The animation shows a full period of the pulsar PSR B0531+21 with its two maxima resolved in time. Due to the low brightness and the recording technique the image quality is unfortunately not very high. This animation was taken on 17.01.2021 by Martin Fiedler at the observatory Radebeul (Germany) with a 24" f/3.3 GoTo Dobson and a QHY174M-GPS camera in combination with an OVNI-M night vision device. Summarized, here the single frames of a 15 second video sequence with 5ms exposure time each were successfully merged according to the pulsar frequency. If needed I can also provide an animation without the time insertion.
http://galerie2.astroclub-radebeul.de/i ... Medium.gif
Here is some additional information about the creation of the video:
According to my research on the pulsar, it has two maxima in the 33.5ms of its period, one bright and one fainter, both lasting about 5ms. This results in the minimum necessary exposure time (5ms). In addition, the pulsar is very close to another star and must be spatially resolved accordingly. So I extended the focal length, which also helps to recognize the pulsar, because the high background brightness due to the nebula is reduced.
On a 5ms frame you can't recognize stars with 15 or 16 mag even with a cooled CMOS camera + night vision device, for that the noise too high. The only option, as is often the case, is to stack enough images to get the faint pulsar out of the noise. That's when I had the idea to use our QHY-174M-GPS camera, which is actually there to measure star occultations by asteroids and can set an accurate GPS time stamp. With this, one could assign to each frame its exact location in the period of the pulsar, and then stack them precisely. For this purpose I equipped the telescope with a Barlow lens, the OVNI-M night vision device without lens/eyepiece and the QHY174M-GPS camera with a 50mm f/1.4 lens. The NVD was running at maximum gain and also the camera was at the limit with the gain but cooled down to -40°C.
Then I had to get the timestamp from the single images, rounded appropriately, into the file name. For this I searched for a long time for a useful OCR software. It would be easier if the SharpCap would do it right away (somehow I couldn't manage that). Then I built an table and took the timestamp of the first image as the beginning of the period. Then you need the exact period of the pulsar (0,033781965), which changes every year, I didn't know that before. Now I always add the full frequency to the first timestamp for the complete recording period. This resulted in the end in a column with nearly 500 values (times) in my case. Then you have to do the same for the remaining 33ms of the period (each new column +1ms). Then I built a program to find the frames matching the period times and copy them into the corresponding folders. In the end I had 34 folders with about 80 frames in each. I stacked them all manually in Fitswork, the flashing of the NVD makes an automatic detection of even the brightest star impossible. Since the star was also not clearly visible everywhere, I could usually only really use 50-60 images.
The chance that this really works out I had rather classified as low, it was completely unclear whether the time stamp of the camera is accurate enough and whether the OVNI-M can even resolve 5ms in time. But it really worked and in the animation you can see the main and the secondary maxima cleanly separated. The image quality is grotty but it worked. With a bigger telescope, better seeing and less noise I'm sure it would work without stacking, but I don't have it. Nevertheless I am very satisfied and happy with the results.