There should not be any difference between LabVIEW calling your drivers
to take a Thermocouple measurement and VB.NET calling your drivers to
take a Thermocouple measurement.
Can you reproduce this effect? Did it just happen that one
time? If you reboot and try it does the problem represent
itself? Was anything touched on the hardware between one run and
the next? Did you have a cold front move thru your test
facilities between tests?
Problems like this can usually be traced to something repeatable, so if
we can figure out what happened and make it happen again, then we can
(hopefully) figure out how to fix it.