PHD2 Bad Pixel Map and Lodestar

[Joel Short]

This is probably more of a Lodestar issue, and I have asked this of Terry at SX.  But since I'm seeing it with PHD2 I figured I'd ask here as well.  With the warmer weather has come a slew of new hot pixels on my Lodestar.  Seriously, so many hot pixels that I'm wondering if it is finally going the way of the dodo (this Lodestar is probably 7yrs old?).  The problem I'm running in to is when I take a bad pixel map and apply it, what I'm left with are about 5 donuts.  The hot pixel clusters are so large that the bad pixel map only accounts for the central portion of the really bright/condensed spots.  

Terry's response was that there are a lot of hot pixels but that it didn't seem terribly unusual and perhaps I need to take better darks for the BPM.  So...I'm not sure what would help to do that. Would increasing the exposure time and/or number of exposures help?  Any other words of advice?

joel

[Andy Galasso]

Hi Joel,

We may be able to give better advice if we can see your bad pix map master dark and your bad pix map files.

Open your most recent PHD2 debug log and search for "Loading defect map file".  That will tell you the folder location and file name for the defect map (something like "PHD2_defect_map_1.txt"). The master dark file will be named PHD2_defect_map_master_1.fit with the number matching the number in the defect map file name.

Andy

[Joel Short]

Here's a link to the files:  https://www.dropbox.com/s/mne9rgi9mkwgt8f/darks_defects.zip?dl=0

FWIW, I also took darks and comparing the defect map to the master dark, the defect map looks much better.  I also note that I took the darks and BPM during the day when it was warmer outside.  Last night with the cooler temps the hot pixel "donuts" were not as apparent. 

joel

[Andy Galasso]

Joel,

I suspect you may be using the SX ASCOM driver to connect to the camera and are using the Gaussian Blur option. Is that correct?  I believe this is having the effect of blurring a single hot pixel into the surrounding pixels.  This make the bad-pixel map more difficult to build and less effective since it is assuming isolated hot pixels; it does not handle clusters very well. You could try disabling the Gaussian blur option in the ASCOM driver setup.

Another experiment to try would be to use the builtin driver, Starlight Xpress SXV.  This operates a little bit differently from the ASCOM driver in that it bins the pixels 2x1 to avoid the interlacing issues that I suppose led you to turn on the Gaussian blur. When you build a bad-pixel map with the builtin driver you will be selecting the actual pixels and not the post-processed pixels (gaussian blurred or deinterlaced).

Andy