This is the brightest asteroid to cross Santa Cruz in many years. Watch out for saturations. Adjust gain and recording level to get a nice black sky and bright star. The duration is also a long 2.4s. That means this is an excellent opportunity to find even tiny moons. A single integration drop to zero even at 1x is a detection, if winds and clouds behave as they're expected to - i.e. not happen. However, one problem will be moonlight. The gibbous moon is only 3 degrees above the target.
The bright sky will mean you'll want to probably adjust the gain down. A dark sky is what you want to see. 7.0 magnitude is still so bright compared to even a moonlit sky, that it should be possible to adjust so that the contrast is very high.
This is a slow moving, small asteroid. The 2.4s duration is due to being near the stationary point in its Earth-centered motion, not large diameter. The shadow is moving slowly, so to get the accurate predicted moment of occultation, you need to accurately position your icon on the location you observed at.
Alt=64, Az=134 in Aries, 3 degrees below the gibbous moon.
|
|
I observed and recorded from the same site at Czechoslovakia 43 minutes earlier, at Cabrillo College parking lot not far from the weather station. The sky was wicked bright from the gibbous moon only 3 degrees from the target. I used 1x, and low brightness and low contrast setting. I did not have to worry about "sawtoothing" because my star was so bright that it spread across several rows and columns of the detector.
I did not adjust gain; I left it at Watec 910hx max=41 as usual. The star was a bright-ish gray ball but it did not look to have saturated pixels. And indeed on reduction it did not show any saturated pixels during the PyMovie run. All pixels looked light gray on the on-site recording Lenovo screen, not white. Sky was adjusted to look dark gray. The two nearest neighbor stars to the right were still visible, though rather fainter. The brighter of those two was good enough to use as a ref star and also as a tracking star.I asjusted the mask size to 3.2px which enclosed all the target star lit pixels but little else. The aperture box at 21 px. I let PyMovie use re-centering.
This (target) small gray ball was remarkably constant in brightness during the long recording, except at the event moment it faded to less than half brightness for maybe 1/2 second was my guess, or a little more.
Initial Reductions Immediately Below Are Incorrect.
If you, the reader, want to save time and get straight to the final solution and avoid the initial attempts, then go on down to "Putting it All Together"
Reductions:
I reduced it twice initially, both times before realizing this was a double star. The first time I used the fainter star (the only other star on the video) as tracking and ref. Then I realized that since the target never totally disappeared, I'd probably do better if I made the target a tracking star too. That light curve looks better. There's a slight downward tilt to the raw light trend, but that is removed when using the ref star, faint as it is. Clearly, this star did NOT disappear during the 0.8s of the occultation. I don't know if it was due to being a double. The OWc page shows the target star at 0.20 mas, and the asteroid at 5.17 mas, so that's a 25x ratio in angular diameters and doesn't suggest a partial occultation is all that likely. But the light curve doesn't suggest steps like you'd expect for a double star. Instead, more like a disk-on-disk look to it. The reduction below is square wave, for simplicity. I will have to look at Kirk's light curve to have more to say...
The event was 6s early. That's quite a long ways. Lucky the error was mostly long-track and not cross-track. We could have very easily had a miss, with this error. Predicted UT time on the OWc page was 6s later than observed.
BELOW IS ASSUMING A SINGLE STAR AND SQUARE WAVE LIGHT CURVE. THIS FIRST ANALYSIS IS NOT THE ANALYSIS TO BE REFERRED TO IN THE FUTURE
magDrop report: percentDrop: 61.0 magDrop: 1.023 +/- 0.147 (0.95 ci)
DNR: 4.33
D time: [02:46:26.1916]
D: 0.6800 containment intervals: {+/- 0.0058} seconds
D: 0.9500 containment intervals: {+/- 0.0151} seconds
D: 0.9973 containment intervals: {+/- 0.0530} seconds
R time: [02:46:27.0035]
R: 0.6800 containment intervals: {+/- 0.0058} seconds
R: 0.9500 containment intervals: {+/- 0.0151} seconds
R: 0.9973 containment intervals: {+/- 0.0530} seconds
Duration (R - D): 0.8119 seconds
Duration: 0.6800 containment intervals: {+/- 0.0089} seconds
Duration: 0.9500 containment intervals: {+/- 0.0218} seconds
Duration: 0.9973 containment intervals: {+/- 0.0645} seconds
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
====================================================================================
Dave Gault kindly took a look at this, once I received and uploaded Kirk's light curves below. Kirk's light curves clearly indicate a double star is involved. His R shows a prounouced step. The D may show a very short step also. However, the integration of this interpretation into a coherent story with my light curve is still under consideration. My interpretation, in trying to merge with Kirk's observations, is that I had a mixture of a double star and partial occultation of the remaining star combined perhaps with some diffraction, to keep ~12% of the total light still visible at maximum occultation.
Kirk's data shows a much better determination of the brightness of star A and star B. Since we have identical Watec 910hx cameras and telescopes, I used his levels to guide my own analysis. Kirk's full un-occulted level on his light curve is 95, and when only star B is showing, for about 1s after the total occultation of both stars, his level is 28 on average.
Star A = 71% of total light
Star B = 29% of total light
Assume Star A disappeared first, leading to a sloping level for seeing star B alone, of 0.29 x 260 = 75.
NIE Test: 14.5 sigma
magDrop report: percentDrop: 60.1 magDrop: 0.998 +/- 0.176 (0.95 ci)
DNR: 4.47
D time: [02:46:26.1753]
D: 0.6800 containment intervals: {+/- 0.0054} seconds
D: 0.9500 containment intervals: {+/- 0.0138} seconds
D: 0.9973 containment intervals: {+/- 0.0420} seconds
 |
 |
Looking at the total recording, my light curve shows both point to point noise, and a slower (a second or two) noise waves, even with the horizontal and median PyMovie filtering. On this basis, I decided the slight drop (about 13%) and leveling before the first D was noise, as was another drop about a second after the final R. If true, the sequence was then ABAB, not ABBA as was true for Kirk's light curve. Stars appear to be close to the path limit for me. |
Here is my assumed sequence of events as best I believe can be justified, given that Kirk's data very clearly shows Star B = 29% of the total light, and also that the "star" is listed in OWc as only 1/25th the diameter of the asteroid, corresponding to 0.09s duration of occultation. The sloping aspect of star A's level is smaller if you consider the possibility that this 1x recording still somehow has "sawtoothing". |
Pause: Here's the arguments that Star B did not disappear early and reappear late as DaveG is suggesting:
1. The total duration of the Star B occultation is then 3.4s. That is much longer than the maximum 2.2s for this asteroid prediction. HOWEVER, I note that Kirk's Star A and B events are almost this long. Bad diameter prediction for the asteroid? I'm open to changing my mind here, but I would like to see if a consistent Star A & B position for both my and Kirk's data is possible.
2. The drip in brightness DG attributes to Star B disappearing is too small; ony ~13%, to be Star B according to Kirk Benders very good data showing it was in fact 29% of the total light, not 17%
3. There are ~second(s) duration low amplitude waves in my light curve of similar brightness drops of order 10% or so, during my calibrated light cure, and it's more reasonable to then attribute the early dip and the second dip after the full R (which DG blurs together into a single level and thus accepts that that level difference is just noise), to just noise. I therefore take the unocculted level as the level outside of the much more obvious occultation, rather than an average over the minutes of data outside before and after the occultation.
4. By making that assumption, I then get brightness levels consistent with Kirk's brightness levels; a 70% drop as Star A disappears, then the beginning of Star B's drop to zero but which is interrupted by Star A's rise out of its occultation. This then shelves fairly obviously on close up, to a level that is 30% below the unocculted level, as it should if this is Star B reappearing
... Onward - to continue my analysis to get the timings, then...
Disappearance of Star B is judged to be 1/2 way down the light curve from the level determined from the occultation of A, which is consistent with Kirk Bender's Star A brightness of 71% the total combined brightness. I was unable to get a long enough number of data points for the time from the D of star B and the beginning of emergence of star A to get a timing determination from PyOTE. PyOTE kept trimming out the entire light curve with each attempt to limit the interval to just a very few points. So instead, I am determining the time graphically by finding the moment when the image fades to 0.5 of the faded (missing Star A) level, which is at level=0.5 x 85 = 42. This is at
Star B's D = 2:46:26.6785 UT
from graph below.
PyOTE determined the occulted level after Star A disappeared as shown, but Bender's 29% of full value, applied to the consistent level before the D, gives a lower value = 71. I took the average of these two levels as the basis for then deciding when Star B was half occulted, and that is the uncorrected time for Star B's disappearance. Then the drop to value=30 at minimum, and then Star A began its reappearance... |
The shelf of the light curve on the way back to full brightness is at 2/3 of the unocculted brightness, which argues that what happened is that the brighter star (Star A) was the first to reappear. So I am processing this as a ABAB double star event. If Dave Gault judges this is impossible to reconcile with Kirk's light curve, I'll re-consider. I note that PyOTE's determined drop (rise, actually) was 70.3%, which agrees with Kirk Bender's light level for star A.
NIE test:
magDrop report: percentDrop: 70.3 magDrop: 1.320 +/- 0.325 (0.95 ci)
DNR: 2.54
R time: [02:46:26.8235]
R: 0.6800 containment intervals: {+/- 0.0058} seconds
R: 0.9500 containment intervals: {+/- 0.0164} seconds
R: 0.9973 containment intervals: {+/- 0.0370} seconds
 |
 |
NIE Test:
magDrop report: percentDrop: 40.1 magDrop: 0.557 +/- 0.119 (0.95 ci)
DNR: 2.94
R time: [02:46:27.0035]
R: 0.6800 containment intervals: {+/- 0.0065} seconds
R: 0.9500 containment intervals: {+/- 0.0186} seconds
R: 0.9973 containment intervals: {+/- 0.0692} seconds
 |
 |
Observed a positive from Moran Lake parking lot.
I recorded at 1x from Moran lake in the same spot after Czechoslovakia. I only have the target visible on my recording. At the time I tried setting the gain so another star was visible, but the target still was bright white on the monitor and I was concerned it would be saturated, so I turned the gain down until I the target wasn't white. So I could only use the target for tracking. In pymovie I watched the mask during the event and it got the D but then it lost tracking and didn't capture the beginning of the R, although it did recover tracking the target a little later. During the event the level went to background, the target was not visible during the whole event like yours was. I ran pyote on it anyway and it got a 3.0878 event (earlier than predicted), the D is somewhat gradual. The R is sharper but as I said it lost tracking momentarily for the R, so those timings shouldn't be used. There's some bad timestamps but they are far apart from the event area.
The target did not move between the D and R, so I did not need tracking at that interval, but pymovie will not analyze without a tracking aperture. I tried turning off re-centering but it still lost tracking on the R. However, pymovie supports two-point tracking, for recordings where the target drifts through a line but the target disappears. You go to a frame, add a tracking aperture on the target, set the aperture as the beginning point, then advance to another frame, center the aperture on the target and set it as the end point. It will then linearly interpolate tracking between the two points. So I put a tracking aperture and set it to one point just before and one point after the event, and did an analysis from 2 seconds before to 2 seconds after those two points. I analyzed and watched the mask and it stayed on the target for the entire analysis and it captured the R. The resulting D is somewhat gradual like before, but the R has a distinct step, above background, then returns to the baseline level target curve. There's even a little dip in the R step, probably noise, but maybe a graze? Also, there are high one point spikes right before the D and right after the R. This is a 1x interlaced recording so there is some sawtoothing as it alternates between odd and even lines, so those spikes could be noise but they are coincidentally right at the D and R.
RN: But, could there be sawtoothing for such a bright star which spills over several rows and columns in its image? For me, I was not able to get the target star to look like a bright sharp dot, no matter how hard I tried to focus it. The star was very bright and large. My solution to avoid saturated pixels was to adjust the brightness/contrast in IOTA VC2.4, but since Kirk uses a miniDV camcorder, not direct to digital, this option is not available and he resorted to dropping the gain down to accomplish the saturation issue.
I ran pyote for a square wave solution and it got an event containing both steps for a 2.8739s duration. See what you think, graphs attached, files with a 1 are the analysis of the full 5 minute recording with regular tracking, files with a 2 are with two-point tracking analysis. Double star? Partial eclipse? Diffraction spikes? Graze? This has it all- maybe?
I see your (RN's) event seems to have a little dip after the main event, not sure if this is also in your other stars, but it would fit within the duration of my event, and you weren't that far from my chord relative to the size of the asteroid.
(This is Kirk's analysis but not yet using the fact it is clearly a double star)
magDrop report: percentDrop: 91.2 magDrop: 2.638 +/- 0.343 (0.95 ci)
DNR: 2.55
D time: [02:46:22.5925]
D: 0.6800 containment intervals: {+/- 0.0059} seconds
D: 0.9500 containment intervals: {+/- 0.0193} seconds
D: 0.9973 containment intervals: {+/- 0.1056} seconds
R time: [02:46:25.4664]
R: 0.6800 containment intervals: {+/- 0.0059} seconds
R: 0.9500 containment intervals: {+/- 0.0193} seconds
R: 0.9973 containment intervals: {+/- 0.1056} seconds
Duration (R - D): 2.8739 seconds
Duration: 0.6800 containment intervals: {+/- 0.0091} seconds
Duration: 0.9500 containment intervals: {+/- 0.0295} seconds
Duration: 0.9973 containment intervals: {+/- 0.1117} seconds
 |
 |
 |
 |
 |
 |
 |
My initial Model guess above has....
RN: Star A occultation lasted 0.648s. Star B occultation lasted 0.325s
KB: Star A occultation lasted 3.0 s . Star B occultation lasted 2.0s
Is this consistent with KB's timings?
NO! It requires an unphysical long skinny/ bulbous appendage on one end of Troja. I experimented and could not find a shape and occ durations for our 2 stars x 2 observers that was physically plausible. However, Dave Gault's original proposal I also played with, and does make a reasonable solution. The asteroid isn't a perfect ellipse, but a general asteroid shape and tracks such as I've drawn are a reasonable approximation to what we observed... I was bothered by the long axis 2.2x longer than the nominal circular diameter, but if the asteroid is significantly ellipsoidal (it's small, a size where more oblong irregular asteroids are more common), then the equivalent circular diameter and albedo are not that unreasonable.
Jan 10. My guesstimated solution playing around manually but using my D and R as 3.4s for Star B and 0.65s for Star A's graze...
|
My manual trial/error guesstimated solution, again. Flip top over bottom (I made no attempt to relate to the true sky) and it looks quite similar to Dave Gault and Dave Herald's solution. |
Dave Herald optimized solution fit to a perfect ellipsoid and got this, rather similar to mine. He'll include the double star parameters in a new paper on double stars discovered by occultation. |
I get an occultation for Star B of 3.51 seconds, and for Star A of 0.81 seconds and is a graze...
D: magDrop report: percentDrop: 13.0 magDrop: 0.151 +/- 0.030 (0.95 ci)
NIE test: 4.2 sigma
DNR: 0.97
D time: [02:46:25.0235]
D: 0.6800 containment intervals: {+/- 0.0775} seconds
D: 0.9500 containment intervals: {+/- 0.2620} seconds
D: 0.9973 containment intervals: {+/- 0.6184} seconds
R: magDrop report: percentDrop: 13.4 magDrop: 0.156 +/- 0.038 (0.95 ci)
NIE test: Sigma=4.1
DNR: 0.93
R time: [02:46:28.5034]
R: 0.6800 containment intervals: {+/- 0.1052} seconds
R: 0.9500 containment intervals: {+/- 0.3971} seconds
R: 0.9973 containment intervals: {+/- 0.9771} seconds
Duration (R - D): 3.4799 seconds
 |
 |
 |
 |
magDrop report: percentDrop: 55.7 magDrop: 0.884 +/- 0.149 (0.95 ci)
NIE test: 16.1 sigma
DNR: 4.32
D time: [02:46:26.1934]
D: 0.6800 containment intervals: {+/- 0.0071} seconds
D: 0.9500 containment intervals: {+/- 0.0198} seconds
D: 0.9973 containment intervals: {+/- 0.0682} seconds
R time: [02:46:27.0035]
R: 0.6800 containment intervals: {+/- 0.0071} seconds
R: 0.9500 containment intervals: {+/- 0.0198} seconds
R: 0.9973 containment intervals: {+/- 0.0682} seconds
Duration (R - D): 0.8101 seconds
Duration: 0.6800 containment intervals: {+/- 0.0107} seconds
Duration: 0.9500 containment intervals: {+/- 0.0294} seconds
Duration: 0.9973 containment intervals: {+/- 0.0758} seconds
 |
 |
The shelf on the rise out of the deep occultation, which I initially attributed to star A being fully out and Star B still behind the asteroid, now is seen as due to a mountainous irregularity on the edge of Troja, as part of a grazing occultation. This isn't far-fetched at all. However, the final appearance of Star B is still a bit puzzling, since it looks that the full brightness does return and then stabilize beore dipping again for a half second. These half-second dips are not unusual for the unocculted light curve in my recording, but then the final rise to full brightness is only 13%. I get 13% of the light for Star B, which is only half the brightness of Star B seen in Kirk's recording. I don't have an explanation for that. It can't be a grazing effect because Star B is now seen as traversing near the long axis center of the asteroid, not a graze at all.
My IOTA VTI "went crazy" before the event, and got no data. It was likely a miss from that far north of the path. I may have caught the northern edge.
RN: In fact, my and Kirk's timings show there was a south shift, so Karl would have had a miss.
