Most of the effects of box rotation can be accomplished with intelligent scan patterns. Box rotation would also add to the complexity of the data analysis. -- BrianMason - 03 Oct 2007
I (RMP) agree. It seems none of the receivers on APEX have a box rotator. We should ask them (Karl Menten) if they have suffered from this.
I asked Karl Menten about this and got the following answer: "The bolo cameras, which are in the Cass cabin, don't have a derotator. CHAMP+ (7x350 micron+7x450 micron array), which is in a Nasmyth cabin does. CHAMP+'s derotator has 3 modi (Eq, Horizontal, and cabin) to keep the array's footprint constant in the respective coord system. In the present implementation the scanning angle is calculated and set at the beginning of each scan. To avoid electric path changes in the IF cables etc. the angle is kept constant throughout a scan." I also discussed this issue with Lars Ake-Nyman, who until recently was APEX station manager. He did not know why CHAMP+ was unique in having a derotator, except for the possible need for long integration, although neither Lars nor I could think of what actual science case would require long integrations. -- JeffMangum - 10 Nov 2007
If the polarization properties of the antenna are related to the antenna optics (as I believe they are), then they will be fixed to the (az,el) frame of reference of the antenna. Therefore, a feed rotator would not help to control this.
Just some historical perspective:
The 8-beam 1mm array used on the 12 Meter Telescope in the mid-1990s had a box rotator whose purpose was to fix the (Az,El) frame of the telescope to the sky frame (usually (RA,Dec)). When this decision was made, it was assumed that mapping would be done in the (now old-fashioned) point-and-shoot mode, which necessitated parallactic angle compensation.
By the time the array was installed on the telescope, OTF was well developed for the single-pixel receivers, and OTF became the main (for all practical purposes only) observing mode used with the 1mm array. The only use for the box rotator was to slightly rotate the array so that no two beams followed each other during an OTF raster, allowing for sampling of larger fields for a given amount of time.
The box rotator on the 1mm array was a continuous source of failure for this system. Problems with positional repeatability, mechanical failure, etc. lead to the belief that, if we were to build such an instrument again, we would not include a box rotator.