TangraTests

Tangra Tests Notes

Dec 17, 2011
For all, 'no integration', 'no preprocessing', start at frame 1. Frame 96 Julie pans, must manually reposition circle, and I stop analysis when it loses tracking again after it's clearly centering on the bus trail. blue circle of guiding star first goes to the tail at frame=120, and quite often by frame=150. tolerance always at max=3.5px.

A='aperature photometry', P='psf photometry, 'X = optimal extraction photometry', T='tracked...', U='untracked', F='Full or almost full disappearance', W='wind or strong shaking', K='strong flickering', S='stop on lost tracking', and n=not selected, and following Pavlov suggestion, gtr=(obj1 guiding star=target, obj2 occulted star=random background spot manually positioned).

object	A/P/X phot, radius of circle	settings	psf fit area	output filename	Frame=30 Sig-Back, noNorm	Frame=30 Signal only, noNorm	Frame=30 Background only, noNorm	comments
bus	A,2.5px	TWKS gtr	9px		2120	3011	891	tops out Fr~70, by F~150 blue circle often trails head,
bus	A,2.5px	TFWKS gtr	9px		2120	3011	891	tops out Fr~70, by F~150 blue circle often trails head
bus	A,3.0px	TWKS gtr	9px		2501	3807	1306	tops out Fr~70, by F~150 blue circle often trails head
bus	A,4.0px	TWKS gtr	9px		2712	5074	2362	tops out Fr~125, blue circle often trails head by Fr~150
bus	P,2.5px, avg backgrnd	TWKS gtr	psf fit=9px		1833	2728	895	tops out Fr~120, flattop psf past Fr=35, val=2000
bus	P,3.5px avg backgrnd	TWKS gtr	psf fit=9px		3547	5346	1799	tops out Fr~120, flattop psf past Fr=35, val=2000
bus	P,2.5px, psf fit backgrnd	TWKS gtr	psf fit=9px		2106			psf fitting background causes wild photometry once sat pixels exceed circle
bus	P,2.5px, psf mode background	TWKS gtr	psf fit=9px		2106			psf fitting background causes wild photometry once sat pixels exceed circle
sao 99876 v=6.9,sp=K	A, 2.5px, avg backgrnd	TWKS gtr	psf fit=9px	S99876-A25.csv	avg=2121			Fr=43-87 to get closest to unvignetted
sao 99876 v=6.9,sp=K	A, 3.5px avg backgrnd	TWKS gtr	psf fit=9px	S99876-A35.csv	avg=2813			Fr=43-90 to get closest to unvignetted
sao 99876 v=6.9,sp=K	A, 4.0px avg backgrnd	TWKS gtr	psf fit=9px	S99876-A40.csv	avg=2921			Fr=43-90 to get closest to unvignetted
sao 99876 v=6.9,sp=K	A, 6.0px avg backgrnd	TWKS gtr	psf fit=9px	S99876-A60.csv
sao 99876 v=6.9,sp=K	A, 6.0px mode backgrnd	TWKS gtr	psf fit=9px	S99876-A60.jpg, S99876A60s.jpg for signal only	avg=2775
bus, from julieVD2	A, 6.0px mode backgrnd	TWKS gtr	psf fit=9px	Bus2-Ap6pxModeUnc.csv Ap6pxModeVig.csv corrected for bad pix and vignetting				F=182 and beyond has saturated pixels

Tentative Conclusions
* using "full or almost full disappearance" doesn't affect Bus photometry, other things equal; Hristo says only use for occn's.
* bus gets psf flat-topped by saturation by frame 38 or so. Use frame=30 for testing above. Click on light curve point to see details of that point.
* Hristo says use 'background mode' which is 3xMedian - 2xMean like in DAOPHOT and more tolerant of contam by nearby's
* Use 6pixel circle, as a good compromise between contamination, and getting all of bus during unsaturated portion.

Determined vignetting using 24 time integrated circles r=6px at various radii.

Generated new avi; julieVD2.avi which has its first frame when bus first appears, and last frame past saturation/merging of bus and capsule.

Bus2-Ap6pxModeUnc.csv needs to have bad records deleted, when tracking failed (a couple of dozen times or so). Frame delete list, from exam of light curve...
101,102,122,136-9,147,148,159-162,223,225,227-236,238-240,264,273-286,293+

To multiply a column in excel by X...
Put X in an empty cell. Copy it.
click on top member of column, then shift-click on bottom member to highlight entire column
right-click inside the column Paste Special and choose multiply and OK.and the numbers will all get multiplied

Extinction correction: Airmass X = 1/(sin(alt) +.025exp(-11sin(a))) and there is .28 mag of extinction per airmass. Altitude of the comparison SAO 99876 was 13.068deg. Altitude of the bus was 11.65 deg during unsat time.

Frame 140 of JulieVD2 has the time stamp in transition between 13:51:47 and 13:51:48.

Star counts all corrected for vignetting, averaged over ~60+ frames, as many as I can before losing signal with panning

star	V	SpType	Radius during data	Vignette Corr multiplier	signal+sky	sky	signal	altitude
SAO 99840	6.7	A	r=0.06	0.98	7571	5998	1573	11.65
SAO 99924	7.7 (7.5 GSC)	F	r=0.15	1.00	6595 7039	5345 5671	1250 1368	15.15
SAO 99876	6.9	K	r=0.37	1.10	8840	5780	3060	13.07
SAO 99953	(7.8 in GSC)	K	r=.20	1.0	6588	5338	1250	15.63
SAO 99928	7.7 in GSC, 7.5 in SAO	F	r=.55	1.28	6583	5545	1038	13.38
SAO 99886	6.9	F	r=0.50	1.22	7377	5590	1787	11.33
SAO 99933	7.7 in GSC	K	r=.05	1.0 (a few sat pixels)	7189	5491	1698	15.18
Denebola + V=8.9 F star	2.2 (2.1 in GSC)	A3	r=.15	1.0 (lots sat)	12760	6056	6704	11.40

Capsule reductions

First appears to me at F=269. Saturation of pixels begins at frame 315. Very difficult to get Tangra to track the capsule. Success starting at frame 301, going to 342 after which tracking and data goes bad. Capsule pixel saturation starts at 315. Excel file of results is cap-Ap3px.csv

star	V	SpType	Radius during data	Vignette Corr multiplier	signal+sky	sky	signal	altitude deg	comments
SAO 99953	7.8 in GSC	K	.15	1.00	2706	1272	1434	15.63	20% bigger than 6px version
SAO 99924	7.7,7.5	F	.14	1.00	2753	1371	1382		same as 6px version

jump to frame 283 in julieVD2, check 'field rotation' and 'stop if lose tracking', .5px tolerance, slow but reliable as tracking, 3px aperture size for aperture photometry.

Capsule Light Curve: Capsule in black, Bus in gray. Frames 324,325 dropouts due to tracking loss on Tangra, and all after Frame=342. Bus pixels strongly saturated for all points, and uses an undersized aperture. Capsule frames affected by a few saturated pixels beginning on Frame 315.