This is related to satellite (optical) tracking but someone here may have a solution. Please allow this post.
I have taken photos, with a DLSR, of satellites for several years and have been a member of Seesat for a while. I can get great photos and get .fits files and build observations. But they are not consistent and not reliable. I think that the problem is that I take ten second (normally) exposures but can’t get the start of the streak accurately enough. The Nikon D300s records time but I can only get the time to the tenths of a second (I think) and I am not sure that the time synchs from the computer to that accuracy.
I am looking to build a small dongle that plugs into the “hot shoe” and uses the flash signal (when the front shutter clears the focal plane) to record the time from a GPS chip. That should allow me to record time to the millisecond.
But I analyze orbits and am not an engineer and don’t build circuits.
Anyone interested in helping?
Charles
astronomer, C++, orbit analyst, photographer
If I understand you correctly, you want to send a timestamp when a input is triggered ?
If we assume a simple microprocessor and a gps chip using nmea, we get timestamps sent every second, and the message can take tens to like a hundred ms to be transmitted. This can also be synchronized with the pps signal that is in the microsecond domain. Between these the microprocessor needs to keep track itself.
When the input is triggered, the current timestamp can be sent to the host via usb-serial.
Hi Charles, I think your solution could work, but maybe it is easier to change the camera to an USB astro camera like the ZWO ASI174 MM or better. Maybe you can reuse your Nikon manual focus lenses, through a lens adapter. And use the STVID software of Cees Bassa. It only runs on Linux, so that would require a Linux PC or Laptop. I use a Raspberry Pi 5 mini computer for this. I control it remotely through VNC.
Of course, this doesn’t answer your original question and proposes a completely different way of working. The computer synchronizes the time to a NTP time server. Measurements have shown it will be within 0.2 seconds accurate.
Once it works it will provide you the IOD lines fully automatically, which makes it possible to do a lot more (every night) observations with less work.
Link to stvid software:
Also within Satnogs there is work in progress for an STVID, but I haven’t used it yet:
Thank you!! I am working on that as a separate effort but I do think that if I could get a DSLR to work it would open up tracking to anyone (probably two people!!) who would like to use a used DSLR camera. They are available very inexpensively.
I have exchanged email with Greg Roberts, Marco Langbroek, etc about their systems.
Almost certainly this and even more flexible systems have been developed but I have looked for a couple of years and have not found anything.
I live in Houston, Texas not far from a huge petrochemical industry and the space industry and I must assume that lots of people have needed to take photos with accurate time stamps but I have looked and looked and can’t find anyone.
For my Canon DSLR I’ve used gphoto2 to operate the shutter over USB, and use the computer clock to log the start of the exposure. As there will inevitably be a delay between sending the trigger and the shutter opening, I used audio recordings to determine this delay. For this I recorded the sound of the shutter opening, and the computer playing an audible beep when it triggered the camera to open. Using this approach I was able to determine a delay of about 0.16 seconds between the trigger and the shutter being completely open.
Using the same method, I found then when setting the exposure time to 10 seconds, the actual time between the shutter opening and closing was more like 10.06 seconds, which was corroborated by the length of the satellite trails in the images.
In the end I only used the DSLR for objects in high orbits, where the timing is less important.
I strongly agree with Eelke that for LEO orbits using a video or CMOS camera will give vastly superior results over a DSLR, simply because at 10 or 25 FPS you obtain so much more positional and timing information that a DSLR can not be competitive.
Cees - Yes, working with David Brierly I can tell that a ten second exposure is NOT exactly ten seconds. The big improvement needed is to get an accurate time of the start (or stop) of the exposure. When I started this I thought that there must be several solutions - but I have not been able to find any. I am hoping to get a very accurate time for the start of the streak and then work on finding an accurate duration.
The video camera is a far better solution but maybe if we had a DSLR option, some people (who could easily afford and set up a DLSR) could get an simpler path to contributing.
I mainly do analysis and so this takes me away from that, but it has been very interesting.