Tried an imaging session last night with latest 3.6.8 Stable. All functions seemed to work fine - except for flat generation. The last three jobs in my sequence are flats on the three narrow-band filters. First one ran OK, but now every single attempt to generate a flat at all hangs KStars right after the "Turn light box light on..." The progress indicator in the lower right of the Capture Module says "Preparing..." I get the "force quit" dialog from Ubuntu after a short time. Curiously, the log warns about "Always Reset Sequence" enabled, but it is not enabled (screen shot). I have a FlipFlat and have the optical train set correctly (screen shot). If I remove the FlipFlat from the optical train config and close the cover and turn on the light manually, the sequence runs fine. Auto flat calibration and generation used to work great - was adding the cap and light to the optical train new for 3.6.8? Seems to be a significant issue there.

Hi, have sometimes the same situation my camera refusing to capture..capture status just displaying "preparing..." . No obvious root cause in logs ...or
Restarting power and Kstars is not always enough to retrieve a fully operational capture module.
Effective on 3.6.9 and previous.

Try to uncheck the "take flat at the same focuser position" under the filters menu...

Hi, i don't use filters, so that's probably why i don't see this option in the ekos "capture" module ?

Best regards.

In my experience it was checking the option depicted in the figure that triggered an error similar to yours... if you do not have filters... well, maybe at code level the problem is the same but is triggered somewhere else

Thanks for the screenshot. I confirm i don't have access to the menu "filter". This error is unpredictable, didn't understood yet what was trigerring it. Yesterday it was working, now it's not...In this example i try just to take Dark..i don't care about focusing...

it seems that opening then closing the telescope cap (snap cap driver) activate the capture...even if it was already close...
Looks like an initialisation bug or something this kind...