time since the first trigger.

Sequence mode has three main applications:

  1. To acquire data at a high sampling rate when the input waveform has long periods of dead time.
  2. To acquire data with minimum dead time between acquisitions.
  3. To use the trigger time-stamp table to understand the event timing.

For an example of using sequence mode to analyze signals with long dead time, consider capturing several packets of an I2C data signal and measuring the time between packets and the rise time of the data (Figure 5).

Figure 5 Analyzing and I2C data signal for the inter-packet delay and the rise time of the data signal. Source: Arthur Pini

Four data packets were acquired at 20 ms/division at a sampling rate of 5 MS/s. Cursors measure the delay between packets at 43.5 ms. The duty cycle of the I2C signal is small. So, there is a great deal of memory used to show the dead time between packets, resulting in a low sampling rate. Measurement parameter P2 measures the rise time of the signal. It shows a rise time of less than 291 ns with a yellow warning indicator. The sampling rate is 5 MS/s, or a sample period of 200 ns, is the cause of the measurement warning.

Using sequence mode, we can acquire each packet in its own memory segment. The number of segments is model-dependent and is a function of the size of the acquisition memory. In the oscilloscope used, up to 2000 segments can be stored. The packets are much shorter than the whole data stream, and a higher sampling rate can be used (Figure 6).

Figure 6 A sequence mode acquisition captures ten data packets at a sampling rate of 500 MS/s. It also measures the time between packets using the sequence mode trigger time stamps. Source: Arthur Pini

The number of segments acquired is user-selected; in this example, ten packets were acquired with a sample rate of 500 MS/s (2 ns sampling period) horizontal scale of 200 ms/division. A zoom trace, Z1, horizontally expands segment 1. The rise time measurement now shows 28.19 ns, with the green check status icon indicating a valid measurement. The higher sampling rate provides more than adequate time resolution for the measurement.

Below the display grids, the sequence mode time stamps read the time of each trigger using the oscilloscope’s real-time clock. It also reads the time of each segment trigger from the first trigger as well as the time between triggers with a resolution of one nanosecond. The time between segments is nominally 43.6 ms. In this example, sequence mode eliminated the very long dead time between data packets and improved the measurements.

A second measurement provides an example of using sequence mode to minimize dead time between acquisitions. Looking at the signals generated by an ultrasonic range finder. The range finder emits a burst of five 40 kHz ultrasonic pulses each time a measurement is made (Figure 7).

Figure 7 Each measurement of an ultrasonic range finder emits five 40-kHz bursts. After each burst, it waits for an echo from the target before emitting the next transmitter burst, stopping after the fifth burst. Source: Arthur Pini

When a measurement is started, the range finder emits a 40-kHz burst. It waits to detect an echo before emitting the next burst and continues the process until five ultrasonic pulses have been transmitted. In real-time acquisition mode, trying to acquire five individual bursts would not work. After each acquisition, the oscilloscope would pause to display the data acquired and not be ready for the next burst in the series. That is where sequence mode has an advantage. It segments the acquisition memory into a user-set number of segments and acquires one acquisition into each segment without pausing to display the data until all the segments are filled or the acquisition is manually stopped by the user. The latency between acquisitions of each segment is less than a microsecond in the oscilloscope used. During the acquisition, it records the trigger time of each segment. Figure 8 shows an example of the five-segment acquisition.

Figure 8 The sequence mode acquisition of ultrasonic range finder signal with each of the bursts in its own segment appears in the upper left grid. Source: Arthur Pini

Sequence mode waveforms can be displayed in any of five different ways. The display shown is called an adjacent display. The segments can also be overlapped for comparison. They can be displayed as a waterfall display, with each display segment offset vertically. A perspective display provides a three-dimensional view with segments displayed with both vertical and horizontal offset. The final display type places each segment in its own grid and is called a mosaic display.

Each segment can be operated on as an independent waveform. Using zoom traces, each segment can be displayed independently of the others. In Figure 8, zoom traces Z1-Z5 show each of the segments. Any of the oscilloscope’s measurements and math functions can be applied to the segments independently. As an example, the math trace, F2, applies a 2 kHz bandwidth second-order Butterworth band pass filter centered about 40 kHz to segment 1. Averaging can be applied not only to individual segments as they are acquired, but it can also be computed across all the segments in a single sequence, as shown in math trace F1.

The sequence mode trigger time stamps table shows that ultrasonic bursts occur with a nominal 4.2 ms period and confirms that none were missed.

RIS, Roll, and Sequence modes

These examples provide evidence of the usefulness of RIS, Roll, and Sequence modes. They show how you can extend the capabilities of your oscilloscope.

 Arthur Pini is a technical support specialist and electrical engineer with over 50 years of experience in electronics test and measurement.

Related Content



Source link