https://pfspipe.ipmu.jp/jira This file is an XML representation of an issue en-us 8.3.4 803005 13-09-2019 https://pfspipe.ipmu.jp/jira/browse/INSTRM-1392 Instrument control development <p>Determine the blackspot positions in the PFI coordinate system, in units of <tt>mm</tt>, as a function of elevation angle.</p> <p>Comment by RHL:</p> <blockquote> <p>How is this to be carried out? Using the 71M (and if so, I think we need to include the effects of parallax. I don't know how Kiyoto determined his spot positions, but even if we validate and use his code we need the result <em>in mm</em>)</p></blockquote> INSTRM-1392

Determine blackspot positions on PFI at different elevation angles

Task Normal In Progress Unresolved yuki.moritani hassan EngRun Mon, 11 Oct 2021 16:09:24 +0000 Wed, 2 Aug 2023 09:12:23 +0000 0 4 <p>Just in case... The images were already taken using during Sep run (with focal plane illuminated by the dome room lights). </p> <p>What is the plan for making progress on this one?  Can <a href="https://pfspipe.ipmu.jp/jira/secure/ViewProfile.jspa?name=kiyoto.yabe" class="user-hover" rel="kiyoto.yabe">Kiyoto Yabe</a> provide some details on how he processed the 71M data, and can we provide a file with the known design or as-built positions?</p> <p>When I was in Hilo in September I was confused by as-designed black spot files. <a href="https://pfspipe.ipmu.jp/jira/secure/ViewProfile.jspa?name=chyan" class="user-hover" rel="chyan">chyan</a> gave me <tt>data/pfi/dot/dot_design_asrd_20210422.txt</tt> which I couldn't make line up, and then <a href="https://pfspipe.ipmu.jp/jira/secure/ViewProfile.jspa?name=yuki.moritani" class="user-hover" rel="yuki.moritani">yuki.moritani</a> gave me another file which made more sense.  Do we have a file that we can commit to git and which we believe to be correct?  Either as-built or design, but definitely in PFI mm.  I should remake the plots that backup my claims in this paragraph!</p> <p>(I thought I had already done somewhere before but) I have shared my notebooks to measure the dot position at Subaru here:</p> <p><a href="https://github.com/Subaru-PFS/pfs_utils/tree/u/kiyoyabe/dot_measurements_subaru_202107/notebooks/kiyoyabe" class="external-link" rel="nofollow">https://github.com/Subaru-PFS/pfs_utils/tree/u/kiyoyabe/dot_measurements_subaru_202107/notebooks/kiyoyabe</a></p> <p>I was just starting from designed dot position in mm (dot_pos_final.txt, which is exactly the same as dot_design_asrd_20210422.txt I believe), but this was just to make an initial guess of the position in pix on 71M. So, I think that the transformation in the beginning of the notebook is meaningless.</p> <p>Then I search the dot in disk shape around the initial guess, but the measurement is not good in some dots, so by using the measured data points, I calculate the transformation (including camera distortion) between designed dots (in mm) to dots on 71M (in pix).</p> <p>Then I did the second run and refine the transformation and get the final dot positions (in pix) and write down to the csv file (I did for multiple visits and averaged in deed):</p> <p><a href="https://github.com/Subaru-PFS/pfs_instdata/blob/master/data/pfi/dot/dot_measurements_subaru_20210710_el90_rot%2B00_ave.csv" class="external-link" rel="nofollow">https://github.com/Subaru-PFS/pfs_instdata/blob/master/data/pfi/dot/dot_measurements_subaru_20210710_el90_rot%2B00_ave.csv</a></p> <p>Just for your reference, the measured values are reasonable matched to the images all over the field like this:</p> <p><a href="https://github.com/Subaru-PFS/pfs_utils/blob/u/kiyoyabe/dot_measurements_subaru_202107/notebooks/kiyoyabe/dot_meas_example1.png" class="external-link" rel="nofollow">https://github.com/Subaru-PFS/pfs_utils/blob/u/kiyoyabe/dot_measurements_subaru_202107/notebooks/kiyoyabe/dot_meas_example1.png</a></p> <p> </p> <p>Thank you for the details.  Are you able to produce the positions of the black spots <em>in PFI mm</em>?  Otherwise, we have to redo all this work to convert back from 71M to PFI coordinates, to allow us to work in MCS pixel coordinates.</p> <p> </p> <p>I added a notebook to convert to mm here:</p> <p><a href="https://github.com/Subaru-PFS/pfs_utils/blob/u/kiyoyabe/dot_measurements_subaru_202107/notebooks/kiyoyabe/getDotPositionOnPfiCoordinate_yabe_subaru_20210710.ipynb" class="external-link" rel="nofollow">https://github.com/Subaru-PFS/pfs_utils/blob/u/kiyoyabe/dot_measurements_subaru_202107/notebooks/kiyoyabe/getDotPositionOnPfiCoordinate_yabe_subaru_20210710.ipynb</a></p> <p>The resulting locations are in:</p> <p><a href="https://github.com/Subaru-PFS/pfs_utils/blob/u/kiyoyabe/dot_measurements_subaru_202107/notebooks/kiyoyabe/black_dots_mm_kiyoyabe.csv" class="external-link" rel="nofollow">https://github.com/Subaru-PFS/pfs_utils/blob/u/kiyoyabe/dot_measurements_subaru_202107/notebooks/kiyoyabe/black_dots_mm_kiyoyabe.csv</a></p> <p>My result is a bit different from those in the file in `pfs_instdata`</p> <p><a href="https://github.com/Subaru-PFS/pfs_instdata/blob/master/data/pfi/dot/black_dots_mm.csv" class="external-link" rel="nofollow">https://github.com/Subaru-PFS/pfs_instdata/blob/master/data/pfi/dot/black_dots_mm.csv</a></p> <p>by about 200 um, which may be consistent with the difference we saw last time, but I'm not sure.</p> <p>So, could you please check by using my new location to see whether things make sense?</p> <p>If the new file works, I will push it to more proper place (maybe `pfs_instdata`?).</p> <p> </p> <p> </p> <p>I finally gave up working with 71M for the dot position in mm because I cannot remove non-negligible residual after taking into consideration the non-telecentricity of 71M, instead I found a way to deal with the on-telescope dot data with MCS camera. Although it is not perfect, the result looks reasonable. The attached plot is the position and radius of dots on MCS image (randomly selected; red for the new and blue for the old one), which is converted from PFI mm to MCS pix using PfiTransform class and the new list of dots (in PFI mm) and looks better than before. </p> <p>This is the result for EL=90 deg. and I also have lists for other elevations but I don't think they are reliable because the original images are not good.</p> <p>I'm still working on cleaning up notebooks and making the final products with reasonable filename. I hope to push them to the repository soon.</p> <p><span class="image-wrap" style=""><img src="https://pfspipe.ipmu.jp/jira/secure/attachment/14469/14469_dot_meas_example_20210926.png" style="border: 0px solid black" /></span></p> <p>Princeton plan to obtain data during the May 2022 engineering run as a function of elevation angle for the dot-crossing exercise (<a href="https://pfspipe.ipmu.jp/jira/browse/PIPE2D-1018" title="Plan for data acquisition May 2022 run" class="issue-link" data-issue-key="PIPE2D-1018"><del>PIPE2D-1018</del></a>). This ticket will be addressed by that activity. This ticket may be closed and replaced by a new ticket to capture that activity.</p> <p>It sounds reasonable to me.</p> <p>Updates: During 2022 November run, dot position was measured at EL=60 and EL=30.</p> <p>The data are stored on pfs_instddata:</p> <p><a href="https://github.com/Subaru-PFS/pfs_instdata/tree/master/data/pfi/dot" class="external-link" rel="nofollow">https://github.com/Subaru-PFS/pfs_instdata/tree/master/data/pfi/dot</a></p> <p> </p> <p>Arnaud has written dot position data taken above to <tt>pfs_instdata</tt>.</p> <p>I quickly compared measured position</p> <p><span class="image-wrap" style=""><a id="15446_thumb" href="https://pfspipe.ipmu.jp/jira/secure/attachment/15446/15446_quickComparison_dotPosition_20221130.png" title="quickComparison_dotPosition_20221130.png" file-preview-type="image" file-preview-id="15446" file-preview-title="quickComparison_dotPosition_20221130.png"><img src="https://pfspipe.ipmu.jp/jira/secure/thumbnail/15446/_thumb_15446.png" style="border: 0px solid black" /></a></span></p> <ul> <li>As Arnaud mentioned during the engineering run, difference in position between the three EL is less than 50um.</li> <li>Compared to the measured position at EL=90 (September 2022), ~20 um systematic shift is seen in Y, which is orthogonal direction to gravity. It might be caused by measurement.</li> </ul> <p>We need to discuss if we need to define the dot positions depending on EL. If 50 um of difference is OK, using the position at EL=90 will be fine.</p> <p>We discussed this ticked at ICS/PFI+MCS meeting on 9 Dec 2022 and agreed to check repeatability first.</p> <p>Since the engineering runs in 2022 December and 2023 February were canceled, I move this to backlog now.</p> <p>Dot position was measured at EL=90, 60 and 30 during the July 2023 run. However, it was found that POpt2 was not set correctly at EL=60 and 30. (i.e. Hexapod and ADC was not synchronised with EL), Measured position was clearly affected by distortion of POpt2 (ADC). So repeatability of EL dependency should be checked to the next run.</p> <p>On the other hand, repeatability at EL=90 can be compared among the data from the runs in Nov 2022, Apr/May 2023 and July 2023.</p> Blocks INSTRM-1388 Relates PIPE2D-1018 Development Rank 0|zzsx5s: Sprint PreEngRun4 EngRun3Cleanup Story Points 2.0