Occultation of TYC 1856-00745-1 by (87) Sylvia and by its satellite Romulus
Analysis of scans
E. Frappa, A. Klotz, P. Dubreuil, V. van Vlijmen, A. Carbognani

Cord analysis at Euraster.net

1. TAROT observations

TAROT used to observe star occultations by asteroids. More than 50 positive occultations have been recorded for few years.

The analysis is based on the fit of the light curve. The fit input parameters are (magnitude magV, time of disapearance D, time of reapearance E) for Sylvia and Romulus. The time sampling of the fit is 10 times that of the data to be well oversampled. We first fit the PSF spread parameter according the fall and the rise of the Sylvia occultation. Then, the fit of the light curve take account the spread due to the PSF. This method allows to determine (D,E) at a precision (at 2 sigma level) below the time sampling.
Time sampling = 0.247 sec/pixel
TYC 1856-00745-1 magV = 10.7
  S1 = 01:36:53.0
  E1 = 01:38:21.5
  S2 = 01:38:38.0
  E2 = 01:40:06.5
(87) Sylvia magV = 12.6
  D = 01:37:26.6  (+/-0.1 sec)
  R = 01:37:39.5 (+/-0.1 sec)
  duration = 12.9 sec (+/-0.2 sec)
Romulus magV > 14.0
  D = 01:39:00.73  (+/-0.08 sec)
  R = 01:39:01.17 (+/-0.08 sec) (see chi2 fit correction below)
  duration = 0.44 sec (+/-0.11 sec) (see chi2 fit correction below)

Air: Temperature = 12.6 (deg. Celcius),
Humidity = 22 (percent)
Wind: Speed = 0 (m/s)

Zoom on the occultation light curve showing the 2 sigma uncertainty:

All the above determinations of R and D were fits made by eyes. The Romulus occultation profile is affected by noise. So, we tried to compare the eye fit  (o) with the chi2 minimization fit (*):

The corresponding values for the Chi2 minimization are:

  D = 01:39:00.73  (fit sampling 0.01 sec)
  R = 01:39:01.21  (fit sampling 0.01 sec)
  duration = 0.48 sec  (fit sampling 0.01 sec)

Only the R value is found to be late compared to the eyefit. As a consequence, the duration becomes 0.04s longer.

2. Pierre Dubreuil observations

Pierre Dubreuil (Aspremont, France) observed with a Celestron 8 F/D=3.3 + Audine Kaf401. The telescope mount is an EQ6 driven by ASCOM. Pierre choosen to record four scans (#1 to #4 below). Geographic coordinates of the observation are lon=7°14'31 E lat=+43°46'58" alt=480m.

Scan #1 => from 2013-01-06T01:33:13 to 2013-01-06T01:37:13.000
Scan #2 => from 2013-01-06T01:38:17 to 2013-01-06T01:42:17.000
Scan #3 => from 2013-01-06T01:43:23 to 2013-01-06T01:47:23.000
Scan #4 => from 2013-01-06T01:52:48 to 2013-01-06T01:56:48.000

Unfortunately, the main body occultation occured between the end of the first scan and the beginning of the second scan.

Fortunately, the Romulus occultation occured during the second scan. The occultation is well seen in a faint trail just above a bright one:

It was not easy to extract the the star profile and to calibrate the time. It has been possible by the help of a non trailed image acquired juste before the first scan. Anyway, the absolute calibration of the time is not better than +/- 0.5s due to the round of the second of the DATE-OBS keyword in the FITS header. The determination of the (D,R) time was performed using a simulation as for the TAROT data (see previous section). The PSF has been extrated using the bright star of the scan.
Time sampling = 0.388 sec/pixel
TYC 1856-00745-1 magV = 10.7
  S = 01:38:43.6
  E = 01:39:10.0
Romulus magV > 13.5
  D = 01:38:57.60  (+/-0.25 sec)
  R = 01:38:59.30  (+/-0.25 sec)
  duration = 1.70 sec  (+/-0.35 sec)
The precision do not include the +/-0.5 second due to the round of the DATE-OBS keyword which affects only D and R, not the duration. Moreover, additional offset for D and R should occur due to the method of the computer clock updating (not mentioned by Pierre). We estimate the total time offset on D and R to be +/- 2s.

3. Victor van Vlijmen observations

Victor van Vlijmen (Plascassier, Grasse, France) observed with a Celestron 11 F/D=10 + Sbig ST 4000. The telescope mount is an EQ6 driven by ASCOM. Pierre choosen to record one scan from 2013-01-06T01:37:45 (few seconds after Sylvia main body occultation) to 2013-01-06T01:39:45 (includes the Romolus occultation according to TAROT). Geographic coordinates of the observation are lon=6°58'00 E lat=+43°38'42" alt=250m which is about 7.5 km south of the TAROT Calern cord projected in the sky plane.

The signal to noise ratio (SNR) is very low because the time sampling of the trail is only 0.068 sec/pixel. The light curve was binned by a factor 3 to increase the SNR. The PSF was taken into account to fit a model on the data.

The conclusion is that there is no occultation longer than 0.3 second centered on 01:38:58 to 01:39:03. The following graph shows the zoom centered on the date of the occultation seen at Calern offseted by -0.66s to take account for the geographic coordinates difference with Calern. The red model is an occultation 0.3s convolved by the PSF of the image.

4. Albino Carbognani observations

Albino Carbognani (Italy) observed with a Ritchey-Chretien 810mm F/D=7.8 + FLI Pro Line back illuminated (OAVdA). Albino choosen to record one scan (motor stopped during 87 seconds). Geographic coordinates of the observation are lon=7°28'42" E lat=+45°47'22" alt=1670m.

Scan #1 => from 2013-01-06T01:37:08.7 to 2013-01-06T01:38:35.7

Only the Sylvia occultation is detected (satellite are not supposed to pass over the site). The signal to noise ration is very low due to the presentce of a lot of clouds in the sky. Even with a 810mm diameter telescope the magnitude 10.7 star is very faint. The time axis was resambled by a binning 3 to increase the signal to noise ratio. The analysis was made with the same method as previous sections:
Time sampling = 0.216 sec/pixel (after a digital binning 3 along the time axis)
TYC 1856-00745-1 magV = 10.7
  S = 01:37:10.2
  E = 01:38:34.0
Sylvia magV = 12.6
  D = 01:37:33.30  (+/-0.28 sec)
  R = 01:37:51.00  (+/-0.28 sec)
  duration = 17.70 sec  (+/-0.40 sec)

5. Model

From the two cords, we can project the times of Romulus as distances in the plane of the sky. From two cords, only a spherical model can be constrained by the observations. We are obliged to add 0.90s second to the times (D,R) of P. Dubreuil to be consistent with a sphere. This time offset is consistent with the uncertainties due to the round of the DATE-OBS keyword due to the errors in the beginning and end of the trail.

We find a best fit using a diameter 23.0 km and the centrality 11.1 km northern to the Calern path.
Taking account for the 2 sigma level errors, the diameter lies between 19.0 and 28.3 km and the centrality is between 9.3 and 13.7 km northern to the Calern path.
Note this is only a preliminary result. Two other cords will be added later to constrain the shape of Romulus.
Note that the end of the Romulus shape was at 500m south to the TAROT Calern position !

6. Predictions

from http://www.euraster.net/pred/20130106-Sylvia/index.html

Sylvia orbital elements from S. Preston (updated 2012 Dec 22, 00:01 UT)
Sylvia diameter from ISAM (254 km at the moment of occultation)
Star position from UCAC4
Satellite position from Miriade (IMCCE)

Google map prediction path for Romulus