UT20080219-22 were four nights of excellent, stable weather. 51 HARP data were logged and are shown below:

Mean collimation offsets (CA,IE) are (+15.5", +36.1") - to be compared with the nominal (+15", +35") (see /jac_sw/itsroot/src/harpb/XML/instrument_harpb.ent). RMS scatters in (dS,dZ) are (1.9",3.2"), with no strong suggestion of any systematics. The correction to the flexure term, TF made last time seems to have worked OK.

However, as before, the performance, especially in elevation, is v.poor. The reasons may be several-fold: eg :

1. the K-mirror misalignments
2. the deviations of the basic HARP pointing model from that of RxA
3. uncorrected thermal effects (which hit the elevation residuals more than the azimuth residuals)
4. anomalous refraction (ditto)

The recent analyses of HARP "AZ-EL" fivepoints and subsequent jiggle-maps taken with an "RA-DEC" configuration for the K-mirror might provide an answer to the first part, thus allowing a determination of the second:

After correcting these 51 HARP data for the K-mirror coefficients derived on 20080222, the adjusted residuals are plotted below against azimuth (left) and elevation (right):

RMS scatters in (daz,del) are now (1.8",2.9") which is only marginally better than without the correction. However, there are four rather odd elevation residuals associated with targets approaching transit in the north (these comprise 2 of 4 measures of UCam, plus 2 of 4 of CRL618), although in general there is (again) no suggestion of systematics that might be amenable to extraction using TPOINT.

NB: Without the worst three only the scatters become (1.8",2.3") so these three may be claimed to lie more than 3 sigma from the (true !) mean . . . and may be removed !? The raw scatters of the corresponding original 48 data are (2.0",2.7"). No further improvement is achievable by adjustment of the TPOINT parameters.

Conclusion
The K-mirror solution derived recently - together with some juducious pruning of allsky pointing data - may indeed be able to improve the all-sky pointing of HARP to a level of about 2" rms in each coordinate. Another ~20 "RA/Dec" data are requested in order to confirm the validity of the current data and/or justify the elimination of the above 3 (or 4) or any further possible outliers.