On Apr 26 Nick Rees modified the antenna servo. Previous error tolerances of 40" in elevation may have been too tight : if the error is ever greater than this the servo essentially 'gives up'. He also (?) put in an software fix that will null the servo offset assuming it is well behaved. No trimming of offset in the hardware is planned.

A CW run with the 'fixed' TMU inclinometer was made during the afternoon of the 26th and again during the (foggy) evening of the 28th :

    Datasets El Dirn     HST     mean leg temperatures   Humidity
                        start      start   middle   end       %
 (  20020327 90 cw       21:56      4.3      3.8    3.3      80 
    20020424 90 cw       21:00      5.7      4.7    3.7      45 )
    20020426 90 cw       14:15      4.0             5.3
    20020428 90 cw       20:30      6.8             5.8     100

The temperature data from the fixed TMU inclinometer seem OK again (!?) so data were processed into a track model in the conventional way. (A single spike in the temperature data of the 28th was removed) The models resulting are compared below with the model from the 27th of March :

• On the 26th the spikes are reduced in number to only 4 really, and these reach no more than 2" in amplitude. The temperature was similar to that on the 24th - so the thermal component of Justin's theory is not in play, here. The azimuths of the residual spikes are approx 13, 37, 217, 244 which correspond to (wheel,joint) = (2,9/10) (4,3/4) (2,3/4) (2,4/5). It is also noteworthy that the spikes in quadrant 5 are of lower amplitude than their mates in q1. The cable-wrapping system in the plinth area provides tensional differences on the antenna in q1 & q5 : otherwise one would imagine it should behave identically therein.

• There is also a strong wheel ripple signature (one cycle per 8.8degrees, peak-to-peak amplitude of about 1") absent from the data of the 24th. Such a signature was once considered a prelude to a wheel breaking, but has since been interpreted as a realignment of the rotation axis - a process facilitated perhaps by unloading !?

• The data of the 28th were taken when the antenna was slightly warmer, and are free of 'spikes' - which would support the thermal component of Justin's theory. Comparison of the data from each side of the antenna shows that the wheel that has undergone realignment is on the right.

• Updates have been made to the plots of symmetry and strain gauge data. Both show corresponding improvements from the situation on the 24th.

• The new model was installed at 03:30 HST 29 Apr 2002.