Occultation of 12.4 Star by Pafuri

Nov 29, 2015

Preston skymap and info
Predictions . Cabrillo Dome is at track +21.67km north

FOV in the Orion Q70 eyepiece in 10" Meade

Camcorder LCD field


Ted, Ray, and Rick Baldridge all said they'd not be trying it. Derek and I were the only two I knew of who definitely were going for it. I'd thought about sleeping at the observatory for this 3:43am event, but decided I'd get more work done if I instead stayed home and got up early to drive down. Got up at 2:10am, and was on the road by 2:25am, arrived at my office and printed the charts above. It was cold on campus, but much colder when I got to the observatory. Hoar frost covered everything, and water puddles had a coating of ice. Dragged out the 10" and gear and thank God our lone working hairdryer was still working. I periodically unfroze my hands with it, and kept dew off the 10" corrector plate. I left the ZR45mc camcorder in my car until 20 minutes to go. I used Hamal and Regulus on opposites sides of the sky to 2-star align the 10", and then sent it to the triangle of bright stars shown on the first chart. It was on the extreme right edge of the FOV, so targeted to within half a field. I was OK with that. I left the stars in the Q70 eyepiece and focused them carefully, then removed the 2" diagonal with Q70 and screwed on the Meade f/3.3 and Celestron 1.25" adapter, slipped in the Watec 910hx, oriented with menu wheel on lower left as per my standard. Then slowly rotated the focusing knob CCW until the stars came into focus. The 2x setting was not good enough to see 12th magnitude and I adjusted it to 16x, which was going to be enough for magnitude 12.5. That's 8 field integration or about 1/4 second time resolution. Not great for this 4 second event, but it was pretty dim and the 85% moon was only 34 degrees above it and brightening the sky.

I found the target just in time to start the recording about 2 minutes ahead of the predicted time. I watched carefully and the star did indeed wink out. My hands were cold and hurting and I was having flashbacks of being at Lick Observatory 2 mornings earlier for the Low Key Hill Climb, where it was 32F as well. I packed up quickly, took one scene shot after I'd packed everything, and drove back to the warmth of home.

The LiMovie Analysis... here is the .csv light curve and the raw .avi video is here (45 sec centered on the 4 sec event)

I set the photometry circle to 4 pixels, and the sky torus to 6-25 pixels for the target (faint, in center), and the comparison=tracking bright star above. I linked tracking and verified the target circles did move with the tracking circles and didn't "get lost" during the occultation.

The LiMovie light curve. Not sure why the yellow and pink curves are offset since they are the same star with the same photometry circles, as verified during the play-through/capture of the data. The target is in blue, with a clear occultation. In hindsight, I should have integrated shorter

The target light curve, with the occultation centered more or less

Zoomed in. The 8-frame grouping is clear. Each point is a frame analyzed. The values within an 8-frame group differ only in the photometry noise.


The Occular Analysis

I chose min and max occultation frame lengths by looking at the obvious occultation and going a few frames beneath, and then more than a few frames longer. The star has no color information so I left it blank in the magnitude panel. I checked off 'block average' and N=8 points, and verified on the blocked vs unblocked light curves that the proper offset was 0 frames. The S/N was 3.37 in the solution and occultation confidence was 100%. The duration was 3.9 seconds, and a little closer to the centerline on the north side was Tony George, who had a 4.0 sec event.

The IOTA report is here and it was sent in at 4:30pm Nov 29, 2015.

Occular's light curve, including the 8-frame integration out of the videocam

Fairly tight D and R solutions for the 100 trials

Occular's "final report"


Now the plot twists a bit. Tony George got a positive from the track just south of me by a few km, and gets a very similar light curve which shows the same small hump at the early stages of the occultation, as if the star were a binary. Derek may have a miss, or may have only 1 of the components occulted (the dimmer one), only more work will reveal. The plot thickens!....

Sky plane including Derek's track. If the asteroid's round, Derek should have an event. Perhaps only one component of new binary?

My .csv light curve analyzed by R-OTE software. Gets same timings as Occular. And quality of light curve is no better than mine from LiMovie/Occular.

Tony George's light curve. Both of ours seem to show a ~0.8 second period of a faint companion after the D, and then a slightly reduced brightness for a similar time after the R, as if a faint companion star were arranged roughly along the direction of the asteroid movement.

I forgot about taking a documentary shot, so intent on packing up quickly. I snapped one just before getting into my warm car to get back home


Dec 3; I tried sending the full 400meg version of my .avi file, but after all night, it stuck and failed. Instead, I re-downloaded a trimmed version from the camcorder, to just the 45 seconds centered on the event. I sent the trimmed version to Tony, who reduced it his way, and got the same timings, and depending on assumptions, either there was a hint of a rise at the beginning of the occultation, or not. It's just not convincing as a signature of a binary, so that'll remain off the table. Derek's tape is apparently too noisy to read; he set the videocam to no integration. The rule of thumb is to make sure the target is visible, but just faintly, if you are adjusting integration. With 8-frame integration the star was faint, but I think I could have gotten away with 4x frames and perhaps gotten better time resolution. It would not have been visible on the camcorder LCD screen, the the brighter neighbor star would have reassured me it was still in focus and on-screen during the event. Occular's reported timing accuracy does seem to be too optimistic. Given as 0.033 sec (1-sigma), but R-OTE says 0.04sec for 1-sigma.

Turns out that it's not possible to do linked tracking with no-psf tracking in the version of LiMovie (0.9.97a2) I had. Tony sent me a newer version 0.9.98a2 (the "update software" button in LiMovie instead deletes LiMovie and doesn't download an update!). With this, I produced these tests. Tony and Brad's experience is that non-psf tracking/photometry gives better results than using psf. That's not my experience so far. But here's my tests...All use linked tracking so the target circles follow the tracking star circles. I verified this is true even with non-psf set. I also have see that LiMovie resets parameters to defaults even after you've set them ! How rude! So I have to recheck after all stars are set, just before hitting 'start', to ensure proper check boxes are set still.

aperture tracking and photometry, with small 3/5/25 pixel circles

aperture tracking and photometry, with more standard 5/7/25 circles. I see no difference compared to the smaller circles

aperture tracking and photometry with larger 7/9/25 circles. Curves look noisier, as was predicted by Tony. I'd have thought they'd be less noisy as more star signal would remain inside the inner circle, but apparently not so.

PSF tracking + aperture photometry, with 5/7/25 circles. This looks like lower noise than aperture tracking.

PSF tracking + PSF photometry, with 5/7/25 circles. This has the best looking occultation bottom, but the noise in the comparison star is a little larger than using aperture photometry.

Tony used my trimmed .avi file and his decisions on software parameters to generate these two alternative light curves

The distribution of solutions. Note the error distribution is steeper than Gaussian. 1, 2, 3-sigma vertical bars are indicated. 1-sigma error is 0.040sec


Bottom line; A good first guess is to go with PSF tracking and aperture photometry, for future events.