You need to test, we're here to help.

You need to test, we're here to help.

18 June 2013

Back to Basics: Random Interleaved Sampling

Figure 1: This image illustrates the general principle underlying RIS.
Figure 1: This image illustrates the
general principle underlying RIS.
Modern oscilloscopes come with all kinds of bells and whistles, and users might be tempted to invoke them for all sorts of situations. But not every whiz-bang feature of an oscilloscope is applicable all the time. Rather, some features are great in the right applications but disastrous in others.

A case in point is random interleaved sampling (RIS). RIS is an acquisition mode that has been in use for some years now; most oscilloscope makers offer it (some call it equivalent-time sampling, repetitive equivalent-time sampling, or sequential sampling). In part, RIS came about because users of digital oscilloscopes missed the persistence displays found in analog instruments. Persistence displays were great for showing any kind of variability in repetitive waveforms, but the weakness was that there was no way to perform processing on the persistence display. Meanwhile, digital oscilloscopes gave users the unheard-of ability to capture and analyze a single acquisition, but digital oscilloscopes could not analyze a persistence display. RIS was invented to bridge the gap, in a sense. It still does not permit processing of persistence displays, but it does produce waveforms from multiple repetitive acquisitions.

There is a good amount of information available on what RIS is and how it is executed, and this post will not go into great detail on that score. Suffice it to say that RIS delivers a higher effective sampling rate than a given instrument is capable of in other acquisition modes (Figure 1). RIS leverages the time-to-digital converter (TDC), which is the hardware used to place the trigger position correctly between samples. The TDC charges with constant slope from the beginning of the trigger to the next sample. Then it discharges with a constant, but much lower, slope to provide very fine time resolution.

In RIS mode, TDC values are used to bin successive acquisitions and to select and arrange each acquisition for interleaving. The TDC values are totally random as long as there is no relationship between the input waveform and the sample clock of the oscilloscope. Once the acquisitions are taken and binned, they are interleaved together to form a single waveform with a high sample rate.

Where RIS Works Well


RIS is an acquisition mode that comes with some constraints on when it should be used. For one thing, the input waveform must be repetitive. For another, the trigger point must be identical for every sweep. RIS assumes that the same analog waveform is being triggered at the exact same place on every acquisition.

That all sounds quite restrictive. So why use RIS? It's very useful in signal integrity measurements where a repeated stimulus, such as a TDR pulse, is applied to a circuit. In such scenarios, the repetition rate of the stimulus typically allows the circuit to relax to the same state between pulses. RIS also lends itself to applications such as device characterization, where automatic, highly precise measurements of timing and threshold crossings are required.

Basically, RIS fits any application where the trigger event is asynchronous with the sample clock and generates a repetitive waveform. As long as the user is not measuring waveform variability (such as jitter), RIS will generate a low-noise, single waveform result that is suitable for post-processing.

Where RIS is Unhelpful

Table 1: A summary of applications in which RIS should not be used with some suggestions for alternative measurement techniques.
Table 1: A summary of applications in which RIS
should not be used with some suggestions
for alternative measurement techniques.

Conversely, numerous scenarios and/or conditions mitigate against use of RIS (Table 1). It would not be productive or useful in applications such as jitter measurements, eye-pattern analysis, or noise measurements. In all of those cases, the goal is to isolate, measure, and analyze anomalies in the signal of interest. Use of RIS for these types of measurement tasks will result in severely distorted waveforms with no usefulness whatsoever. RIS is not, by any means, an acquisition mode that is useful in all measurement tasks. But when used properly, it is a very helpful tool for generation of an extremely clean waveform that provides highly precise results.

For a more complete treatment of this topic, a white paper on the Teledyne LeCroy website is a great reference source.

No comments: