Sorry I not sure where to post a feature request. So I post another one here.

The original one: www.indilib.org/forum/development/3531-a...-greater-than-0.html

Last night, I tested the meridian flip in Ekos and found a problem.

I set up a series exposures with duration 900s and the auto meridian flip was enable, then I chose a target which was near the meridian. When the hour angle was greater than 0 (this is the default value) , the scope would not flip automatically until the current exposure was finished.

In my opinion, the current exposure should be abort and dropped when hour angle is greater than 0, then perform the meridian flip immediately and restart the exposure which has been abort before. The reason is that if an exposure is very long (For example 3600s or more, my CCD can do such a long time exposure), the scope cannot track accurately when hour angle is greater than 0. In addition, the scope is in dangerous position and it might hit the tripod.

Based on my view, I urge that to stop current exposure and perform meridian flip in Ekos when auto flip is enabled and telescope hour angle is greater than 0. Then restart the exposure which has stopped before.

EDIT: There seems to have a better way to deal with this problem: comparing the time to meridian(hour angle is 0) with next exposure duration. If the exposure duration is greater than the time to meridian, wait and do not start new exposure until auto meridian flip is completed.

Does anyone come across the same problem with me ?

Your proposed solution might work. I think very few folks if any would have this problem unless they do very looong exposure and they're already at limit when hour angle is zero. In my case, I can safely track 2 hours past the meridian, and I usually do not capture longer than 10 minutes so it was never an issue for me.

Perhaps it can be "yet another option". Since imaging at the meridian is quite valuable data for many and not a single second should go by underutilized.

I agree that the imaging data on meridian is quite important. I do photometric work and know that the data from meridian usually has the best quality due to a lower airmass.

However, I will still be grateful if you can add an option to flip immediately and pause the exposure if the duration is longer than the time to meridian. Since I am working with a remote telescope, I must keep the telescope in a safe position as much as possible.

In addition, shooting a target with a exposure duration longer than 600s is quite common for me as I use Strömgren u filter to measure the Balmer jump, where the light flux is quite low in near UV (due to the Balmer jump the stellar flux is significantly lower and the atmosphere absorb more light in NUV). It is necessary to get a longer exposure to achieve the same SNR compared with those broad band system.

My opinion on this is that it is up to the remote installation to make sure the mount does not hit the tripod. There is a good pull request for EQMOD on github which should be generalized.

Then, yes, it is up to the client to abort the exposure when a meridian flip must be executed. But when? The capture tab provides the HA limit, so this could be used as a first limit (considering second is the driver hard limit).

But that means the started exposure is lost. In order to retain that imaging time, an additional implementation is required to be able to pre-flip the mount. If the telescope driver provides that feature, then the capture tab could use it when starting capture.

To extend even more, the scheduler needs to properly adjust the capture plan to avoid the situation where the flip occurs in the middle of an exposure. Not easy because of the variability of the job durations, but achievable. Specifically, it requires the scheduler to stop a job, do something else, then return to that paused job. Technically, it is already able to abort and restart jobs, but not fit another job in between as the jobs are currently scheduled one after the other (per creation or per altitude).

To summarize:
1. Implement INDI driver under-the-counterweight mount security measure when remote RA limit is exceeded (east or west pier side).
2. Force exposure abort in the capture tab when HA limit is exceeded (meridian flip will then automatically ensue).
3. Implement INDI driver pre-flip capability.
4. Implement pre-flip request in the capture tab, with opposite RA value.
5. Implement rescheduling of job when an exposure is bound to be aborted by a meridian flip (if we don't have 3/4, this one is slightly more complex).

I miss cases on guider management and job repeated batches here, but probably this will just work?

-Eric

Maybe it doesn't require such dramatic changes. The capture job can go to "waiting" until the HA limit is reached and then it flips. As far as the scheduler is concerned, the capture never stopped.

In order to improve the robustness of the scheduler, I plan to add operation timeouts. The waiting capture got me in the tripod last time unless there would be a better inter-tab communication scheme?

-Eric

Do you plan to implement the waiting feature in the next version of Ekos ?

Hi knro:

How about using the exposure time, read out time and delay time settings to estimate the time to meridian, though this method still has its limitation as download time cannot be predicted.