Hi Alan,
OK, that error comes when we ask the Agilent API to convert back to drift time from the CCS value it gave for a particular (DT, mz, charge) observation.
Specifically, it returns CCS=437.31581033776683 for (DT=75.802466126052, mz=286.820459998333, charge=3) but when asked for the DT for (CCS=437.31581033776683, mz=286.820459998333, charge=3) value it errors out.
I'm guessing that drift time is outside the useful range, and indeed there's not much signal for that ion and we're probably just picking up noise.
So, two issues - Agilent really ought to be able to round-trip that calculation, and Skyline probably ought not to try to determine a drift time for such a weak signal. Not much I can do about the first, but we can get smarter about the latter.
In the meantime, you can probably move forward by temporarily removing the charge 3 ions from the document before training the IMS library.
Best regards,
Brian