David's Astronomy Pages CCD Image Notes - Session 88 (2005-06-06)

Notes (S87)

Notes (Main)

Home Page

Notes (S89)

Goto Images from 2005-06-06

Parfocal Issue - Blue Filter

The set of RGB filters that I use with my CFW8a filter wheel are meant to be parfocal (ie all come to focus at the same position), however it is has been noticed that star images through Blue filter are distinctly less sharp than through other filters, with focus position set using the Clear Filter.   This observation was tested by analysing FWHM values of stars in M92 taken through different filters.  The results showed that whilst FWHM peaked at around 3 arc secs through Red, Green and Clear filters, it peaked at around 5 arc secs through Blue filter. This seemed to confirm that star images were indeed less sharp through Blue Filter.   This conclusion was confirmed visually by looking at two available repeat images in each filter.

M92 (Hercules)
CCD Colour Image (LRGB) Luminance 3x20s (average), Blue 3x30s (average) with  x1.15 weighting Green 3x20s (average), Red 3x20s (average) 2005-06-06  23:38 h UT (#88023-34) (altitude of M92 at time of image 74 deg)
20s, Red Filter (#88029), 20s, Green Filter (#88032),   30s, Blue Filter (#88026), 20s, Clear Filter (#88024)
Red Filter Image (20s exposure), #88029		Blue Filter Image (30s exposure), #88026
FWHM Values measured on earlier images of M27  shown similar problem

In a separate test,  1 sec exposures of a mag +5.5 star were taken through each filter. Analysis of these images suggested that images through Blue were equally sharp.

  Red Filter      FWHM=  3.7, Mag=  8.65, S/N= 102.5
  Green Filter    (image failed)
  Blue Filter     FWHM=  3.7, Mag=  9.09, S/N=  80.2
  Clear Filter    FWHM=  4.1, Mag=  6.47, S/N= 286.6
 

Back to Top

Image Linking 

Significant issues with Image Linking were evident during Sessions 86 and 87  (2005-05-13 and 2005-05-17). This was initially put down to having a relatively narrow field of view (7x12 arc mins when operating scope at f/10.3) and a relatively sparse star field in the regions being imaged in Virgo.  Later analysis showed that there appeared to be a transient problem that was seemingly associated a new hotfix version of CCDSoft, where by the 'Show Inventory' command  will not find the usual number of objects and thus cause Image Link process to fail due to insufficient number of stars in the image.   

Eventually it was discovered that resaving the Source Extraction settings in CCDSoft,  allowed Show Inventory and thus Image Linking to again work properly.      

Whilst this was improved image link success,  there was nevertheless still an underlying problem, whereby if I  I tried to image link with the just the GSC catalog, linking would sometimes fail as I might not have enough GSC stars in view, whilst if I image link with the UCAC2 or USNO B1.0 subset catalog on I found that image linking would fail if the exposure wasn’t long enough to image faint stars ( brighter GSC stars are ignored when linking using UCAC and USNO stars).   To solve this problem I modified my CCD Imaging scripts so that they would first try imaging linking with the GSC catalog and then if this didn’t work it would try image linking with UCAC2/USNO catalogs. The script does this It does this by calling TheSky to turn the relevant stars catalogs on & off as appropriate, using following code

' Turn on  USNO B and UCAC catalogs 
'  ----------------------------------------------------

Chart.DocumentProperty (skDocProp_PlotUSNOBSubset) = 1 
Chart.DocumentProperty (skDocProp_PlotUCAC) = 1 
Chart.Refresh

' Turn off  USNO B and UCAC catalogs 
'  ----------------------------------------------------
Chart.DocumentProperty (skDocProp_PlotUSNOBSubset) = 0 
Chart.DocumentProperty (skDocProp_PlotUCAC) = 0 
Chart.Refresh

Initial trials show that I could get up to 90% success using Dual Attempt Image LInking. Subsequently this number got up to 95% after optimising CCDSoft’s source extraction parameters. I

Polar Alignment

Despite the image link problems 20 objects were nevertheless successfully acquired and imaged over a 2 hour period in session 2005-05-17 using automated scripts. The Pointing Errors (ie the distance between the planned target object and the image centre ) were relatively small, lying in the range 0 to 3.7 arc minutes, indicating that the Polar Alignment of the Telescope is holding up pretty well since ts last re-alignment made back in Dec 2004.  

Back to Top