Summary of LBC IQ 22,23,24 June 2013

This page was created to collect plots made from the 3 nights devoted to studying LBC IQ issues: 2013 June 22,23,24 UT and to discuss these.

Data for LBC Collimation Models

Background: The LBC Collimation Models that are in use were determined from months of archival data (2010 December to 2011 June, in /data/RB_CollTables_link/ and /d1/kuhn/RBCollTables/). These use the mirror positions (PSF total collimation) recorded in the headers of the last extra-focal pupil images before dofpia exits. A 0.8 mm offset was added to the mirror Z position (since these were extra-focal pupils), and the X,Y,RX,RY positions were range-balanced, offline, to predict a new set of X,Y,Z,RX,RY positions as a function of T301 and elevation. No copointing could be done. I think Andrew incorporated the copointing, offline, later?

• Script to do the offline range-balancing is on ohia in ~/iraf/Mytools/oflrangebal.cl

Although the last extra-focal pupil images are, by intention, spherically aberrated, the final Z4 exit correction that accompanies the Z11 exit correction sent to remove the aberration is typically small enough that it should not significantly effect the trend of Z with elevation and temperature. Indeed, from the FPIA logs, we see Z4(exit) is typically ~1000 nm (which corresponds to delt_Z~0.0379 mm). Z4(exit) is computed within DOFPIA as a function only of seeing, so it cannot have much range (Z4(exit)~500-1400 nm or 0.02-0.05mm delt_Z for estimated seeing 0.6"-2").

Z4(exit) = -1.75*Z11(exit) = -1.75 * (FinalZ11zero + FinalZ11scale*seeing), where

Channel	FinalZ11Zero	FinalZ11scale
Blue	-100	-350
Red	-75	-350

To check how well the collimation model was doing, for all object observations made during several nights in the spring of 2013, I plotted the mirror Z position against elevation. For objects observed through transit, the track of mirror Z position along which the object rose did not overlap with the track along which the object set. (Attach example plot). Is this due to hysteresis or uncorrected thermal effects?

To address this question, on two nights (23 and 24 June), we obtained data for co-pointed and range balance collimation models. Our procedure was to:

1. at the start of the night, as usual, clear active optics. Active optics was not cleared during collection of data.
2. Slew to a target with well-known coordinates. USNO-B1.0 stars were used for these targets.
3. Run dofpia to focus and collimate
4. Obtain a pair of 1-sec images of this USNO-B1.0 star.
5. Run lbcrangebal
6. Obtain a pair of 1-sec images of this USNO-B1.0 star again.
7. Repeat 2-6.

The temperature used by the collimation model is T301, the steel temperature at the upper C-ring extension. The range in T301 over the time the data were collected was:

date	T301 (deg C)
20130623	12.5 - 9
20130624	11.8 - 9.4

On 23 June:

• elevation sequence made almost 2 cycles down up and the tracks were (datafiles plotted use the names below: "down 1", "up 1", "down 2", ...):
  • down 1: 85, 75, 65, 60, 55, 50, 45, 35, 25, 15 (dofpia stopped, no data), 20, 15
  • up 1: 15(obs not repeated), 25, 35, 45, 50, 55, 60, 65, 75, 85
  • down 2: 85(obs not repeated), 75, 65, 60, 55, 50, 45, 35, 25, 15
  • up 2: 15(obs not repeated), 20, 30, 40 (twilight)

On 24 June, the elevation sequence repeated that for 23 June, but made only one cycle down and up.

• • down 1: 83, 78, 67, 60, 55, 50, 45, 35, 25, 15
  • up 1: 15 (obs not repeated), 25, 35, 45, 50, 55, 60, 65, 75 (had to close for wind after this).
  • reopened, cleared active optics, and focussed and collimated at 77 deg
    • use the Z data point corrected by 0.8mm on plots (black). *This point falls closer to the original downward track when the T301 term is removed (see sx_lut_20130624.jpg and dx_lut_20130624.jpg).

Dofpia: On both nights, dofpia took many more iterations at low elevation (poor seeing) than at high elevations, and during each of these iterations, it added Z11, Z22 and removed Z4. Is it driven by the fitting of the outer diameter (Z4) or by the ratio Din/Dout (Z11)? We think Z11, and the need to reach the condition that the ratio: Din/Dout = 0.2(blue) or 0.26(red). (Noting here that Din/Dout should be 0.16/0.2 for an unaberrated pupil, but slightly larger for the spherically aberrated pupil to which dofpia converges - and then exits with Z11/Z4 to return to the unaberrated image.) FPIA logs for all three nights are attached: Red/Blue2013-06-22/23/24.Log

The mirror positions along each elevation track are plotted against elevation for SX and DX and 20130623 and 20130624. There are 8 plots total (all in /data/HubMoves).

• sx_lut_20130623.jpg, dx_lut_20130623.jpg, sx_lut_20130624.jpg and dx_lut_20130624.jpg:
  • X,Y,Z,RX,RY as read from the header, i.e. with the T301 correction made by PSF included. The collimation models are plotted for the temperature range, T301 = 9 (blue), 10.75 (black) and 12.5 (red).
• sx_lut0_20130623.jpg, dx_lut0_20130623.jpg, sx_lut0_20130624.jpg and dx_lut0_20130624.jpg:
  • X,Y,Z,RX,RY from the header and with the T301 correction subtracted. The collimation models are plotted for T301=0.
• On all of these the filter focus offset, -0.02 mm for V-BESSEL on LBC-Blue(SX) and +0.069 for V-BESSEL on LBC-Red(DX) is subtracted from the header value (this would be necessary to combine data taken through different filters, but for these nights on which the data are taken through the same filter, it is a constant offset for every datapoint).
• The mirror positions in the header are the total collimation positions which are the sum:
  • Total = Global_offset + Lookup_Table(elev) + Temp_Corr(T301) + Instrument_Offset + Active_Optics + (Manual_Pointing + Pointing)
    • Lookup_Table is a function of elevation only
    • Temp_Corr is a function of T301
    • Instrument_Offset.
      • for Z, this is filter focus offset
      • for X,Y,RX,RY this holds the cumulative corrections sent by lbcrangebal.
  • So for lut0:
    • • Z = Z(hdr) - Tcoeff*T301 - filter_focus_offset
      • X = X(hdr), Y=Y(hdr), RX = RX(hdr) and RY=RY(hdr)
  • And for lut0:
    • • Z = Z(hdr) - filter_focus_offset.
      • X = X(hdr), Y=Y(hdr), RX = RX(hdr) and RY=RY(hdr)

Some observations from the plots:

• The tracks down,up,down do not overlie one another. This is in Z as well as in Y/RX and X/RY and for both SX and DX.
  • High elevation points are separated by delta_Z = 0.35mm, low elevation ones by delta_Z = 0.2 mm. These are with the T301 correction removed. With it, the spread is larger. The spread is similar for both SX and SX.
  • Since the data are co-pointed, trends in Y/RX and X/RY are expected to be coupled for SX and DX.
  • delta_Z = -0.35 mm, the spread in Z at high elevation, corresponds to diff(Z4)~ -9200 nm and delta_Z = -0.2 mm, the spread in Z at low elevation, corresponds to diff(Z4) ~ -5300nm
    • There is a Z4/Z11/Z22 matrix which dofpia uses to account for crosstalk between the application of these Zernike aberrations to M1. So if you apply 1000 nm Z11, you would measure not only the Z11 applied but also Z4 and Z22. If we assume the Z4/Z11 and Z4/Z22 gains used by dofpia, -1.75 and -8: Z4 = -1.75 * Z11 + -8 * Z22. ...
  • Plots of Zernikes Z4, Z5,6 Z7,8 Z11 and Z22 following these down/up/down/up tracks are attached.
    • • between the 1st and 2nd high-elev point:
        • Z11 decreased by 100 nm for DX, about same for SX.
        • Z22 decreased by -250 for both.
        • Z4 decreased by ~ 6000 nm (SX), ~7000 nm (DX).
      • between the 1st and 2nd low-elev point:
        • Z11 increased by ~ 1000 nm for DX, ~500 for SX
        • Z22 decreased slightly, but by ~<100 nm, for both.
        • Z4 decreased by 4800 nm (SX) and 7000 (DX)
  • The track in mirror Z position is reflected, at least partly, by the total Z4.
  • The changes in Z11 and Z22 are smaller than what is expected based on the Z4/Z11/Z22 gains and the change in Z4... other factors involved...
  • Note that the Z4/Z11/Z22 FPIA gains are not the same as those used for M1 by the Gregorian active optics.

Best-focus

During 23 June 2013, we obtained through-focus sequences at 3 different elevations, both along the first downward track and along the first upward track. We obtained these at PA=0 and PA=180 (see Focal Plane Tilt, below).

These focus sequences were analyzed with IRAF.obsutil.starfocus to fit the FWHM vs focus data for a number of stars within the central 5 arcmin x 5 arcmin about the rotator center. The results are in the tables below and plots attached.

Analysis procedure was to:

• create simple FITS file of LBCCHIP2 image only:
  • imcopy lbcb.20130623.hhmmss.fits[2] lfcb.20130623.hhmmss.fits
    • SExtractor runs on MEF or simple FITS, but how can I tell it to run only on extension 2 of a MEF? So I created a simple FITS file.
• run skybox, a hacked version of skyseeing (itself hacked from allseeing), which is in my LBTtools on ohia. It works on a simple FITS file. It is run on the middle image of the focus sequence (the one supposedly at best focus) to create a catalog of stars in ./Misc. The program uses SExtractor to create the catalog. The output catalog is given a name like Bhhmmss.cat.
• run readcat.pl (in /data/20130623/Misc) to read this catalog, select stars within a 5' x 5' box around the rotator center (average rotcen for both Blue/Red taken to be 1056, 2918, so region is [390:1723, 2252:3585] and output a image_cursor file, Bhhmmss_starfoc5.cat, in the format used by starfocus.Catalog files stored in subdir under /data/20130623/Misc.
• run starfocus
  • starfoc @bfoc0a.list focus=focusoff rad=7 sbin=5 swid=5 iter=2 logfile="bfoc0a.log" imagecur="Misc/Bhhmmss_starfoc5.cat"
  • delete stars, starting with those having very high ellipticities or fwhm. For some I deleted a focus point if it looked like it was systematically high (seeing blew up?) and compared the result with and without this point. Though at 25 deg, the seeing did vary (the best-fwhm varied) from ~0.95-1.35", the
  • log files are sometimes named with hhmmss.
• data files (tables below from these) contain best-focus from starfoc, FWHM corresponding to best-focus, sequence name, filename of middle image of sequence and OB name (truncated, so RB_rV_superfoc7_pa should end with pa180). estseeing is seeing estimated by FPIA on the last pupil image obtained before dofpia exited.

Blue: /data/20130623/bestfoc_1k_x_1k.dat

bestfoc	bestfwhm	seq	filename	dd dm ds	uty utm utd hh mm ss	lbcobnam	estseeing
-0.016	2.74	bfoc0a	lfcb.20130623.042659.fits	+74 53 02.91	2013 06 23 04 26 59.451	RB_rV_superfoc7	1.32
-0.033	3.10	bfoc180a	lfcb.20130623.043357.fits	+74 41 15.69	2013 06 23 04 33 57.512	RB_rV_superfoc7_pa	1.32
+0.022	3.02	bfoc0b	lfcb.20130623.051208.fits	+55 13 17.69	2013 06 23 05 12 08.036	RB_rV_superfoc7	1.05
+0.0017	3.20	bfoc180b	lfcb.20130623.051902.fits	+55 09 50.86	2013 06 23 05 19 02.616	RB_rV_superfoc7_pa	1.05
+0.032	4.52	bfoc0c	lfcb.20130623.061507.fits	+25 47 06.75	2013 06 23 06 15 07.682	RB_rV_superfoc7	1.23
+0.07	5.64	bfoc180c	lfcb.20130623.062204.fits	+25 47 55.64	2013 06 23 06 22 04.642	RB_rV_superfoc7_pa	1.23
+0.086	5.21	bfoc0d	lfcb.20130623.071804.fits	+24 57 42.91	2013 06 23 07 18 04.562	RB_rV_superfoc7	1.23
+0.138	5.42	bfoc180d	lfcb.20130623.072457.fits	+24 58 42.76	2013 06 23 07 24 57.236	RB_rV_superfoc7_pa	1.23
+0.008	3.45	bfoc0e	lfcb.20130623.080835.fits	+55 08 32.43	2013 06 23 08 08 35.276	RB_rV_superfoc7	0.95
-0.003	2.85	bfoc180e	lfcb.20130623.081529.fits	+55 05 17.58	2013 06 23 08 15 29.133	RB_rV_superfoc7_pa	0.95
+0.0110	2.75	bfoc0f	lfcb.20130623.084848.fits	+75 03 13.75	2013 06 23 08 48 48.104	RB_rV_superfoc7	0.85
+0.016	2.73	bfoc180f	lfcb.20130623.085539.fits	+74 52 06.29	2013 06 23 08 55 39.805	RB_rV_superfoc7_pa	0.85

Red: /data/20130623/rbestfoc_1k_x_1k.dat

bestfoc	bestfwhm	seq	filename	dd dm ds	uty utm utd hh mm ss	lbcobnam	estseeing
0.022	2.83	rfoc0a	lfcr.20130623.042625.fits	+74 53 32.43	2013 06 23 04 26 25.588	RB_rV_superfoc7	0.94
0.011	2.74	rfoc180a	lfcr.20130623.043327.fits	+74 41 55.49	2013 06 23 04 33 27.416	RB_rV_superfoc7_pa	0.94
0.025	2.90	rfoc0b	lfcr.20130623.051138.fits	+55 13 25.25	2013 06 23 05 11 38.151	RB_rV_superfoc7	0.68
0.0259	3.02	rfoc180b	lfcr.20130623.051833.fits	+55 10 02.35	2013 06 23 05 18 33.463	RB_rV_superfoc7_pa	0.68
0.046	4.26	rfoc0c	lfcr.20130623.061442.fits	+25 47 05.30	2013 06 23 06 14 42.463	RB_rV_superfoc7	1.05
0.057	5.94	rfoc180c	lfcr.20130623.062134.fits	+25 47 52.57	2013 06 23 06 21 34.041	RB_rV_superfoc7_pa	1.05
0.092	4.62	rfoc0d	lfcr.20130623.071735.fits	+24 57 40.86	2013 06 23 07 17 35.744	RB_rV_superfoc7	1.18
0.095	5.64	rfoc180d	lfcr.20130623.072425.fits	+24 58 39.01	2013 06 23 07 24 25.635	RB_rV_superfoc7_pa	1.18
0.03	3.78	rfoc0e	lfcr.20130623.080806.fits	+55 08 39.54	2013 06 23 08 08 06.229	RB_rV_superfoc7	0.65
0.026	3.33	rfoc180e	lfcr.20130623.081455.fits	+55 05 32.80	2013 06 23 08 14 55.213	RB_rV_superfoc7_pa	0.65
0.051	3.0	rfoc0f	lfcr.20130623.084819.fits	+75 03 36.81	2013 06 23 08 48 19.760	RB_rV_superfoc7	0.56
0.062	2.88	rfoc180f	lfcr.20130623.085507.fits	+74 52 51.74	2013 06 23 08 55 07.588	RB_rV_superfoc7_pa	0.56

The best-focus determined from starfocus is seen to depend on seeing or elevation. I concentrated first on plots of best-focus against seeing, given the behavior of dofpia at poor seeing (note: look through FPIA logs. Does dofpia always do this, at any elevation, for est-seeing >~1.5"?)

Attached jpgs, bestfocus_vs_seeing.jpg and bestfoc_vs_seeing_zoom.jpg, plot the seeing (bestfwhm converted to arcsec) on the X-axis and the best-focus (mm), FPIA-estimated seeing (arcsec) and elevation on the Y-axis.

• • Thick red line connects points, (best-focus,seeing)
  • Light magenta line is linear fit to these data (best-focus,seeing)
  • green line connects points, (estseeing,seeing)
  • Thick green line is estseeing=seeing
  • cyan line is linear fit to (estseeing,seeing)
  • Blue points mark elevation vs seeing.

Linear least squares fits to the data of: best-foc[mm] = seeing[arcsec] * a + b and est-seeing = seeing[arcsec]* c + d yield:

chan	a	b	c	d	filter focus offset	filter
blue	0.16	-0.11	0.35	0.81	-0.02	V-BESSEL	according to fit, filter focus offset of -0.02 mm corresponds to seeing 0.56", 0.05 mm for 1"
red	0.075	-0.018	0.66	0.28	0.036	r-SLOAN	according to fit, 0.7" seeing -> 0.036 mm

Notes from plots:

• Both the best-focus and estimated seeing have a stronger dependence on seeing for blue than for red channel.
• The best-focus determined from starfoc does have a lot of scatter at poor seeing. Could the fit be affected by seeing variations during the 5-min superfoc sequence?
  • The 7-point curves, with delta_focus = 0.06mm, pretty well described a parabola or "V". There were one or two focus points which lay above the curve and I deleted these in star-foc when it seemed warranted.
  • The second 75 deg field (b/rfoc0/180f.list) had many stars, a huge and unreasonable, range of FWHM and ellipticities measured by starfoc. I need to figure out what starfoc was doing here.
• This type of plot: best-focus vs seeing;
  • illustrates the scatter in the measurements of filter-focus offset, that it is larger for Blue than for Red.
  • That there is a trend in best-focus with seeing may be explained by the behavior of dofpia in poor seeing conditions.
    • There may be also elevation and temperature dependences.

Does adding the offset between best-focus and nominal filter focus offset explain the difference between in the down/up tracks? No. This offset is very small (0.1 mm max) compared to the difference ~0.2 - 0.5 mm between the down/up Z vs elevation tracks.

Focal Plane Tilt

Procedures and Programs:

• Created IRAF program "startilt" (again modified from others, namely allseeing and iqplot) to:
  • read a list of images taken in a superfocus sequence
    • find the one nominally at 'best-focus' (the focus offsets for V-BESSEL and r-SLOAN are hardwired into the program)
  • create simple fits files, one per extenstion, from this best-focus MEF LBC image
  • run SExtractor on each simple fits file.
  • read the catalog file output by SExtractor, select stars meeting certain criteria, and write out a list of x y coordinates for these stars.
  • create a script (SF.cl) to input to starfocus
    • starfocus is picky --- it sometimes crashes when it comes upon an 'invalid flux profile'. Therefore it is not run within "startilt"
  • cl < SF.cl to run starfocus on each image extension and create a logfile.
• Use perl script, "fptilt.pl" (currently only on my laptop: /home/olga/Programs/lbcIQ/IQProgs) to parse the log output of starfocus and return best-focus, fwhm, ellipticity referenced to the detector coordinates using the header CRPIX1,CRPIX2 values. Where x,y are coordinates in the focal plane and xc,yc are detector coordinates:
  • for chips 1-3, x = xc - CRPIX1(chip) and y = yc - CRPIX2(chip)
  • for chip 4, x = -yc + CRPIX2(chip4) and y = xc - CRPIX1(chip4)
  • and CRPIX1, CRPIX2 are, for the 4 chips:

chip	LBC-Blue		LBC-Red
chip	CRPIX1	CRPIX2	CRPIX1	CRPIX2
1	-1087	2924	-1044	2913
2	1035	2924	1078	2913
3	3157	2924	3200	2913
4	-1709	2271	-1719	2239

Pupil images (LBC and MODS)

Skirt:

• The 'skirt' was seen on LBC pupils throughout these nights. See the example images from 20130622.
  • In the images linked below, the extra-focal pupils are on the left, intra-focal on the right. The blue (SX) are on top and red (DX) below.
  • LBC pupils at high contrast to show "sunburst" around edge of extra-focal blue pupils
  • LBC pupils at lower contrast to show "skirt" and orientation of swing arm
• MODS pupils taken immediately before also show the skirt.
  • MODS extra and intra focal pupils show fuzzy edge - all SX mirror
    • b and r 5: delta_Z = -1.2 mm, t=30-sec. Diameter ~ 182 pixels on r (22")
    • b and r 7: delta_Z = -1.6 mm, t=60-sec. Diameter ~ 234 pixels on r (29")
    • b and r 8: delta_Z = +1.6 mm, t=60-sec.

Orientation of LBC extra-focal pupil:

• Pupil Orientation
  • "Diving" board attached to outer rim of mirror cell near MV hoses, so, nearly 180 deg from the instrument platform, but a little closer to the base of the swing arm than to the 'back', near the gallery. The board can be seen in the extra/intra-focal pupil images attached.

-- OlgaKuhn - 28 Jun 2013