Asteroid occultations: drift-scan method

[lukebl]

Hi all. I want to attempt some asteroid occultation measurements, using the drift-scan technique. That requires you to be able to start and end the exposures at a precise time. Does anyone know how to do that automatically to ensure accuracy (rather than clicking manually)? I am using an SXVF-H9 cam and capturing using Astro Art 3, which doesn't seem to allow you to set an automatic start time for the image.

I also need to be able to analyse the subsequent star trail, to determine a light curve for the period of the exposure. Most websites require you to have Maxim DL, which I can't afford, so are there cheaper or free alternatives?

[lukebl]

Sorry for bumping this post. Maybe my initial post rambled a bit!

Basically, I just wanted to know if there was a way of setting my Starlight Xpress camera to automatically start capturing an image at a precise pre-determined time, rather than relying on me clicking at the right moment. Any ideas?

[Phil Fargaze]

Would this link be of any help? http://tinyurl.com/nadoema

Also there is the Occultation Observing Manual which is a pdf download, drift scan is on page 184. http://tinyurl.com/ott7vws

[lukebl]

Thanks for that, Phil. Interesting stuff. I've studied those sites before, but it still doesn't answer my main query as to how I can set my camera to automatically start capturing an image at a specific pre-determined time, and without having to use the unaffordable (to me) MaximDL.

[Phil Fargaze]

I do asteroid occultation observing myself but I use the method of taking an AVI with GPS video time inserter. That is recorded onto laptop with a frame grabber and VirtualDub software.

Light curves can be produced with either Tangra or Limovie. All the mentioned software is freeware.

You might want to try the UK based yahoo discussion forum for asteroid and lunar occultations for an answer to your original question which is here http://groups.yahoo.com/neo/groups/UKoccultations/info

[glappkaeft]

Thanks for that, Phil. Interesting stuff. I've studied those sites before, but it still doesn't answer my main query as to how I can set my camera to automatically start capturing an image at a specific pre-determined time, and without having to use the unaffordable (to me) MaximDL.

You can time the exposure manually but you need to know that the FITS time stamp on the image is accurate to a few 0.01 s or at the least 0.1s. This requires that the computer controlling the camera to have a accurate clock (we built a local time server that uses a GPS with a PPS output signal) and to calibrate the shutter latency (time between the command is sent and the shutter actually opens). Unfortunate the only way I know of to calibrate shutter latency is to use MaximDL.

The reduction can be made using IRIS and then analysed using any number of programs.

[lukebl]

Thanks for that info. I had subsequently found that the FITS data attached to the image accurately records the start time of the image, so it should be pretty accurate.

I use atomic.exe (http://www.worldtimeserver.com/atomic-clock/) to synchronise the computer clock. Is this sufficiently accurate for asteroid occultation timing?

[glappkaeft]

Thanks for that info. I had subsequently found that the FITS data attached to the image accurately records the start time of the image, so it should be pretty accurate.

 

How did you determine that? We measured a delay of 0.42 seconds with a camera without a mechanical shutter using a special calibration routine. That latency would have been enough too make it hard to make comparisons between different observers.

 

I use atomic.exe (http://www.worldtimeserver.com/atomic-clock/) to synchronise the computer clock. Is this sufficiently accurate for asteroid occultation timing?

Synching the clock using NTP over the Internet is considered an unreliable method since the time reported can vary quite a bit due to latencies especially if there is lots of traffic or a wireless internet/LAN involved. Nettime is another NTP program that reports the offset between of the old and the new time. With our local time server the error (updates every 60 s) is 0-1 millisecond but over the internet I often see 10-50 ms or even a few tenths of a second sometimes. If you can reliably get a accuracy of a few 0.01 s it should be OKish to use NTP but this need to be tested on site, preferably every time.

Euraster prefers either "GPS one Pulse Per Second", "Phone time signal (wired phone)", "Permanent radio time signal", "Local atomic clock" or "Permanent NTP synchro with local atomic clock". That is because they are both very accurate and reliable.

[lukebl]

Now I'm completely confused.

If I can't trust either the time stamp on the FITS image or the accuracy the computer clock even with synchronisation software, then I don't see how it's possible to accurately time anything with my current setup!

[glappkaeft]

If you get consistent results (low offsets between synchs) you can trust the computer clock. The FITS time stamp is another issue, it might be precise but is it accurate? You need to check this and correct for any difference. This can be done by photographing a computer screen where the current time is continuously displayed and take into account refresh rates, double buffering and so forth. Maxim has a nice automatic routine for this built-in and I programmed my own Java app to do the same as a double check but it is very manual. I have not been able to calibrate my analysis method with a times pulse or similar so I'm not really sure how accurate it is.