Geminiclose
Gemini Password Reset

Please provide your username.

A password reset confirmation email will be sent to your email address.

Username
  • Projects

 
ACP-1421 - Miscellaneous AutoFlat.vbs Improvements
Estimated: 12 Hours 0 Minutes   Logged: 5 Hours 0 Minutes   Time Left: 7 Hours 0 Minutes
Print  |  Send  |  Refresh  |  Url
Type: Enhancement Enhancement
Priority: High High
Severity: Major Major
Status: Closed Closed
Resolution: Complete
Resources: Bob Denny
Reported By: Bob Denny
Component: AutoFlat Script
Fixed In Version: 8.3
Versions Affected: 8.0
Start Date: 11/3/2018
Percent Complete:
Created:  1/6/2016 4:27:36 PM
Revised:
 9/4/2019 10:14:38 AM
Closed:  9/4/2019 10:14:11 AM
Resolved:  9/4/2019 10:14:32 AM

Via email from Eric Dose:

On the forum, I noticed that you might be working on AutoFlat.vbs for ACP's next
release. It turns out I've modified an AutoFlat.vbs version for my own use and
have tested the small but effective changes (screen flats only) for 2 months
with no problems at all. It sports the following enhancements for screen flats
(some of which will carry over to sky flats as well):

Already done: 1. PerFilterBrightness now allows spaces in the list of brightnesses. 
See in the config file how it helps clarify things. I think you may already be 
working on this change, too.

Already done: 2. There's new white-space trimming on the Alnitak-handling code as
well. The Alnitak controller fails on encountering spaces in certain places, and
in a few unusual cases I've noticed spaces do creep in. My tiny mod prevents this,
ungracefully but effectively.

3. An aggressive and much faster-converging algorithm to get proper exposure
times. It usually trims 1-3 test cycles off *each* filter's startup, or
sometimes much more--the current script can be problematic when the tests start
with the images saturated, where very small exposure steps can use up 8-12
cycles to get to the specified ADU level. This is of course a killer for sky
flats and dense filters (merely an inconvenience for screen flats).

4. I take the initial, small-frame test flats halfway from ccd's center to its
corner, rather than at center. This is most important when vignetting is
substantial (big ccd chips)--for these cases, moving out a bit trims a couple of
convergence cycles off each filter. My image-placement math may not be 100%
optimal but it has been working for my 3K x 2K ccd.

5. In a more major and personal extension, the user can set a pre-flat wait
period with a new, entirely optional #PREFLATDELAY directive. I've done this to
be 100% sure I'm thermally transferring any ghosting from bright stars in the
light last images to the ccd's read cells. After the delay, it takes 3 quick
darks without saving to read the chip, which does the actual ghost clearing.
This is automating what I did manually in my asteroid-hunting period. And
obviously this mod is meant for screen flats, as a delay is not so realistic for
twilight flats. If the new directive is absent from the flat plan, there's no
delay, though in my present version the quick darks persist (unless the code
line nPreFlatDarks = 3 is set to = 0).


So, I've tried to mark all my mods with comments containing my initials EVD so
you can Find them (but I'm guessing you'll diff against ACP 8's script). I've
also attached an example config file and flat plan compatible with the modified
script.

Please feel free to use any or all (or none) of these mods, without reservation.



Attachments

Download AutoFlat vbs.txt 1/6/2016 114.97 KB Delete
Download AutoFlatConfig.txt 1/6/2016 5.65 KB Delete
Download flats_VRI_9.txt 1/6/2016 178 Bytes Delete