“PERIODIC ‘Task’ EVERY ‘X’ FRAMES” a success

[Candieman]

Robin,

Last night was the first time in 17 nights in which the smoke and clouds were thin enough to allow testing (but not real imaging). I seized the opportunity to finally test the new “PERIODIC Autofocus EVERY ‘X’ FRAMES” that I had asked for a few weeks ago. Here is the relevant portion of my sequence script:

PERIODIC Autofocus EVERY 12 FRAMES
SET EXPOSURE TO 2
SET GAIN TO 300
GUIDING STOP
AUTOFOCUS OFFSET 200 TO -200 STEP COUNT 11 BACKLASH 500
GUIDING START
SET EXPOSURE TO 20
SET GAIN TO 111
FRAMETYPE Light
END PERIODIC
FRAMETYPE Light
GUIDING DITHER EVERY 3 FRAMES
CAPTURE 30 FRAMES REQUIREGUIDING True
LIVE MODE
PERIODIC Autofocus STOP

I just did a series of very short exposures (20 seconds each) with dithering every 3 frames and autofocus every 12 frames just to put this new script through its paces without taking a lot of time. This did exactly what it was supposed to do, and now I’m planning on making this my “standard” method for imaging, although I’m not sure if the line “FRAMETYPE Light” needs to be within the PERIODIC task, since I have one immediately after the “END PERIODIC” statement. Which one is “more correct”? Also, any suggestions or corrections that could improve the performance of this portion of my script are welcome.

One issue I noticed while this was running was with the Estimated ETA for this sequence. At the start of this script, it gave me a time that was ~15 minutes away, which made sense, since I was asking for 10 minutes' worth of exposure time, plus 10 stops to dither, and 2 automated refocus steps. However, sometime not too soon after it started, I noticed that it no longer gave an estimated completion time. The message on the screen switched to: “Cannot estimate ETA for this sequence”, and that stayed there until the sequence script reached the end. Even when the script was doing the last exposure, it still said it couldn’t estimate the completion time, which was just seconds away. This is not the first time I’ve seen this happen, so it's not the new “PERIODIC ‘Task’ EVERY ‘X’ FRAMES” command. I scanned through the log file from last night, but I didn’t see anything in it concerning the estimated time to completion. Not having an estimated completion time is hardly a showstopper, but I’m just wondering if you have seen this too, or if others have had it happen to them. Perhaps there is an easy fix to this.

At any rate, the new command works like it should, and I'm happy with that. Thanks for implementing it.

Rich

[admin]

Hi,

good to hear that it worked as planned

The one thing I would suggest with the sequence is to use the 'Save the camera settings and restore them when this block finishes' step inside the periodic. That lets you remove the two steps that return the exposure/gain to their previous values and ensures that the periodic block will do the right thing regardless of the previous gain/exposure. I think you can remove the light frame from the periodic too - that won't be changed by anything in the periodic.

Now to the ETA issue...

Currently the estimation for end time does not take into account periodic tasks of any sort, so the capture 30 frames step will take considerably longer than expected. Once the time taken so far goes beyond the estimate for the whole of the rest of the sequence then you get the 'cannot estimate message'. There is some adjustment for excess time taken, but that only happens at the end of a step (basically spotting that a step took 30 seconds when estimated at 20 so add 10 seconds to the ETA). Probably something I should look at, but quite complex as it would require each step to estimate time left at any point.

cheers,

Robin