2.1   Radius of Curvature and Conic¶

The Radii of Curvature (ROC) of M1 and M3 were measured at 21 C at the Richard F. Caris Mirror Lab of the University of Arizona [3]. The specificiations were also for T = 21 C [2]. 11.5 C is the average operating temperature on the summit, based on LTS-54 [7]. We therefore correct the measured ROC of M1 and M3 to T=11.5 C and use those values in the integrated Zemax model. The conversion was done with Finite Element model, where the main impact factor is the Coefficient of Thermal Expansion (CTE) of the glass. The values below are also found in LCR-1036 and on this confluence page.

The conic constants of M1 and M3 were also measured at the Richard F. Caris Mirror Lab of the University of Arizona [3]. The specifications are found in LTS-50 [2].

2.2   M1M3 Polished Surfaces¶

The M1M3 surfaces were measured at the Mirror Lab interferometrically under the test tower. There were localized features on the surfaces, called crows’ feet [1], which were measured using the Slope-measuring Portable Optical Test System (SPOTS). Final surfaces including both the lower order polishing errors and the crows’ feet are shown in Figure The polished M1M3 surfaces as measured by the Mirror Lab and the Rubin Observatory teams. (The units are in microns)..

Figure 1 The polished M1M3 surfaces as measured by the Mirror Lab and the Rubin Observatory teams. (The units are in microns).

The M1M3 surfaces as measured are along surface normal. In Zemax, the as-built surfaces are described as grid sag surfaces. To get the surface sag from surface normal measurements, a slope correction has to be applied [10] to the surfaces shown above.

When using grid sag surfaces in Zemax, one of the recommendations is that any low spatial order components, if significant, are better modeled as Zernike coefficients, with the residual represented by the grid sag [6]. Given the small magnitude of the surface errors (<20 nm for M1 and M3), we do not utilize the Zernike coefficients. All the M1 and M3 polishing errors are described by their respective grid sag files.

2.3   M1-M3 Relative Positioning¶

The measured value of the M1-M3 vertex separation is shown in Table M1M3 vertex separation.

According to the M1M3 final report by the Mirror Lab [3] and Document-19421 [4], when viewed in the M1M3 Coordinate System (CS) [11], the best-known M3 optical axis is displaced from M1 optical axis by 0.4 mm along -155 degree direction. In the Zemax CS (ZCS) [11], M3 is decentered from the M1 optical axis by (0.363, -0.169) mm.

For the wedge, Document-17506 [3] doesn’t give the orientation of the M3 wedge relative to M1. Document-19421 [4] reports 0.0161 mm Total Indicator Runout (TIR) along -100.0 degree direction in the M1M3 CS. However, in calculating the x and y-tilts, it should be noted that the decenter of the M3 optical axis discussed above (0.4mm along -155 degree direction in the M1 CS) is actually a rotation around the M3 center of curvature, aka, an aspheric slide. This was confirmed by Mirror Lab personel via email. The x and y-tilts in Zemax need to have two parts:

• a rotation of M3 surface around the M3 center of curvature, along -155 degree direction in the M1 CS, by 0.4/8343.9 radian = 4.8E-5 radian = 2.7E-3 degree. 8343.9 mm is the ROC of M3.
• a tilt of M3 surface around -10 degree direction in the M1 CS (so that the highest point on the perimeter is in the -100 degree direction), by 0.0161/(2508*2) = 3.2E-6 radian = 1.8E-4 degree. 2508 mm is the outer radius of M3. Note that this tilt angle is ~1/15 of the rotation due to decenter.

Projecting these two components to the x and y axis and adding them up gives -1.34E-3 degree of x-tilt and 2.52E-3 degree of y-tilt. Converting into ZCS we get x-tilt of 1.34E-3 degree and y-tilt of 2.52E-3.

In Zemax, the M3 decenters and tilts can either be modeled using coordinate breaks before and after the M3 surface or by defining the tilt/decenter properties of the M3 surface [6].

3.1   M2 Polished Surface¶

The M2 final polished surface was measured interferometrically by Harris. Since M2 is a convex mirror, measurements were first made on multiple segments of the mirror, which are subsequenty stitched together. The linear discontinuities are due to separate ion-figuring runs.

Figure 2 The polished M2 surfaces as determined by the Harris and the Rubin Observatory teams. (The units are in microns).

Early in the M2 polishing process, it was discovered that the M2 surface has a “swirl” feature. This is due to the variation in material properties causing differences in responses to ion figuring. The M2 surface map shown above is the sum of the interferometrically measured M2 surface and the synthesize swirl feature. The swirl feature has a RMS of ~5 nm.

Like the M1M3 surfaces, the M2 surface as measured is along surface normal. To get the surface sag from surface normal measurements, a slope correction has to be applied [10].

4.1   L1L2¶

We have the L1 and L2 TWE maps. Do we have data on L1L2 alignment??? Is there an L1L2 combined TWE???

Figure 3 The raw TWE maps as received from the camera team (L1 and L2 data came in units of microns; L3 data came in units of meters and have been converted to microns).

coordinate systems used for the TWE maps??? I am assuming the same coordinate systems have been used as for the filters, is that true???

pixel size???

A positive number in the data array means the optical path going through that point is longer, i.e., there is larger phase delay.???

4.2   L3¶

5.1   Transmitted Wavefront Maps¶

Figure 4 The raw TWE maps as received from the camera team (The units have been converted from meters to microns).

The center of each data array is the center of the filter. Raw TWE data came in the units of meters. A positive number in the data array means the optical path going through that point is longer, i.e., there is larger phase delay.

The pixel resolution is provided for each filter in the table below.

5.2   Throughputs¶

Need internal reflection data to reproduce graph in LCR-1705 ???

6.1   L1-L2 Separation¶

6.2   L2S2 to Filter S1¶

The alignment of the L1L2 assembly to the rest of the camera (filter, L3, and the detector) is a static variable. This is essentially a hexapod with 6 degrees of freedom (DOFs) on which we can do an one-time adjustment when the camera is fully assembled. Before that this relative positioning contributes 5 optimization variables. The rotation around the optical axis is not used in system optimizations. Also note that the spacing was 349.58 mm in v3.3 optical design, and the as-built spacing is going to be roughly 3 mm less due to LCR-646.

6.3   Filter S1 to L3S1¶

The thickness of the filter varies by band. But the distance between filter S1 and L3S1 is fixed at 72 mm. This was 69 mm in v3.3 design. The filter was moved toward L2 by 3mm, due to LCR-646.

6.4   L3S2 to Detector¶

The spacing between L3S2 and the detector plane is fixed at 28.821 mm. (What is the as-built value???) The original v3.3 design had it at 28.5 mm. This change is the result of a system wide optimization after we found out that the M2 ROC was slightly out of spec. The decision was documented in LCR-1036 and on this confluence page.