Since the last report on HARP pointing, the method of acquiring pointing has been changed from being receptor-centred to being array-centred. 155 data have been logged during these 12 days over Christmas and are displayed below with only average collimation offsets removed:

There is no separation of data (by k-mirror position ?) as seen last time. The scatters (rms) in (dS,dZ) (after removal of two obvious outliers) are (4.3", 3.3") and there is some systematic effect in the dS-vs-Z plot (centre-left) that needs addressing. Part of these (large-ish) scatters might be the use of global collimation terms; nightly variations in collimation are quite possible. The remainder of the scatter - above and beyond the nominal (2",2") we might expect from a 'good' model - is undoubtedly due to the lack of accounting for the k-mirror and HARP-centre misalignments, but it is somewhat reassuring that the coefficients we seek in Per's expressions

    -dAz = x0*sin(2K)     + y0*(1+cos(2K)) - x0'*cos(E-2K) - y0'*sin(E-2K)
    -dEl = x0*(1-cos(2K)) + y0*sin(2K)     + x0'*sin(E-2K) - y0'*cos(E-2K)

20070111
K-mirror positions for these data ( 155 in all) were extracted from the JCMT_LOG.log files and the coefficients in Per's formulae

    -dAz = + xo*sin(2K)         -dEl = + xo*(1-cos(2K)) 
           + yo*(1+cos(2K))            + yo*sin(2K)     
           - x'*cos(E-2K)              + x'*sin(E-2K) 
           - y'*sin(E-2K)              - y'*cos(E-2K) 

The residuals do bottom out at the above position, but the minima are shallow; see below how the residuals in azimuth(left) and elevation(right) vary with the value of the 4 coefficients:

Either

• these formulae do not describe the error functions of the k-mirror
• there are other more dominant terms still to be incorporated, or
• the residuals are dominated by noise rather than any systematics such as these. (Note that there is no information regarding the state of vibration/oscillation of the SMU during these measurements).

Two other terms have been suggested :

     dAz = -x2*sin(K) - y2*cos(K)
     dEl =  x2*cos(K) - y2*sin(K)

It still seems necessary to return to basics - starting with a redetermination of the RxA pointing model.