Autofocus Error

Hello, this is my first session in Prism and I am having some issues getting autofocus to work. I take a 5 second exposure, select a star, but when I run it I get this error.!

Screenshot 2021-02-10 195107|690x259

I have tried taking new exposures and selecting new stars but it just keeps happening. Also, I am starting from good focus.

welcome to the club. Did you use the automatic focus or direct auto focus ?

I used automatic. I am coming from NINA imaging software. I basically took a good v-curve there and used both sides of the curve to set my staring and end points.

What hardware are you using by chance?

ASI294mm and a Moonlite. I think I figured it out though. I had to increase the half window size. It seems at the default of 35 the window was smaller than the star. I increased to 200 and autofocus ran. Still need to tune the setting a bit but it did a successful run. Now to try and get platesolving to work!

1 Like

excellent news.

Pro tip: Binning helps plate solving

Got platesolving working, now Im trying to get guiding calibrated.

How do I enable dithering when running through the automatic observation?

Hi,

In my experience, this is caused by a hot pixel on the camera within the small frame used by autofocus. The algorithm appears to look for the brightest point as the star’s centroid and gets fooled by such pixels.

My solution has been to ensure the focus star is not near such pixels when I start (I can see the hot pixels in the image capture).

There may be ‘clever’ ways round this like:
(i) Dynamically subtracting darks, or using cosmetic file list of hot pixels
(ii) Tweaking the algorithm so it ignores single pixels (they surely cant be stars!)

…But I dont know how to access them, or if they even exist.
If not, perhaps a suggestion for next release?

Thanks for the reply. I was able to get autofocus to work by increasing the half window size. I left it default as the autofocus video had suggested, however I think that setting may need to change depending on your imaging tried. At 35 (the default setting) the window doe the guide star was extremely small. Not until I increased the half window size to about 100 was the window large enough to fit a guide star and begin guiding on it.

It’s been a journey but in slowly figuring Prism out. The only 2 issues I am having at this point is being unsure if Prism is actually dithering when using the Automatic Observation and the fact that Prism keeps randomly close. No error message or anything, it just disappears and thats a huge problem because I image while I sleep.

I experienced some random ‘hanging’ sometimes but, as I was actively driving it (not asleep!), I could restart. Prism is good at restarting quickly from where you left off…

I have notced that, when things dont shut down normally, Prism sometimes leaves a process running in the background. Maybe worth checking to see if you have any rogue processes?

It was running on a freshly rebooted Windows session so I would think there wouldn’t be any rogue processes. I checked the log folder in the a Prism directory and it was empty.

When say it restarts where toy left off, so you mean Prism resumes the state it was in before it hangs? For me when I reopen the program I have to start from scratch.