Data should be taken with a considerable overlap in azimuth in the
1
/5
quadrant. This allows a check on the zero-point stability of
each inclinometer. A set of Point-Wise data may take 2
to obtain and
temperature changes of a degree in that time will not be uncommon. Given the
uncertainty regarding the usefulness of the temperature probes the best that
can be done then is to assume that any zero-point change has occurred as a
linear drift (with time). The overlap data then allows a comparison between
measures at the same azimuths (eg 20degree and 380degree) from which to
establish the baseline correction.
Tests on OUTCW1 showed that the choice of the overlap subsection had little
effect on the determined slopes of the baselines provided that it was reasonably
large, say >10degree. Even selecting an azimuth range particularly full of
track structure (eg. 30-50degree = 390-410degree ), produced similar results.
The changes seen between quadrants 1 & 5 were :
Data Ch 1 2 3 4 5 6 7 8 9
OUTCW1 -14 -34 - -14 14 6 99 -30 5 mV
OUTCCW1 0 -7 - 40 21 -1 9 -19 -3
CW25 -11 -12 - 28 23 2 65 -31 4
Chs 6 & 9 show the corresponding (assumed linear) change in temperature
(4mV 
1degree).
Even if RLB is ignored it would probably not induce
errors of more than 2" (40mV) in any critical component of the model (currently
Chs 1 2 5 & 8). Chs 4 & 7 are not currently used in constructing the track
model, but the large changes seen here may be cause for some future concern.