This is a very bright event, but shallow drop of 0.22mag. I'd recommend it at 1x to maximize the number of points, help on saturation. Will want to dial down Gain from 41 normal maximum to something smaller. Duration is 19s. Got to confess there's little scientific value to this one. We have photo'd this one up close and know it well. Any moons probably would have been seen by now. But it's good practice for shallow drop events.
Alt=29, Az=122 in Aquila
|
|
I recorded it for 3 min before and 3 min after the event. I did it at 1x, from driveway, and lowered the computer brightness and also the gain, gain=36. I cut the brightnesss on the IOTA VC 2.4 menu and also the contrast. The star looked bright, but I figured saturation would only affect occasional pixels.
There was evidently more variable aerosol or perhaps poor tracking than I expected. The noise was pretty much corrected for by the reference star, but I'm tempted to re-analyze it with a larger mask and combine fields (not analyze in field mode). I have not done that for the timings below. I set the D and R intervals by eye, to avoid false positives due to remaining noise. I've noticed that even when setting the D and R intervals, PyOTE can still make up its own mind on where it finds the best event. In this case, however, with multiple trials, it always settled on the correct range and the predicted moment was right in the center of the detected 21s occultation.
NIE Test: 14.1 sigma
magDrop report: percentDrop: 13.1 magDrop: 0.152 +/- 0.017 (0.95 ci)
DNR: 0.50
D time: [04:43:51.3535]
D: 0.6800 containment intervals: {+/- 0.3261} seconds
D: 0.9500 containment intervals: {+/- 1.2336} seconds
D: 0.9973 containment intervals: {+/- 3.1239} seconds
R time: [04:44:12.3934]
R: 0.6800 containment intervals: {+/- 0.3261} seconds
R: 0.9500 containment intervals: {+/- 1.2336} seconds
R: 0.9973 containment intervals: {+/- 3.1239} seconds
Duration (R - D): 21.0399 seconds
Duration: 0.6800 containment intervals: {+/- 0.5717} seconds
Duration: 0.9500 containment intervals: {+/- 1.7006} seconds
Duration: 0.9973 containment intervals: {+/- 3.6894} seconds
 |
 |
 |
 |
 |
 |
 |
Analysis #2: Bigger 4px mask and use 2x, combining the fields into their frames.
I then re-did the reduction. In PyMovie I let it group fields into frames - no "analyze in field mode" - and I raised the mask size to 4px instead of 2.4px, since the sky was rather dark and the target quite bright, there was little danger of sky noise being a worry, and the advantage is that I could be more certain all the target photons were counted. The result was clearly better. After minimizing the metric interval (chosen to be the entire light curve) in both smoothing length and time offset, the occultation stood out more clearly. The D and R timings had much smaller error bands, most important.
magDrop report: percentDrop: 15.1 magDrop: 0.178 +/- 0.017 (0.95 ci)
DNR: 0.86
D time: [04:43:51.3735]
D: 0.6800 containment intervals: {+/- 0.1852} seconds
D: 0.9500 containment intervals: {+/- 0.7146} seconds
D: 0.9973 containment intervals: {+/- 1.7718} seconds
R time: [04:44:12.3734]
R: 0.6800 containment intervals: {+/- 0.1852} seconds
R: 0.9500 containment intervals: {+/- 0.7146} seconds
R: 0.9973 containment intervals: {+/- 1.7718} seconds
Duration (R - D): 20.9999 seconds
Duration: 0.6800 containment intervals: {+/- 0.3387} seconds
Duration: 0.9500 containment intervals: {+/- 1.0076} seconds
Duration: 0.9973 containment intervals: {+/- 2.2053} seconds
 |
Without the reference star, identifying the occultation would have been impossible. |
Better confinement throughout the light curve, makes the shallow but long integration fairly straightforward to see by eye. |
This second analysis provided significantly tighter error baands on the timings. Still the same D and R and duration as the first analysis. |
NIE test was better too; 18.9 sigma instead of 14.1 sigma with the 1x and smaller mask analysis |
Observed from home and got a recording at 1x. May have seen the fade on-screen.
 |
 |
 |
 |
 |
 |
I consider this a first-cut reduction for Karl. His duration is much longer than the 21s that Kirk and I saw, and his raw data looks much better than his PyOTE calibrated data. Why? Well, the ref star he used was much fainter and did not have saturated pixels, while Juno I note did have some saturated pixels. The long slow drop to the D and even longer slower escape from the R make me highly suspicious the D and R timings were widened by PyOTE because of the false gradual fade and brighten. This fade and especially post-R brightening are not seen in the PyMovie data, which is the first red flag that should be raised. It's a big claim to have that long brighten and fade, given that even at my home elevation there was not clouds or fog or hint of such. I think these reductions should be done on the raw PyMovie data after importing to PyOTE and not do a calibration or maybe chose a better ref star? I see natural drop spots after the PyOTE D and rise spot before the PyOTE R which would give timings more realistic. I also note that in my and Kirk's data, the predicted center of the event was perfectly in the center of the detected occultation - as it should for such a well observed low number (3)! asteroid. This is not so in Karl's data above. I hope to see a re-reduction and will re-post it when done.
This event brought back memories of my first asteroid occultation, back on Dec 11, 1979 as a new PhD student at UCLA. I organized a group of grad students to drive out to near Castaic Junction, at Oak Flats NW of Los Angeles, and recorded the occultation with visual WWV/tape recorder old technology. Meanwhile, back at the rooftop observatory of the Astronomy building at UCLA, Dan McKenna led the photoelectric detection of the occultation. This was a well observed event, one of the best for asteroids at that time, and published in the AAS Journal PDF
