J. Kissel, R. Lane, T. Sadecki

After the blade tip height adjustment of BSFM01, which (sadly) requires the removal of all the BOSEM Coil/Sensor/IRLED assemblies around the M1 mass, they needed to be reattached and aligned. Attached are the results of the alignment: for both V and Y, we show the magnitude of the transfer functions and coherence between Vertical and Yaw Euler basis drive and all six OSEM basis sensors. Given all of the difficulties in adjusting the position the OSEMs with respect to their flags described below, and after several iterations on each, we had settled for "at least 30 dB of decoupling between to the sensitive and independent OSEM basis sensors." We were able to achieve this for both V and Y (though some are better, and in most it's *easy* to get worse). 

Details:
--------
The alignment process is arduous at best, as we've found from our experience with the QUAD:
(1) The flags are cylindrical, so the left-right adjustment is as sensitive to misalignment as the up-down adjustment (where the up-down, left-right plane is perpendicular to the sensitive axis of the OSEM). This is being fixed with the advent of Matt's flat flags, but we've yet to get any on a working system.
(2) The left-right, up-down range of an OSEM Coil/Sensor/IRLED with respect to a cylindrical flag is at most +/-2 mm from dead center (estimated from D1001794 and D060111), so the centering the flag in this 4 sq. mm box by eye is difficult.
(3) The nominal left-right, up-down adjustment for the OSEM Coil/Sensor/IRLED assembly is two CAM nuts (one for left-right, one for up-down), whose motion is controlled by slots surrounding the opposing CAM nut, carved into the adjustment plate (see D060322). The nuts are only mobile when their respective SHCS is loose; then they maybe rotated with fingers. Ideally, one would loosen one came nut and slide the OSEM along the opposing slot, then rinse and repeat for the other nut. However, both CAM nuts must be loose to enable any motion, and even when loose, there is a good deal of static friction to overcome, often resulting in sudden large motions (larger than the 4 sq. mm box). 
(4) There is a fixed peg and diagonal slot in the corner of the OSEM Adjustment plate, which forces the adjustment motion either along its diagonal, or in a sort of rotation about it. This motion forces one to adjust the up-down, left-right motion simultaneously, making it that much more difficult.

But we know all this.

So in order to get a more quantitative measurement of whether the BOSEMs are centered with respect to their flag, we've performed a test similar to what has been done on the QUAD: at single frequency, drive the suspension in Euler degrees of freedom which are supposed to be independent of others (vertical, yaw), and measure the response in all 6 OSEM basis sensors, moving those sensors which should not couple to the Euler basis drive around until that statement is true. Specifically, we drive in Yaw, which should only be visible in F2 and F3, and move around LF, RT, and SD, then we drive in Vertical, and move F1, F2, and F3. 

The specific frequencies used, 1.1 Hz for Vertical and 1.4 Hz for Yaw, were chosen by taking a spectra of all Euler degrees of freedom and looking for a "lonely resonance" that is seen *only* in V (1.1 Hz) or *only* in Y (1.4 Hz). 

The templates for these transfer functions can be found here:
~/SusSVN/sus/trunk/BSFM/X1/Common/dtt_templates/BSFM_OSEMDiagonalization_VtoLFRT_1p1Hz.xml
~/SusSVN/sus/trunk/BSFM/X1/Common/dtt_templates/BSFM_OSEMDiagonalization_YtoF2F3_1p4Hz.xml