Oscilloscope Basics 5: Acquisition Mode

The oscilloscope acquisition mode determines how the sample points acquired by the oscilloscope from the analog-to-digital converter (ADC) are combined with the waveform points and displayed. The following acquisition modes are the most common:

• Normal or real-time acquisition mode

This is the most basic sampling mode, in which a waveform point is created from one sample point during each waveform pause. This is the most common and produces the best display for most waveforms.

• Average acquisition mode

Average acquisition mode lets you average multiple acquisitions together to reduce noise and improve vertical resolution. Averaging requires a stable trigger and repetitive waveform. A higher number of averages reduces noise and improves vertical resolution.

• Scroll Mode

Roll mode is a triggerless acquisition mode in which acquired data is displayed in a rolling fashion starting on the right side of the display and continuing to the left side of the display while the acquisition is in progress. Roll mode is useful for manually adjusting low frequency waveforms, finding interference in low frequency waveforms, or monitoring power supply voltage power-up cycles. Because Roll mode is a triggerless acquisition mode, there is no trigger associated with it and all trigger functions are disabled. While the acquisition is in progress, new data will continue to scroll horizontally across the screen. The horizontal reference point is set to the right and is the current time. The waveform data points scroll to the left of the horizontal reference point at the current sample rate.

• Peak Detect Mode

All DSOs (digital storage oscilloscopes) and MSOs (mixed signal oscilloscopes) have mixed signal acquisition memory, which is the number of samples the oscilloscope can digitize for each acquisition cycle. If the oscilloscope's time base is set to a faster time/div setting, such as 20ns/div, then the oscilloscope will always have enough memory to acquire waveforms at that setting using the oscilloscope's highest specific sample rate. For example, if the oscilloscope's highest specific sample rate is 4GSa/s (250ps between samples), and if the oscilloscope's time base is set to 20ns/div, then a depth of 800 points of acquisition memory is all that is needed to acquire and display a complete waveform. At 20ns/div, a complete waveform on the entire oscilloscope screen consists of 200ns of time (20ns/div x 10 horizontal divisions). The memory depth required to fill that time, while still sampling at 4GSa/s, is only 800 points (200ns/250ps = 800).

If you set the scope's timebase to a slower time/division setting in order to acquire a slower waveform and have more time to acquire, the scope may need to automatically reduce its sample rate to fill the time required for the waveform. All DSOs and MSOs can accommodate this requirement. For example, let's assume you want to acquire a slower signal and need to set the scope's timebase to 10ms/division (100ms across the screen). If the maximum depth of the scope's memory is 2 M points, the scope will need to reduce its sample rate to 20MSa/s (100ms/2 M = 50ns sampling period)

While this is not a problem in most cases, since a faster sample rate is not required to acquire a slower waveform, what happens if the input signal contains a mixture of slow and fast features? For example, what if the input signal you want to acquire is a 30 Hz sine wave with a very narrow glitch on it? Acquiring the 30 Hz sine wave does not require a faster sample rate, but acquiring the narrow glitch will require a very fast sample rate.

With the Peak Detect acquisition mode selected, the oscilloscope is effectively downsampling the acquired data at a higher sample rate, rather than sampling the waveform at a reduced rate. For example, let's assume that the oscilloscope needs to run at one percent of its maximum sample rate. This is equivalent to the oscilloscope running at the maximum sample rate, but only storing every one-hundredth of a point, which is "ineffective" downsampling. In Peak Detect mode, the oscilloscope analyzes a group of 200 consecutive samples in real time (sampling at a high rate), and then stores only the highest and lowest digitized values ​​of the group of 200 points, that is, only 2 points. This would be a downsampling factor of 100. So you might ask, why not always use Peak Detect mode? There are some trade-offs when using this acquisition mode. First, the oscilloscope's absolute maximum sample rate is reduced. Second, the stored points will not be evenly spaced. This is an important criterion of the Nyquist sampling theorem. By far, for this particular detection application - using Peak Detect mode is a good choice. But for other detection applications, Peak Detect may not be the appropriate acquisition mode.

• High-resolution acquisition mode

High Resolution mode averages the sequence of sample points within a sample, thereby reducing random noise, making the trace on the screen smoother and effectively increasing vertical resolution. This mode, unlike Average mode, does not require a repetitive waveform.

• Segmented Memory Acquisition Mode

Segmented memory allows acquisition memory to be divided into a set of equal-length sub-records that are equal in overall length up to the total depth of the oscilloscope's memory. Segmented memory is useful for applications where data occurs multiple times in bursts separated by dead time, because it maximizes the oscilloscope's memory depth by capturing only sub-records after a trigger event. Before the advent of segmented memory, the best way to acquire and store data from dynamic, continuous trigger events was to store the acquired data from each trigger to the oscilloscope's hard drive. The time required to save each waveform to hard drive greatly limited overall throughput. With segmented memory, the oscilloscope can use true high-speed acquisition memory to save each waveform instead of using a hard drive. This greatly increases throughput and minimizes dead time between cycles.

Tips: How to choose the appropriate acquisition mode

Normal acquisition mode:

• Used for waveforms with frequency components less than one-quarter the sampling rate.

• Capture waveforms that occur infrequently, such as glitches.

• For single-shot waveform events.

Peak Detect Acquisition Mode:

• Quickly discover waveform anomalies > 50ps wide on slow time bases.

• Check whether the waveform is aliased.

• For single-shot waveform events.

Average Mode

• For periodic waveforms, you can use the normal averaging mode to reduce some trigger noise.

High-resolution acquisition mode:

• Reduce and improve the signal-to-noise ratio on non-periodic (single-shot) waveforms. For periodic waveforms, use the normal averaging mode to reduce some trigger noise.

• Improve the resolution of the signal. As the high-resolution interval increases, the number of effective bits also increases (up to a certain number of bits).

Segmented acquisition mode (Normal, Peak Detect, or High Resolution acquisition mode:

• View waveform events that have a low duty cycle, but have high frequency content.

Roll Mode Collection Mode:

• Used when manually adjusting low-frequency waveforms.

• Interference is found in the low frequency waveform.

• Monitors the supply voltage during power cycles.