Donut and 4MAP Study results 12262012¶
During engineering time on 12/26/2012 Alistair took a series of images, out of focus, with a range of primary airpad (4MAP) settings from Roberto which produced various aberration patterns in the primary mirror.
I analyzed the donuts in these images using my standard offline donut pipeline, comparing the aberrations in each of these images against the reference wavefront. An example of this comparison is shown below, where the first row of plots has a histogram of the difference between each aberration and the reference, the second row is a histogram of the difference between each aberration and the reference after the aberration surface has been offset and rotated to best match the reference, the third row is a scatter plot of raw differences now versus the focal plane x & y, and the fourth row is the scatter plot of the differences after the offset & rotation. You can see in this image that the raw differences are quite large for Z5 and Z6 which are the two Astigmatisms, and this image is one that should have a large Astigmatism from the primary. Note that the scatter plots show that there is also a rotation of the Astigmatism surface compared to the reference - this is due to mis-collimation not the primary mirror Astigmatism. Any aberrations induced in the primary should appear at the same level at all points on the focal plane.
Here are results:
Comparing the induced 4MAP astigmatism magnitude (in nm) vs. the Donut measurement:
Comparing the induced 4MAP astigmatism angle (in deg) vs. the Donut measurement:
Comparing the induced 4MAP trefoil angle (in deg) vs. the Donut measurement:
Clearly the angles agree extremely well. The magnitude of astigmatism compares ok - the many points taken at 2000nm have donut values with mean of 1860nm and RMS of 250nm, and the overall correlation is in the ballpark. Sources of scatter could include tracking jitter, which does shift the donut measured astigmatism, dome/mirror seeing, as well as the accuracy of the 4MAP system. Also the donut's reference may have some overall offset. There were also four images taken with 2000nm of trefoil, these have a mean of 1600nm with an RMS of 140nm. In conclusion, the donuts do measure the primary aberrations reasonably well.
Next steps include the building a LUT of primary aberrations as a function of Altitude and Azimuth, for use by the 4MAP active optics, and testing how well the 4MAP system can correct the measured deviations in aberration.
For cognoscenti, all the information is summarized in the spreadsheet below. Note that I haven't yet
analyzed the donuts for Quadrefoil, will do soon.