Issue: Anomalous points can occasionally during occur during autofocusing can throw-off the determination of best focus position.  Session S837 contained at least 2 examples of this.   Whilst this problem has been seen before and AstroMain's autofocusing code modified so that the worse consequences of such rogue points (such as badly out of focus images),  focusing can sometimes be not as optimal as it could.

Description:  See graphs & tables below

Wide Search Focus Profile 1 Profile has successfully picked point of best focus at 5405, but there is rogue point sitting at 3245 which had much lower FWHM (5.42") than would be expected based on FWHM of its adjacent points (25 - 40") In this case focusing wasn't affected.

Measure Focus Profile (Foc1, wide)    2020-12-02 18:07  (Local) | Image                     | Result    | DRO   | FWHM  | HFD   | Temp  | Max   | Mag    | SN   | Elong | | Taking 0.7s box (C)  1/22   Ok        |  6845 | 31.22 |     - |   3.2 |  1841 |  -8.96 |    - |  1.02 |  | Taking 0.7s box (C)  2/22   Ok        |  6665 | 20.90 |     - |   3.2 |  2051 |  -9.51 |    - |  1.00 |  | Taking 0.7s box (C)  3/22   Ok        |  6485 | 17.47 |     - |   3.2 |  2776 | -10.06 |    - |  1.03 |  | Taking 0.7s box (C)  4/22   Ok        |  6305 | 14.54 |     - |   3.2 |  3754 | -10.54 |    - |  1.03 |  | Taking 0.7s box (C)  5/22   Ok        |  6125 | 11.38 |     - |   3.2 |  4718 | -11.34 |    - |  1.02 |  | Taking 0.7s box (C)  6/22   Ok        |  5945 |  7.50 |     - |   3.2 |  7576 | -12.22 |    - |  1.05 |  | Taking 0.7s box (C)  7/22   Ok        |  5765 |  5.26 |     - |   3.2 | 11936 | -12.77 |    - |  1.03 |  | Taking 0.7s box (C)  8/22   Ok        |  5585 |  3.40 |     - |   3.2 | 18985 | -13.17 |    - |  1.04 |  | Taking 0.7s box (C)  9/22   Ok        |  5405 |  2.62 |     - |   3.2 | 27882 | -13.37 |    - |  1.05 |  | Taking 0.7s box (C) 10/22   Ok        |  5225 |  3.08 |     - |   3.2 | 24500 | -13.31 |    - |  1.08 |  | Taking 0.7s box (C) 11/22   Ok        |  5045 |  3.99 |     - |   3.2 | 16917 | -13.08 |    - |  1.06 |  | Taking 0.7s box (C) 12/22   Ok        |  4865 |  6.35 |     - |   3.2 |  9903 | -12.51 |    - |  1.04 |  | Taking 0.7s box (C) 13/22   Ok        |  4685 | 10.05 |     - |   3.2 |  5253 | -11.74 |    - |  1.04 |  | Taking 0.7s box (C) 14/22   Ok        |  4505 | 14.41 |     - |   3.2 |  3754 | -10.97 |    - |  1.04 |  | Taking 0.7s box (C) 15/22   Ok        |  4325 | 20.59 |     - |   3.2 |  2532 | -10.42 |    - |  1.04 |  | Taking 0.7s box (C) 16/22   Ok        |  4145 | 24.52 |     - |   3.2 |  2165 |  -9.77 |    - |  1.04 |  | Taking 0.7s box (C) 17/22   Ok        |  3965 | 29.68 |     - |   3.2 |  1615 |  -9.15 |    - |  1.02 |  | Taking 0.7s box (C) 18/22   Ok        |  3785 | 36.96 |     - |   3.2 |  1283 |  -8.54 |    - |  1.04 |  | Taking 0.7s box (C) 19/22   Ok        |  3605 | 40.42 |     - |   3.2 |  1207 |  -4.89 |    - |  1.04 |  | Taking 0.7s box (C) 20/22   Ok        |  3425 |     - |     - |   3.2 |   946 |      - |    - |     - |  | Taking 0.7s box (C) 21/22   Ok        |  3245 | 24.68 |     - |   3.2 |   854 |  -9.22 |    - |  2.28 |  | Taking 0.7s box (C) 21/22*  Ok        |  3245 |  5.42 |     - |   3.2 |   832 |  -8.17 |    - |  2.08 |  | Taking 0.7s box (C) 22/22   Ok        |  3065 |     - |     - |   3.2 |   701 |      - |    - |     - |

Wide Search Focus Profile 2 Profile has successfully picked point of best focus at 8785 based on peak FWHM. Although the quadratic curve fit is not used for determining Best Focus Position during Wide Search Focusing, the curve is shown on graphs for info and could be used in future as part of QC process.   In this case the quadratic curve is being thrown-off by anomalous points  with FWHM of 13 to 15 at either end of the profile range (these have very low max pixel value and can be excluded with a cutoff set at a 10% of hightest max pixel  seen in all frames)

Measure Focus Profile (Foc2, wide)    2020-12-02 18:12  (Local) | Image                     | Result    | DRO   | FWHM  | HFD   | Temp  | Max   | Mag    | SN   | Elong | | Taking ShCap Frame   1/16   Ok        |  9085 | 13.62 |     - |   2.8 |   820 |  -8.82 |   14 |  2.75 |  | Taking ShCap Frame   2/16   Ok        |  9035 | 14.40 |     - |   2.8 |   948 |  -9.37 |   23 |  1.35 |  | Taking ShCap Frame   3/16   Ok        |  8985 | 23.43 |     - |   2.8 |  1260 | -10.11 |   43 |  1.12 |  | Taking ShCap Frame   4/16   Ok        |  8935 | 13.02 |     - |   2.8 |  2336 | -10.92 |   82 |  1.11 |  | Taking ShCap Frame   5/16   Ok        |  8885 |  7.73 |     - |   2.8 |  4992 | -11.70 |  148 |  1.06 |  | Taking ShCap Frame   6/16   Ok        |  8835 |  4.82 |     - |   2.8 | 11188 | -12.45 |  246 |  1.08 |  | Taking ShCap Frame   7/16   Ok        |  8785 |  3.37 |     - |   2.8 | 17568 | -12.77 |  298 |  1.07 |  | Taking ShCap Frame   8/16   Ok        |  8735 |  5.18 |     - |   2.8 |  9152 | -12.33 |  227 |  1.08 |  | Taking ShCap Frame   9/16   Ok        |  8685 |  8.97 |     - |   2.8 |  3704 | -11.51 |  129 |  1.04 |  | Taking ShCap Frame  10/16   Ok        |  8635 | 13.16 |     - |   2.8 |  2932 | -10.82 |   77 |  1.08 |  | Taking ShCap Frame  11/16   Ok        |  8585 | 24.02 |     - |   2.8 |  1576 | -10.07 |   42 |  1.07 |  | Taking ShCap Frame  12/16   Ok        |  8535 | 36.37 |     - |   2.8 |  1300 |  -9.47 |   25 |  1.06 |  | Taking ShCap Frame  13/16   Ok        |  8485 | 40.51 |     - |   2.8 |   992 |  -8.86 |   14 |  1.09 |  | Taking ShCap Frame  14/16   Ok        |  8435 | 27.99 |     - |   2.8 |   840 |  -9.25 |   20 |  2.60 |  | Taking ShCap Frame  15/16   Ok        |  8385 | 14.22 |     - |   2.8 |   692 |  -8.81 |   14 |  2.54 |  | Taking ShCap Frame  16/16   Ok        |  8335 | 14.62 |     - |   2.8 |   712 |  -8.85 |   15 |  3.94 |

Standard Search Focus Profile 13 Autofocusing routine picked position of best focus at 5969 based on anomalous frame with ?cosmic ray strike with FWHM of 0.98". based on peak FWHM point instead of the usual quadratic curve fitted value Visually the actual best focus is clearly at around 5200-5250 The QC process fortunately rejected the autofocus position and returned focus to the previous best focus at ~ 5380 In this case focusing wasn't as well optimised as it could

Measure Focus Profile (Foc1)          2020-12-02 20:42  (Local) | Image                     | Result    | DRO   | FWHM  | HFD   | Temp  | Max   | Mag    | SN   | Elong | | Taking 2s box (C)  1/14     Ok        |  6079 | 14.57 | 13.59 |   1.9 |   796 |  -9.12 |    - |  1.82 |  | Taking 2s box (C)  1/14*    Ok        |  6079 | 11.58 | 14.08 |   1.9 |   791 |  -9.17 |    - |  1.99 |  | Taking 2s box (C)  1/14*    Ok        |  6079 | 13.47 | 13.75 |   1.9 |   767 |  -9.15 |    - |  1.77 |  | Taking 2s box (C)  2/14     Ok        |  5959 |  0.98 |  4.58 |   1.9 |  4680 | -10.02 |    - |  1.08 |  | Taking 2s box (C)  3/14     Ok        |  5839 |  8.61 |  9.73 |   1.9 |   962 |  -9.55 |    - |  1.11 |  | Taking 2s box (C)  4/14     Ok        |  5719 |  7.38 |  9.25 |   1.9 |   998 |  -9.76 |    - |  1.04 |  | Taking 2s box (C)  5/14     Ok        |  5599 |  5.55 |  7.80 |   1.9 |  1328 | -10.15 |    - |  1.04 |  | Taking 2s box (C)  6/14     Ok        |  5479 |  4.46 |  7.16 |   1.9 |  1597 | -10.45 |    - |  1.11 |  | Taking 2s box (C)  7/14     Ok        |  5439 |  3.88 |  6.19 |   1.9 |  1981 | -10.68 |    - |  1.10 |  | Taking 2s box (C)  8/14     Ok        |  5399 |  3.78 |  6.52 |   1.9 |  1951 | -10.62 |    - |  1.07 |  | Taking 2s box (C)  9/14     Ok        |  5359 |  3.41 |  5.71 |   1.9 |  2204 | -10.71 |    - |  1.04 |  | Taking 2s box (C) 10/14     Ok        |  5239 |  3.57 |  6.03 |   1.9 |  2096 | -10.66 |    - |  1.07 |  | Taking 2s box (C) 11/14     Ok        |  5119 |  3.24 |  5.87 |   1.9 |  2400 | -10.79 |    - |  1.08 |  | Taking 2s box (C) 12/14     Ok        |  4999 |  3.87 |  6.35 |   1.9 |  1858 | -10.62 |    - |  1.04 |  | Taking 2s box (C) 13/14     Ok        |  4879 |  4.34 |  6.68 |   1.9 |  1757 | -10.52 |    - |  1.04 |  | Taking 2s box (C) 14/14     Ok        |  4759 |  6.02 |  7.80 |   1.9 |  1245 | -10.10 |    - |  1.05 |

Fixes and other actions:
Fixes made to Foc1 and Foc2 Autofocusing in AstroMain 3.32.14 to exclude faint points at extremes of profile having max pixel value < 10% of peak max pixel value,  potential peak FWHM that are offset from weighted best focus position by more than 250 steps (Foc1) or more than 150 steps (Foc2) or are particularly sharp points (FWHM < 1.2") that are likely to be cosmic ray strikes rather than stars.
Fixed 2020-12-04 (AstroMain 3.32.14)

Create a means by which a focus profile dataset can be easily read in AstroMain in order to replay a profile under simulator conditions in order to test tweaks to the autofocus code. At the moment practice/sim profiles are hard coded and can only be changed by modifying the code.
Done 2020-12-08 (AstroMain 3.32.14)

Foc2 Wide Search Focus profile (S831) Original Profile with 'poor'  Quadratic curve - AstroMain 3.32.6  (2020-11-21) Profile has picked point of best focus at 8810, but the rogue point at 8410 had very similar FWHM and was thus very close to being 'picked'.

Profile using revised QC check - AstroMain 3.32.14 (2020-12-08) Points at extreme of range with very low max pixel value are excluded from calculation of quadratic curve

Foc1 Standard Search Focus profile (S831) Original Profile unable to determine best focus position - AstroMain 3.32.6  (2020-11-21)

Profile using revised QC check - AstroMain 3.32.14 (2020-12-08) Reasonable Quadratic curve fit to Best Focus Position

Foc2 Wide Search Focus Profile 2 (S837) Original Profile with 'poor'  Quadratic curve - AstroMain 3.32.12  (2020-12-02)  Quadratic curve is being thrown-off by anomalous points with FWHM of 13 to 15 at either end of the profile range (these have very low max pixel value and can be excluded with a cutoff set at a 10% of hightest max pixel  seen in all frames)

Profile using revised QC check - AstroMain 3.32.14 (2020-12-08) Points at extreme of range with very low max pixel value are excluded from calculation of quadratic curve

Foc1 Standard Search Focus Profile 13 (S837) Original Profile affected by cosmic ray strike - AstroMain 3.32.12 (2020-12-02) Cosmic ray strike in focus frame at 5969 ( FWHM of 0.98" ) was incorrectly taken as best focus

Profile using revised QC check - AstroMain 3.32.14 (2020-12-08) Cosmic ray frame result has been ignored