Detailed Notes on Timing and Record Occultations
( for the aspiring occultation observer, in the hope it may be helpful)



Last Updated 2014 August


Chords across (130) Elektra recorded by
UK observers on 2010 Feb 20

Techniques used for the study of time dependant astronomical events, particularly Lunar and Asteroid occultations of stars.

Including some observing equipment.


Driftscan of (130) Elektra obtained at
Great S
hefford Observatory

Contents

a) Preamble
b) Orbit Uncertainties

1. Geographical Coordinates
2. Which Time Signal ?
3. How bright is the star?
4. Stopwatch or chronometer
5. Audio Observation and tape analysis
5a MSF time receiver
6. Reaction Time
7. Video
8. Video / VHS Tape
9. Video accuracy
10. Fast Frame Rate
11. Integration time
11aCombined Visual / Video observation

12. Video Time Overlay
13. Digital Video
14. Video Grabber VHS to PC
15. DV tape to PC
16. Video Grabber and calibrated clock
17. Drift Scan and Manual drift scan with DSLR
18. Primary Time signals sources
19. Secondary Time sources.
20. Reporting

21. Software integration of TANGRA 3, OCCULT 4 and OccultWatcher

Preamble
Successfully recording the time (UT) of disappearance (D) and reappearance (R) of a star provides information on the position and size of the asteroid. With more spread out observers the asteroid shape can be deduced. Detection of D and R with a telescope can be made by eye (visual) or with electronics (Video / CCD). Visual observation needs a 20cm or larger aperture (25-30cm) tin order to see and monitor stars of 10th to 11th magnitude clearly.

Electronic detector/recorder combinations will produce a permanent record of the event for re-play and further analysis and smaller instrument can be used (4-8"). Reliable timing is important, so a UT time signal is recorded with the observation. The techniques described here have been used successfully by many observers. Asteroid occultation predictions can be obtained from internet resources (EAON) and globally. DeskTop software by Dave Herald (OCCULT 4) is used to generate predictions.

The current version of OCCULT 4 is 4.1.0.29 (as of 2014 April 21).

OccultWatcher 3.7 is the recommended Windows software for predictions and planning.
TANGRA 3 is used for the analysis of video
AOTA is a new tool from Dave Herald for automating video analysis and reporting.

<Top>

Orbit Uncertainties
There are small errors in the orbit of an asteroid and the position of the star. This results in a likely shift in the track away from the observer, who then sees a shorter occultation or "miss". Or indeed the shift may work to an observer's advantage if he/she is not in the track but the shift is towards the observer. See this graphic by David Dunham. A miss (or no occultation) is a negative observation and should be reported.

Uncertainty and errors arise from:

  • Star positions (10 to 30 mas) NOTE: The Gaia Mission will reduce star position errors to about 1 mas for 1 billion objects.
  • Orbital positions (10 to 100 mas)

Compared to a main belt asteroid diameter of 30-50mas, the prediction uncertainty can be 2 or 3x the asteroid diameter. The errors are displayed on the predictions.

Example: Occult4 prediction for (386) a 208km diameter asteroid (solid lines). Dotted lines enclose the most probable area for an occultation.

<Top>

1. Observer's Geographical Coordinates
GPS is the preferred longitude and latitude datum for an observer, referred to as WGS84. This has superceded the Ordnance Survey datum OSGB36. The two coordinate systems refer to the same spot on Earth, but the co-ordinate origins are different. Read the Wiki. If you have OS coordinates, the OSGB36 to WGS84 conversion can be done online at http://www.nearby.org.uk/coord.cgi. Alternatively you can read them off Google Earth using the cursor to an accuracy of about 5 meters. I have checked my carefully surveyed observatory's OSGB36 coordinates and the mathematical conversion to WGS84 matched the Google Earth location exactly (to +/- 0.1"). After May 2000 the deliberate inaccuracy introduced into GPS caused by "Selective Availability" was switched off, so that GPS now provides sub 10m accuracy.

Height above mean sea level is also required when reporting events. I give heights to the nearest 10 m above MSL. Look up http://www.streetmap.co.uk 1:50000 for height information, or use GPS.

Map (left): The contours and spot heights are still usable

<Top>

2. Which time signal?
Timing accuracy for occultations is crucial. There are some time signals that should be avoided such as from TV channels and un-calibrated computer clocks. Some phone networks (e.g. Virgin - thank you to Andrew Bate for this information) can produce a time signal, but accuracy might be unreliable. Primary time sources are listed in section-18. These are generated from atomic clocks and transmitted to telephone land-line (in the UK), Radio signal, and GPS equipment. In practice one of these time signal should be used to synchronise another clock which could be electronic or electromechanical with a constant rate, and this is used at the telescope. A recent development is the use of GPS as the time source and this is rapidly becoming the preferred method for video occultations.


GPS video time inserters are now preferred for video. The time text overlay is inserted each video frame to 1ms. See section 12

Visual observers using stopwatches or synchronising clocks to UT,should consider using use a PC application written and maintained by astronomer Hristo Pavlov "Beeper Sync 3.3" http://www.hristopavlov.net/BeeperSync/. This is a well characterised method and provides reliable 1 sec ticks using a wired internet connection. (Wireless might introduce additional time delays - this should be assessed by the user)

<Top>

3. How bright is the star?
Stars occulted by an asteroids are quite often faint (typically 10-12th magnitude). The prediction histogram for star magnitude indicates 90% of occulted stars are 10th-12th magnitude (in 2014). Only a relatively small number are observable by eye.

Occasionally some are brighter; and there are two events (D and R) to be recorded in succession. The interval (or chord length) can range from 1 to 30 seconds. If the asteroid and star are of similar brightness the magnitude drop (dM) is small. Under these circumstances (dM < 0.7) an observer might not see the event, or his/her reaction time may be longer (up to 1 second). Short occultations (less than 1 second duration) are subject to larger timing errors and uncertainty because the duration is close to the observers reaction time. ( Reaction time can be avoided by using video Section-7, or Drift Scan Section-17). The magnitude drop is part of the prediction and is an important consideration that will influence the method of observation.

<Top>

4. Stopwatch and Chronometer
Most visual observers will be familiar with the “stop watch and telephone pips” method of timing. Start the watch (or multi-function chronometer) at the instant of occultation and stop the watch at a known second provided by a telephone or radio time signal. (Note: Use BT land-line in the UK). Lunar Occultations are often timed by this method.

The observation time is then calculated:
Time = UT (pips) – dT – PE

[ dT = watch time, PE = Personal Equation or reaction delay e.g. 0.3 to 0.5 sec)

A lap mode setting can record an interval while the watch still runs. PE can be estimated. See Section 6

<Top>

5. Audio observations and tape analysis
One solution to recording events in quick succession is to use an audio tape recorder with a UT clock tick or radio time signal superimposed. Remember to announce a minute marker at the start and end of the recording. Analysis involves counting all the seconds. This takes time, and the process needs to be double checked, but the record is permanent and can be re-evaluated. Those who have Lunar grazing occultation experience will already be familiar with this method of recording. Some observers have digitised the audio which makes analysis easier.

The iphone and other smart phones can record audio, and the files emailed and analysed with free software such as Audacity 2.0

The author has used a camcorder video ( Sony TRV 22E) as an audio recorder while using the internal clock as a time overlay (synchronised to UT).

<Top>

MSF Receiver
An MSF 60 KHz receiver on a 4x1 inch Veroboard was build in 1974/5 based on articles in Wireless World. This was the result.
The aerial was tuned
with a micro volt meter, but can be made from a standard long-wave coil - see here
Output is pips at 1 sec intervals.


Diagram (above) drawn by John Toone and reproduced from The Salford Astronomer (1977)

<Top>

6. Reaction time (reaction delay or Personal Equation)
An observer's Personal Equation (PE) is subtracted from the recorded observation time. PE is the largest uncertainty in timing by visual means and depends on physiological factors. In the end PE may only be an estimate for a difficult conditions e.g. 0.5 +/- 0.2 sec. For clear events (bright star, good conditions) reaction time can be in the range 0.2 to 0.3 sec depending on the observer. There are methods of estimating reaction time. The dropping ruler method, software simulation or on-line apps. The author has used a stopwatch to time an asteroid and lunar occultation from a TV monitor, and compared it to the same event on Video. The reaction time is consistent for bright stars by direct vision. The estimated times of fainter events required greater PE correction (0.5-1 second). Estimated errors should be realistically reported.

<Top>

7. Video.
An integrating video camera has become the detector of choice for asteroid occultations. Watec120N, 120N+ and Mintron 12V6HC [and now the WAT-910HX], are used by amateurs with excellent results. The camera has other applications such as meteor detection, or as a finder, or as an auto guider. Video produces a permanent unbiased record without reaction time uncertainties. The exposure can also be adjusted to suit conditions (magnitude drop, instrument aperture, seeing), so it is a versatile piece of equipment. There is a readout time ( half the frame or integration time) to be subtracted from video times. See Section-9

Cameras in use (compiled by Dave Herald) - some are now unavailable new.

KPC-350BH
Mallincam MCH Plus color
Mintron 12V1C-EX
Mintron 12V6HC-EX
PC164C
PC164C-EX2
PC165DNR
SCB-2000N
SK-1004XC/S0
WAT-120N
WAT-120N+
WAT-902H
WAT-902H2 Ultimate
WAT-910BD
WAT-910HX

The Watec 902 H camera is used extensively (and costs less) but does not have an integration function. (It runs at 50 or 60 fps). In the US the Supercurcuits camera is available with similar sensitivity. The Current WATEC integrating camera is the 910HX/RC
(as of Aug 2014)


8. VHS video tape
VHS is a simple option to record from the video camera. It will also record a time overlay by inserting a time and date generator such as this unit from Voltec into the video signal. The result is a permanent record of the event. VHS video can be digitised for analysis on a PC, or played back frame by frame. This is laborous. A Video Time Text overlay is highly recommnded (Section 12)

<Top>

9. Video Accuracy and detector "dead" time
Video introduces improved accuracy, with 0.02s time resolution (One field = 1/50th second). Compare this to a visual observer who at best may achieve +/- 0.1s accuracy. Although you will find integrating video observations reported with errors larger than 0.2 sec, this is due to the integration time, and isn't a reaction time. Video doesn't suffer from "detector dead time". This is the time when a CCD camera is not exposing, but is saving the previous image. The time taken to download a 16 bit CCD frame is 0.6 to 1.2 sec (Depends on the computer specification, binning and subframe settings). The download time can be estimated, but a short occultation might not be detected if it happens while the CCD is waiting to record the next image. With video there is no delay other that a time correction which is a constant for a particular camera. Details can be found on G Dangl web site.

Gerhard Dangl's video timing analysis http://www.dangl.at/ausruest/vid_tim/vid_tim1.htm
Gerhard Dangl's M67 IMAGE comparison http://www.dangl.at/ausruest/cam_comp/cam_comp.htm

16bit FITS images can be recorded with minimal dead time if 25%sub frames and 2x2 binning is tried. Consider using this arrangement when seeking the best S/N on a faint objects, and when the event could be of long duration.

<Top>

10.Frame rates
Low light monochrome video has the ability to detect short events (0.04s), like non instantaneous events or other unusual phenomena like double stars. Some CCD camera used in sequential imaging can reduce the dead-time by setting a smaller region of interest.( ROI ). Also the ATIC Titan is reputed to operate at 10fps or faster. A USB2 web cam (e.g. Image Source) can also operate at a range of different exposures (and frames rates). This is untested, and accurate time stamping is a difficulty. Some recent digital web cams can operate at 60 fps (or more). This has not yet been fully exploited for occultations.


USB web cams are a distinct possibility, although the writer has not seen any reported observations. Low sensitivity may be a limitation. The IMAGESOURCE DFK21 (colour with IR filter) does not have the required sensitivity. The author has tested one. A 10th mag star needs > 2 sec exposure in a 12"

Reports are comming through of certain QHY planetary cameras used for occultation work at 5 fps (200ms exposure).

<Top>

11. Integration time using an integrating video camera.
The aim of video observation should be to use the shortest integration time the telescope and weather conditions will allow. This will require some experimentation beforehand. For example a 20cm instrument in good conditions can record 12th magnitude stars in 0.16 sec, and a 30cm to 13th magnitude. So a 20cm can record most of the predicted events from EAON and IOTA using an integrating video camera, and fainter objects by extending the integration time. The author finds that a short integration time of 2 frames ( 2/25 sec) helps to smooth out seeing ripple in bad seeing conditions. 10cm and 15cm instruments can be used with longer video integration times. A maximum of 0.64 second exposure should be considered as giving the best compromise between detecting the event (s/n), and time accuracy in smaller apertures. (5cm - 15cm)

Note: MINTRON cameras should be used at frame integrations that are multiples: e.g. 2,4,8,16. etc and not intermediate rates if these are provided.

ALSO the MINTRON and WAT-910HX count in fields and not frames. 2 fields = 1 frame !

<Top>

11b Combined Video / Visual observing

Displayed on a monitor, video will produce a bright clear images that could be timed with a stopwatch, but it is preferably to record the event either on tape (VHS, Digital Camcorder / Video 8 ) at the same time. When operating at 25fps, the PE has to be taken into account. If using an integrating camera, the combined PE and camera delay are summed.

12. Video Time Overlays
There are two choices, either obtain a time and date generator (TDG) with manually sync to UT, or a GPS time inserter.
A) GPS
The GPS system produces a time stamp on every video frame to a precision of 0.001 seconds. It is the convenience of analysis of a permanent record that makes this method so attractive. Here are some commercial GPS units:

GPSBOXSPRITE2 by Blackbox. Here is a screen shot of the Sprite in use, and You tube demo of the Sprite.

AME-TIM10 by Alexander Meier Electronik. Here is a You tube demo of the TIM10.

Kiwi-OSD system is no longer available commercially, but is used by many.

IOTA-VTI is a new addition to video time insertion via GPS, and is designed to give best results for occultations. It can be purchased from Video Timers and it will be maintained for the future.

<Top>

B) TDG
A
Time and Date Generators (TDG) is reliable but it doesn't’t produce an exact time on each video frame. The TDG clock is manually synchronised to UT ( e.g. from telephone land line, radio or NTP) and to obtain the time of each frame, the user advances the video one frame at a time until the frame in which the disappearance is reached. The fraction of a second can then found. Single frame advance is available on most better video players. This is made easier if the video is digitised and analysed by software e.g. Limovie and TANGRA. Both these programs are free down loads. TANGRA 3.1 is the easier to use and some automated analysis can now performed

<Top>


13. Digital Video Recording.
The author uses a DV camcorder with analogue to digital conversion (AV/in) built in. There are several makes available second hand (Sony and Canon). The video output is plugged into the camcorder AV/in for display and recording. Newer equipment might also do this (unknown). DV cameras operate on their own batteries and are highly portable - Beware that tape heads need to be cleaned. A dirty tape head could spoil your recording. Remember to clean it !

Digital Video (DV) cameras capable of A to D conversion (known or reported to the author by David Dunham and others) are:

Sony: TRV22E, TRV33E, TRV480E
Canon MV600i (an others in this series), MVX25i
Canon ZR camcorder models 60, 65, 80, and 90 (and other "i" models not beyond 300i)

Canon MX2, Canon XL2, Sony VX2000, Sony VX2100, Sony PD 150 and Sony PD170 are camcorders with better quality than consumer level equipment. I'm a permanent user of VX 2100 and PD170 - writes Pawel Max Maksym (pl)

A comprehensive list is here http://www.4kam.com/camcorders_with_av-in_av_input.htm sent to me by Jan Manek (update 2011Sep13)

<Top>

14. Video Grabber VHS to digital
Recording directly to VHS tape and viewing on a monitor or TV is straightforward to set up. This may be sufficient for your needs if the VHS player has single frame advance for analysis. Digitization of the tape can be beneficial for post analysis with Limovie. You might already have a PC with a suitable video card, but at the end of the day, you are going to analyse or archive the video data in a PC readable format (e.g. AVI, DivX). The simplest method is to invest in a USB Video Grabber (Analogue-to-Digital). They are supplied with software for video editing which you might not need, but it would be advisable to install as there may be a feature you need later.

There are software tools available such as the freely downloadable VirtualDub. This will detect the Video Camera (or VHS tape AV-out) and digitise the signal. I select a file name and duration for the recording, set the frame rate to 25.000 fps and record. This could be done during the occultation event provided your PC will record happily – test your system before the event and make any adjustments. Check you have auto file name increment enabled, so that previous files are not deleted.

One point to note is that the bottom of the frame might not be digitised by the video grabber (The Author’s did not). This results in the TIME overlay at the bottom of the screen being lost. The answer is to put the time stamp at the top of the frame or change to the digital resolution in VD.

<Top>

15. DV tape to PC
Currently the author transfers DV tape direct from his Camcorder by firewire to the PC which was bought with firewire installed. Connection is made to the PC and NeroVision Express3 SE is launched when the connection is made. This software sees the whole video frame. Windows Movie maker or VD may do the same.

16. Video Grabber and Calibrated PC clock
Some observers use a calibrated PC clock to time stamp the AVI obtained from VirtulDub. The Author uses free software Dimension 4 which applies a correction to the PC clock by connection to the internet. The displayed PC clock (Windows Date and Time Properties) is not more than 0.2 sec different to UT immediately after synchronisation, and over 24hrs the drift is +/- 0.5 s. Some observers prefer to use GPS than rely on a PC clock. If frames are dropped on a PC, the timing becomed less accurate.

<Top>

17. Drift Scan and Manual Drift Scan with DSLR
The drift scan method is described here by John Broughton. The occulted star is allowed to drift across the detector controlled by software. A stellar occultation appears as a break in the stars trail.

Manual Drift Scan with DSLR
Focus a DSLR and position the star near the East side of the frame so that it will drift across the longest side. Start the audio tape recorder with time signal. Then about 1 min before the event, simultaneously switch off the telescope drive and open the camera shutter. The sound of a “switch throw” or “shutter opening” should be audible and recordable. This marks the start of the exposure. Now close the shutter after a predetermined time (sound is recorded again). The length of exposure should made less than the transit time across the field, with the predicted time of occultation occurring near the centre of drift.

Example:
A 0.5 degree field of view has a drift transit time in minutes of
0.5 x 4 / Cosine (star’s declination) Equation (1)

Field calculation
To calculate the field of view of your camera + telescope combination use this example based on a typical Canon DSLR with CCD size of 13.8 x 20.7 mm:

Field of view in degrees where FL is the focal length of your telescope in mm is:
20.7 x 57.3 / FL Equation (2)

This is for the longest side of the camera frame. So for the Author’s 30cm F/4 with focal length of 1200mm, Field = 0.99 degrees

The start and end of the exposure (taped events) were recorded. The time and duration of an occultation can then be worked out from the trailed image. Its helpful to know the rate of sidereal drift in pixels per second because this will indicate the accuracy of the observation. The Canon chip is 3456 x 2304 pixels

Pixels per second for a Canon 350D and 1200mm FL telescope is:
3456 / [ Equation (2) x 4 x 60 / COS(declination of the star) ] Equation (3)

= 14.4 pixels per sec (Star declination = zero)
and 12.5 pixels per sec at Declination +30

Experiment with the ISO settings so that the frame is not unduly overexposed as this will effect the measurement of the star trails.

<Top>

18. Primary Time Sources
1) Telephone (BT 123)
2) Audio radio reception (Many stations world wide, three are selected)
a. MSF (UK) 60 KHz long wave
b. DCF77 (Germany) 77.5 KHz long wave
c. WWV (Fort Collins US) 5000, 10000, 15000 KHz Short wave (Try 5000 in the UK)

3) Global Positional System (GPS)
a. GPS receiver that contains a one pulse per second clock (1pps) and associated electronics to generate and display precise UT.

4) An observatory clock
a. You have access to a professional system perhaps.
b. A clock that displays UT based on radio reception (commercial or home made).

<Top>

19. Secondary Time Sources (Clocks synchronised to UT with a measured time difference)
1) Atomic wrist watch – time automatically synchronised by a radio time signal. There might be a small internal delay in the LCD display, but an analogue display should be fine.

2) PC internal clock set by an internet time service.
a. Clock controlled and updated by a well established application. (e.g Thinking Man Software Dimension 4)
b. Beeper Sync application that produces pips via internet

3) Other clocks with established errors ( i.e. “running fast / slow by:” ) manually synchronised to a Primary time signal (see above):
a. Video Time and Date Generator (TDG)
b. Quartz crystal controlled analogue clocks (electromechanical)
c. Electronic metronome (pips).
d. Camcorder with an internal time display HHMM ss.
e. Stopwatch or multi-function chronometer started or stopped on a UT second.
f. Digital wrist watch

<Top>

20. Reporting.

a) Asteroids
(UK / Europe)
First report ? Then send to the BAA ARPS occultation coordinator for checking ! Thanks Tim Haymes

Fill in the EAON report form and send it electronically. Examples at http://www.euraster.net/results/index.html
This form can be pre-filled from OccultWatcher to avoid mistakes.

Register with PLANOCCULT http://vps.vvs.be/mailman/listinfo/planoccult

b) Using Occult 4 software and sending to the European Collector

21) SOFTWARE
Release Note from Hristo Pavlov.

OccultWatcher 3.7 is the recommended Windows software for predictions, planning and reporting
TANGRA 3 is used for the analysis of video
AOTA is an automated video analysis and reporting tool, included as part of TANGRA, OW and OCCULT 4

Author: Tim Haymes
Assistant Director (Occultations)
Asteroids and Remote Planets Section
British Astronomical Association.