It can be done but does require a bit of code to accomplish that end goal. Standby for a word salad to describe how it can be done.
We can use "extra" plots that are hidden in the plot legend stuffed down at the bottom out of sight. Think of the "extra" plots as the part of the plot that is out of range. Set their color to red to flag the out of range.
As the data comes in each value that will be plotted has to be evaluated if it is in-range or not. Depending on that evaluation the readings are either included for plotting (when in range) or replaced with NaN to effective "lift the pen" for that plot.
At the same time the out of range values will be used in the "extra" plot.
So for each plotted channel you will have two plots, one for in-range and another for out of range. If all of the reading are in-range the "extra" plot will be of the same length as the normal but all of the values will be "NaN".
If they are all out of range then the normal plot will be all "NaN" and the "extra" plot will have all of the readings.
Now if you get past that work there is only one "detail" that can you attend to which is connecting the gap that will be visible after a good reading and before a bad reading. That can be covered by using the last good point as the first point in the bad plot and vise versa.
Ben
