Isolated Subpopulation Analysis

With Hidden Anomaly Location in version 5 of M1 Oscilloscope Tools, we revolutionized the process by which you locate pathological events in your waveforms. HAL finds a large and expanding set of waveform anomalies automatically just by probing the signal, which, when properly deployed, significantly reduces the risk your organization faces of "something stupid getting through to the customer". Version 6 of M1OT gives you a powerful set of tools we call Isolated Sub-Population Analysis (ISPA) that let you understand two important classes of anomalies... those which repeat, and those which can be associated with events/behaviors on other signals. ISPA provides the critical "back end" of the process to take anomalies discovered by HAL, as well as other anomaly-targeting parts of M1, and figure out what their root causes are. ISPA creates an analysis population that is comprised only of the anomalies of interest.... none of the normal events that comprise the majority of the waveform are present in the isolated sub-population (ISP). The ISP can then be explored and analyzed using M1's very rich collection of tools for doing so... You can look at the ISP vs. time and understand things like...

• Duration of the anomaly
• Number of events during individual instantiations of the anomalies
• Amplitude of the anomalies vs. time
• Intervals between instantiations of the anomalies
• Correlation of the anomalies with events in other domains and on other signals
• etc.

There are a number of ways that you can isolate a set of anomalies for analysis, as shown in the diagram below:

In this initial release, the isolation schemes marked with asterisks are available immediately, with the others slated for release later this summer. For more about the ISPA capabilities in M1 OT Version 6, look at Mike Williams' blog.

ISPA is intended to provide the user with a wide variety of ways in which to select some portion or portions of the raw waveform or processed results and present that data separately in order to facilitate further detailed analysis.

The isolated population can currently be presented as either a time vs. time graph (TimeView) or as a histogram (HistoView).  These measurements are 'connected' to the measurements that provided their data, and will automatically update if the original measurement changes.  The ISPA Views have all the functionality of regular Views; they can have their own markers, sync axes with other Views, and so on.

With the current release, there are three ways to specify a subpopulation for further analysis:

• Markers: This will create a subpopulation using only those events laying outside of the horizontal Markers on a TimeView or the vertical Markers on a HistoView.
• Event Triggered: Users can create their own subpopulations based on conditions they set.  For example, a user can specify “Cycle-to-Cycle Jitter > 125 ps This will create a measurement built on only the events that were >125 ps.
• Exception Tracking: Similar to Event Triggered Subpopulations, this method will allow the user to specify their own conditions.  This method then takes it one step further and tracks the failure margins of that condition over a series of acquisitions. The conditions can be anything from very simple (comparing a statistic to a constant) to very complex (multiple statistics and algebraic expressions).

If you have an active subscription, you can start using this immediately, otherwise please contact sales@amherst-systems.com for more information.