With the assumptions above, the measured rotation of the right and left
A-frames should allow calculation of the roll of the elevation axis.
It would appear that the measurements of roll from the TMU platform are
redundant, or that by comparing the two determinations we may be able to
estimate the rigidity f, and correct the yaw estimated from
the A-frame measures.
Such a comparison reveals the general acceptability of the scale of conversion
from A-frame motion to roll (1" to 0.46"), but shows significant
systematic differences that prevent a more quantitative estimation of f.
The cause of the difference is not clear, but the hypothesis that the A-frame
data are badly affected by shocks (after the antenna goes over a bump)
could explain them.