I have recently begun making use of the NEO (Ra/Dec) ephemerides input mode offered by ACP (http://support.dc3.com/plandoc/neofmt.html). This has generally worked well, but I have run into an strange issue:

Typically I generate plans in which the targets are specified by two Ra/Dec coordinates an hour apart. The software appears to correctly interpolate (or extrapolate) from these if the interpolation time is before the midpoint of the two ephemeris times. After the midpoint, it generates a significantly different result which appears to be the second ephemeris point multiplied by 2, which leads to a failed target acquisition. So if a series of targets are specified with ephemerides spanning 1AM-2AM, the interpolation will be performed correctly until the local time reaches 1:30AM, at which point the interpolation breaks down. Is this a known issue, and is there a workaround I could try (perhaps using 3 points instead of 2)?

I also had a couple of general questions about this input mode:
The documentation above seems to suggest that only ephemeris points on the hour should be used. Is this true, or is it supported to add minutes (presumably by adding a two digit integer between the hour and the right ascension)?
Out of curiosity, how is this interpolation being performed for more than 2 points? I seem to remember reading somewhere that it is using Lagrange Interpolation - is this the case?