Loading ...
Sorry, an error occurred while loading the content.

RE: Re: Is this RBI and how does it affect your imaging strategy

Expand Messages
  • bob_franke
    That would be my STF–8300 to -15° I have to watch my Dragon Dictation closer. -bob --- In ccd-newastro@yahoogroups.com,
    Message 1 of 16 , Sep 16, 2013
    • 0 Attachment
      That would be "my STF–8300 to -15° "

      I have to watch my Dragon Dictation closer. <g>

      -bob

       



      --- In ccd-newastro@yahoogroups.com, <ccd-newastro@yahoogroups.com> wrote:

      Hello Lee,

      Something seems to be corrupted with this thread, so I restarted it.

      Yes, you seem to have a good case of the RBI's. This is why I only cool my STF – 83002 -15°. The colder you go, the stronger the RBI. I also never take flats or darks after an imaging session.

      I have seen comments suggesting that RBI flashing will increase noise. Maybe somebody more knowledgeable will chime in.

      Cheers,
      Bob
      http://bf-astro.com
    • stan_ccd
      Message 2 of 16 , Sep 16, 2013
      • 0 Attachment

        --- <bfranke@...> wrote:

        > I have seen comments suggesting that RBI flashing will increase noise.

         

        Flashing does not remove RBI; it distributes RBI over the entire frame.  Like everything else, the RBI "signal" has Poisson noise, so it does add a bit of noise.

         

        RBI bleeds out slowly so it can be treated as a modest elevation of dark noise that decreases over time.

         

        It is easy enough to measure using a virgin (unexposed) camera (i.e. keep it dark while it gets to operating temp).  For a given exp time:

         

        1) take 2 virgin darks

        2) flash 

        3) take a flashed dark

         

        process:

         

        1) subtract one virgin dark from the other virgin dark, measure STD.

        2) subtract one virgin dark from flashed dark, measure STD.

         

        noise from flashing = sqrt((STD_2*gain/2)^2 - (STD_1*gain/2)^2)

         

        Flashing is a common practice for professionals so it can't be that bad.

         

        Stan

      • stan_ccd
        CORRECTION: Here is a corrected and more insightful method: For a given exp time: 1) take 2 virgin darks. Subtract one from the other and measure STD.
        Message 3 of 16 , Sep 16, 2013
        • 0 Attachment

          CORRECTION:

           

          Here is a corrected and more insightful method:

           

          For a given exp time:

           

          1) take 2 virgin darks. Subtract one from the other and measure STD. virgin_noise = STD*gain/sqrt(2);

           

          2) take 2 flashed darks (flash each dark). Subtract one from the other and measure STD. flashed_noise = STD*gain/sqrt(2);

           

          noise from flashing = sqrt(flashed_noise^2 - virgin_noise^2)

           

          Stan

        • bob_franke
          So how does this affect the workflow? Is it necessary to flash before each exposure, including flats and darks? My STF–8300 does not have flashing
          Message 4 of 16 , Sep 16, 2013
          • 0 Attachment

            So how does this affect the workflow? Is it necessary to flash before each exposure, including flats and darks?

            My STF–8300 does not have flashing capability. Can I simply use my flip flat at full brightness just before taking flats or darks?

            Cheers,
            Bob
             



            --- In ccd-newastro@yahoogroups.com, <stan_ccd@...> wrote:

            CORRECTION:

             

            Here is a corrected and more insightful method:

             

            For a given exp time:

             

            1) take 2 virgin darks. Subtract one from the other and measure STD. virgin_noise = STD*gain/sqrt(2);

             

            2) take 2 flashed darks (flash each dark). Subtract one from the other and measure STD. flashed_noise = STD*gain/sqrt(2);

             

            noise from flashing = sqrt(flashed_noise^2 - virgin_noise^2)

             

            Stan

          • rwright@starstonesoftware.com
            It s my understanding that the 8300 is less susceptible to RBI and thus it s not necessary. Richard ... It s my understanding that the 8300 is less susceptible
            Message 5 of 16 , Sep 17, 2013
            • 0 Attachment
              It's my understanding that the 8300 is less susceptible to RBI and thus it's not necessary.

              Richard

              On Sep 16, 2013, at 9:01 PM, bfranke@... wrote:

               

              So how does this affect the workflow? Is it necessary to flash before each exposure, including flats and darks?

              My STF–8300 does not have flashing capability. Can I simply use my flip flat at full brightness just before taking flats or darks?

              Cheers,
              Bob
               



              --- In ccd-newastro@yahoogroups.com, <stan_ccd@...> wrote:

              CORRECTION:

               

              Here is a corrected and more insightful method:

               

              For a given exp time:

               

              1) take 2 virgin darks. Subtract one from the other and measure STD. virgin_noise = STD*gain/sqrt(2);

               

              2) take 2 flashed darks (flash each dark). Subtract one from the other and measure STD. flashed_noise = STD*gain/sqrt(2);

               

              noise from flashing = sqrt(flashed_noise^2 - virgin_noise^2)

               

              Stan



            • astroimagek
              Bob, RBI isn t generally an issue with the KAF-8300 like it is with some other full-frame sensors, such as the KAF-3200. Christian Buil did a comparison
              Message 6 of 16 , Sep 17, 2013
              • 0 Attachment
                Bob,

                RBI isn't generally an issue with the KAF-8300 like it is with some other full-frame sensors, such as the KAF-3200.

                Christian Buil did a comparison between the the 532 (KAF-3200) and the 583 (KAF-8300) which included measuring RBI. You can read his findings here:
                http://www.astrosurf.com/buil/qsi/comparison.htm

                Test this yourself with your own camera. Checking for RBI is easy. Take an image of a star field with some particularly bright stars. Very bright stars, like the core of M42 can saturate the pixels in seconds. Take an exposure of a couple minutes or even longer. Then take some 30 second dark frames and measure the decay of the "shadow" over time. Measure the effect at different temps, perhaps 0C, -10C and -20C. RBI will persist longer at colder temperatures.

                Something else to consider about RBI even if you're using a sensor that is prone to RBI like the KAF-3200. If you're shooting a single subject during the night, RBI doesn't have a significant impact on your imaging or the data you acquire. By definition RBI will be prominent primarily under saturated stars or a bright galaxy core. Since those stars are going to be saturated anyway, adding a higher "floor" value of a couple hundred ADUs won't make the stars any *more* saturated.

                If however, you're doing precise photometry or you're imaging multiple targets per night, then with a camera like the 532/632, RBI is something you need to be aware of and manage appropriately. Pausing your imaging session for 10 to 15 minutes and capturing a few dark frames to throw away will eliminate the excess charge built up under saturated stars.

                Kevin
              • lmbuck2000
                if you are using RBI pre-flood on the light frames you definitely need to do the same on ALL of the calibration frames. at least for the camera i am using, the
                Message 7 of 16 , Sep 17, 2013
                • 0 Attachment
                  if you are using RBI pre-flood on the light frames you definitely need to do the same on ALL of the calibration frames.  at least for the camera i am using, the process introduces significant patterns in the light frames and if you don't calibrate with the same, it get's ugly.

                  here is what the regular darks vs rbi darks look like.  similar patterns appear in all frame types.

                  http://tinyurl.com/npmc2co

                  i did stan's measurements with results as follow.  the "answer" is very dependent on what area of the chip you measure STD but here are a few examples.   not sure what i take away from these values?


                  noise
                  flash noise
                  STD39.15
                  54.27
                  RBI STD54.82






                  27.81
                  117.7

                  87.75






                  23.86
                  90.04

                  67.99



                  i think the easiest way to see if your KAF-8300 camera has RBI is to take some dark frames right after your light frames.  i had never done that until recently and was surprised to see what happened to those darks with this KAF-16803 chip @ -30C.
                • Tim Stone
                  Kevin thanks for the greet explanation. Question: does RBI play any significant role in flat acquisition? ... -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Tim Stone
                  Message 8 of 16 , Sep 17, 2013
                  • 0 Attachment
                    Kevin thanks for the greet explanation. Question: does RBI play any significant role in flat acquisition?

                    On Tuesday, September 17, 2013, astroimagek wrote:
                     



                    Bob,

                    RBI isn't generally an issue with the KAF-8300 like it is with some other full-frame sensors, such as the KAF-3200.

                    Christian Buil did a comparison between the the 532 (KAF-3200) and the 583 (KAF-8300) which included measuring RBI. You can read his findings here:
                    http://www.astrosurf.com/buil/qsi/comparison.htm

                    Test this yourself with your own camera. Checking for RBI is easy. Take an image of a star field with some particularly bright stars. Very bright stars, like the core of M42 can saturate the pixels in seconds. Take an exposure of a couple minutes or even longer. Then take some 30 second dark frames and measure the decay of the "shadow" over time. Measure the effect at different temps, perhaps 0C, -10C and -20C. RBI will persist longer at colder temperatures.

                    Something else to consider about RBI even if you're using a sensor that is prone to RBI like the KAF-3200. If you're shooting a single subject during the night, RBI doesn't have a significant impact on your imaging or the data you acquire. By definition RBI will be prominent primarily under saturated stars or a bright galaxy core. Since those stars are going to be saturated anyway, adding a higher "floor" value of a couple hundred ADUs won't make the stars any *more* saturated.

                    If however, you're doing precise photometry or you're imaging multiple targets per night, then with a camera like the 532/632, RBI is something you need to be aware of and manage appropriately. Pausing your imaging session for 10 to 15 minutes and capturing a few dark frames to throw away will eliminate the excess charge built up under saturated stars.

                    Kevin



                    --
                    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
                    Tim Stone

                  • astroimagek
                    Tim, In general no, RBI is not generally a factor when acquiring flats because you re specifically not saturating pixels. As always, experiment with your
                    Message 9 of 16 , Sep 17, 2013
                    • 0 Attachment
                      Tim,

                      In general no, RBI is not generally a factor when acquiring flats because you're specifically not saturating pixels. As always, experiment with your equipment to achieve the best results.

                      Regards,
                      Kevin

                      --- In ccd-newastro@yahoogroups.com, Tim Stone <tim.stone.piano@...> wrote:
                      >
                      > Kevin thanks for the greet explanation. Question: does RBI play any
                      > significant role in flat acquisition?
                      >
                      > On Tuesday, September 17, 2013, astroimagek wrote:
                      >
                      > > **
                      > >
                      > >
                      > >
                      > >
                      > > Bob,
                      > >
                      > > RBI isn't generally an issue with the KAF-8300 like it is with some other
                      > > full-frame sensors, such as the KAF-3200.
                      > >
                      > > Christian Buil did a comparison between the the 532 (KAF-3200) and the 583
                      > > (KAF-8300) which included measuring RBI. You can read his findings here:
                      > > http://www.astrosurf.com/buil/qsi/comparison.htm
                      > >
                      > > Test this yourself with your own camera. Checking for RBI is easy. Take an
                      > > image of a star field with some particularly bright stars. Very bright
                      > > stars, like the core of M42 can saturate the pixels in seconds. Take an
                      > > exposure of a couple minutes or even longer. Then take some 30 second dark
                      > > frames and measure the decay of the "shadow" over time. Measure the effect
                      > > at different temps, perhaps 0C, -10C and -20C. RBI will persist longer at
                      > > colder temperatures.
                      > >
                      > > Something else to consider about RBI even if you're using a sensor that is
                      > > prone to RBI like the KAF-3200. If you're shooting a single subject during
                      > > the night, RBI doesn't have a significant impact on your imaging or the
                      > > data you acquire. By definition RBI will be prominent primarily under
                      > > saturated stars or a bright galaxy core. Since those stars are going to be
                      > > saturated anyway, adding a higher "floor" value of a couple hundred ADUs
                      > > won't make the stars any *more* saturated.
                      > >
                      > > If however, you're doing precise photometry or you're imaging multiple
                      > > targets per night, then with a camera like the 532/632, RBI is something
                      > > you need to be aware of and manage appropriately. Pausing your imaging
                      > > session for 10 to 15 minutes and capturing a few dark frames to throw away
                      > > will eliminate the excess charge built up under saturated stars.
                      > >
                      > > Kevin
                      > >
                      > >
                      > >
                      >
                      >
                      > --
                      > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
                      > Tim Stone
                      >
                    • kjcphoto2001
                      If you preflash your light frames, you need to preflash your dark frames. You do not need to preflash the flat frames. RBI is a charge trapped in the substrate
                      Message 10 of 16 , Sep 18, 2013
                      • 0 Attachment

                        If you preflash your light frames, you need to preflash your dark frames.  You do not need to preflash the flat frames.  RBI is a charge trapped in the substrate of the chip and it leaks out slowly.  It will show up in 10 minute darks.  That's why you must preflash the darks if you preflash your light frames.  Flats are usually exposures of a few seconds.  That's not long enough for RBI to show up.  Richard Crisp provides some excellent references on the subject.  He is a regular on the FLI and Narrowband Imaging yahoo groups. 



                        --- In ccd-newastro@yahoogroups.com, <knelson@...> wrote:

                        Tim,

                        In general no, RBI is not generally a factor when acquiring flats because you're specifically not saturating pixels. As always, experiment with your equipment to achieve the best results.

                        Regards,
                        Kevin

                        --- In ccd-newastro@yahoogroups.com, Tim Stone <tim.stone.piano@...> wrote:
                        >
                        > Kevin thanks for the greet explanation. Question: does RBI play any
                        > significant role in flat acquisition?
                        >
                        > On Tuesday, September 17, 2013, astroimagek wrote:
                        >
                        > > **
                        > >
                        > >
                        > >
                        > >
                        > > Bob,
                        > >
                        > > RBI isn't generally an issue with the KAF-8300 like it is with some other
                        > > full-frame sensors, such as the KAF-3200.
                        > >
                        > > Christian Buil did a comparison between the the 532 (KAF-3200) and the 583
                        > > (KAF-8300) which included measuring RBI. You can read his findings here:
                        > > http://www.astrosurf.com/buil/qsi/comparison.htm
                        > >
                        > > Test this yourself with your own camera. Checking for RBI is easy. Take an
                        > > image of a star field with some particularly bright stars. Very bright
                        > > stars, like the core of M42 can saturate the pixels in seconds. Take an
                        > > exposure of a couple minutes or even longer. Then take some 30 second dark
                        > > frames and measure the decay of the "shadow" over time. Measure the effect
                        > > at different temps, perhaps 0C, -10C and -20C. RBI will persist longer at
                        > > colder temperatures.
                        > >
                        > > Something else to consider about RBI even if you're using a sensor that is
                        > > prone to RBI like the KAF-3200. If you're shooting a single subject during
                        > > the night, RBI doesn't have a significant impact on your imaging or the
                        > > data you acquire. By definition RBI will be prominent primarily under
                        > > saturated stars or a bright galaxy core. Since those stars are going to be
                        > > saturated anyway, adding a higher "floor" value of a couple hundred ADUs
                        > > won't make the stars any *more* saturated.
                        > >
                        > > If however, you're doing precise photometry or you're imaging multiple
                        > > targets per night, then with a camera like the 532/632, RBI is something
                        > > you need to be aware of and manage appropriately. Pausing your imaging
                        > > session for 10 to 15 minutes and capturing a few dark frames to throw away
                        > > will eliminate the excess charge built up under saturated stars.
                        > >
                        > > Kevin
                        > >
                        > >
                        > >
                        >
                        >
                        > --
                        > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
                        > Tim Stone
                        >
                      • lmbuck2000
                        kjcphoto -- thanks so much for jumping in. i thought i needed to take rbi-flood bias and flats. you are correct, it appears to be unnecessary. i looked at my
                        Message 11 of 16 , Sep 18, 2013
                        • 0 Attachment

                          kjcphoto -- thanks so much for jumping in.  i thought i needed to take rbi-flood bias and flats.   you are correct,  it appears to be unnecessary.  i looked at my regular bias frames vs. rbi bias frames and they are nearly identical.  same goes for the flats.


                          i calibrated an rbi-flood light frame using an rbi-dark with regular bias and light and it looks great.


                          appreciate the info (and makes things simpler - and faster)


                          Lee


                        • kjcphoto2001
                          Lee, you re welcome. Preflashing flats does add a lot of time to your work flow. Also, if you are not scaling your darks, and making darks that are the same
                          Message 12 of 16 , Sep 19, 2013
                          • 0 Attachment

                            Lee, you're welcome.  Preflashing flats does add a lot of time to your work flow.  Also, if you are not scaling your darks, and making darks that are the same time exposure as the lights you are calibrating, you don't need to apply bias frames.  Adam block covers this in his teaching series on using CCDStack. 


                            I have been told that some people with RBI prone chips do not preflash because it adds noise.  Kevin's point about it not being necessary if you don't move the telescope is a workable solution.  I used my camera without preflash for awhile because I didn't have the driver to engage the feature.  I made some good images, but I did see ghost images of bright stars that I had to clean up.  Bright areas in galaxies and nebula may not appear to show any RBI, but it is there.  It's hidden in the brightness of the object and it may not be detrimental to your final image.  


                            The main problem I noticed was when I made my darks without preflash.  I noticed long swirl marks on my darks when I didn't preflash.  It looked like I had light leaking in from the back of the scope in the observatory.  I tried blocking the back and front of the scope but they still persisted.  And they were the same consistent pattern.  I realized that the chip had trapped charges in the substrate from previous exposure.  I would make my flats first, since they were quick to do and then I would leave the scope to make darks on it's own, since it takes much longer.  The flat frame exposures were preflashing the chip causing RBI to appear in my darks.  I had to wait at least 24 hours with no light exposure for the chip to get back to the point where I could make some clean dark frames.





                            --- In ccd-newastro@yahoogroups.com, <lmbuck2000@...> wrote:

                            kjcphoto -- thanks so much for jumping in.  i thought i needed to take rbi-flood bias and flats.   you are correct,  it appears to be unnecessary.  i looked at my regular bias frames vs. rbi bias frames and they are nearly identical.  same goes for the flats.


                            i calibrated an rbi-flood light frame using an rbi-dark with regular bias and light and it looks great.


                            appreciate the info (and makes things simpler - and faster)


                            Lee


                          • lmbuck2000
                            one reason i thought i needed to preflash the bias frames is i do use them to scale darks AND i use them to dark subtract the flats. fact is, the non
                            Message 13 of 16 , Sep 19, 2013
                            • 0 Attachment

                              one reason i thought i needed to preflash the bias frames is i do use them to scale darks AND i use them to "dark subtract" the flats.  fact is, the non preflash bias frames look identical to the pre-flashed bias and same for the flats.   (your comment made me examine them more carefully - than just my assumption --  we know what assumptions do) 


                              because of image shift from dithering and data shift from filter changes i have decided for cameras that matter, using the RBI pre-flood can only help.  but i do need to get some time to create an image (same object) with and without pre-flash to see if it is actually noticeable in "pretty pictures"


                              Lee

                              --- In ccd-newastro@yahoogroups.com, <ccd-newastro@yahoogroups.com> wrote:

                              Lee, you're welcome.  Preflashing flats does add a lot of time to your work flow.  Also, if you are not scaling your darks, and making darks that are the same time exposure as the lights you are calibrating, you don't need to apply bias frames.  Adam block covers this in his teaching series on using CCDStack. 



                            • Gregory
                              Forgive my ignorance, what is a “flash dark?” Gregory From: stan_ccd@yahoo.com Sent: Monday, September 16, 2013 9:36 AM To: ccd-newastro@yahoogroups.com
                              Message 14 of 16 , Sep 21, 2013
                              • 0 Attachment
                                Forgive my ignorance, what is a “flash dark?”
                                 
                                Gregory
                                 
                                Sent: Monday, September 16, 2013 9:36 AM
                                Subject: [ccd-newastro] RE: RE: Re: Is this RBI and how does it affect your imaging strategy
                                 

                                CORRECTION:

                                 

                                Here is a corrected and more insightful method:

                                 

                                For a given exp time:

                                 

                                1) take 2 virgin darks. Subtract one from the other and measure STD. virgin_noise = STD*gain/sqrt(2);

                                 

                                2) take 2 flashed darks (flash each dark). Subtract one from the other and measure STD. flashed_noise = STD*gain/sqrt(2);

                                 

                                noise from flashing = sqrt(flashed_noise^2 - virgin_noise^2)

                                 

                                Stan

                              • stan_ccd
                                A flashed dark is an otherwise normal dark frame taken immediately after flashing. Flashing floods the CCD with infrared light just prior to the exposure,
                                Message 15 of 16 , Sep 22, 2013
                                • 0 Attachment

                                  A "flashed dark" is an otherwise normal dark frame taken immediately after flashing.

                                   

                                  "Flashing" floods the CCD with infrared light just prior to the exposure, in order to saturates the substrate.  The CCD substrate can retain electrons; i.e. the pixel well is not completely emptied by readout. Those retained electrons bleed out during the next exp(s) to produce a "ghost image" from the prior exp. Flashing destroys ghost images by making the entire CCD a uniform ghost.

                                   

                                  If the camera does not have a built-in pre-flash device then another method is to avoid super-cold temps (colder CCDs retain more electrons in the substrate) or warm and re-cool the CCD (esp prior to taking darks).

                                   

                                  But don't worry about this unless you have a an obvious problem.  Usually ghosts are not a problem unless Vega or other very bright star is the field.

                                   

                                  Stan



                                  --- In ccd-newastro@yahoogroups.com, <fyrframe@...> wrote:

                                  Forgive my ignorance, what is a “flash dark?”
                                   
                                  Gregory
                                   
                                  Sent: Monday, September 16, 2013 9:36 AM
                                  Subject: [ccd-newastro] RE: RE: Re: Is this RBI and how does it affect your imaging strategy
                                   

                                  CORRECTION:

                                   

                                  Here is a corrected and more insightful method:

                                   

                                  For a given exp time:

                                   

                                  1) take 2 virgin darks. Subtract one from the other and measure STD. virgin_noise = STD*gain/sqrt(2);

                                   

                                  2) take 2 flashed darks (flash each dark). Subtract one from the other and measure STD. flashed_noise = STD*gain/sqrt(2);

                                   

                                  noise from flashing = sqrt(flashed_noise^2 - virgin_noise^2)

                                   

                                  Stan

                                Your message has been successfully submitted and would be delivered to recipients shortly.