how to measure signal duration and statistics for each pulse

[unclebump]

The attached vi has data for the maximum value of six different cutting tools while machining a part for 48000+ consecutive cycles. How would I measure the length of time each tool was used before being changed??(when profile makes a large negative transition.) Since the first three tools are related to the same holes in the part, I would also like to be able to get statistics(average) for each individual peak to see how the life of one tool affects the other tools. I am not sure if I need to do some more averaging or a differentiation to detect slope change points?? The peak detection vi's did not work real well since the peaks are not a constant value each time. This is a version 7 file.

[CoastalMaineBird]

Judging by your graphs, what you need is a differential operation:

Break the wire going from the MEAN output to the graph.

Insert a TIME DOMAIN MATH block (from the ARITHMETIC AND COMPARISON palette)

Make sure it's set for DIFFERENTIAL.

Wire the output to your graph.

See all those negative spikes? That's the transitions in your data. Put a threshold on those, and you'll determine where the changes are.

[unclebump]

That is very close to what I need, but it is missing a few of the transitions. I am wondering about maybe a decibel conversion of the raw data and/or using the signal processing toolkit.

[unclebump]

Here is what I came up with so far. Once I have the ranges for each set of data, a statistics box should give me the rest of the data that I want to evaluate.