DLR Linearity tests and AIP4Win

Hi Ed, Did you had a look at "Camera Characterization" under DSLR Documentation and Reduction ? It's not finished but cover a good % of the subject. I would say if you use a recent CMOS DSLR there is ingeneral no issue of non linearity ( <1%) with these devices, the only question is to know the clipping level at the lowest ISO that correspond to themax capacity of theimager.At heigher ISO the clipping isthe rangeof the ADC. The imager saturation (in fact clipping)isin general near the top of the ADCrange(i.e 14300 / 16384 for a Canon 450D) 1) I am not using AIP4Win but the mention " Star minus Sky ADU" seems ok as it shouldmean this is the ADU intensityat the star locationminus the sky background level. You could also use IRIS for that purpose, it's available on the web for free and install in one click ( Analysis / Aperture photometry ). http://astrosurf.com/buil/us/iris/iris.htm 2) In general the green saturates first as the sensitivity of the DSLR is much lower in red and blue ( ~50%) In case of very blue source (some LED) it could be the blue. By the way it's better to separate the colors. If you use the solution that is to vary the exposure time be aware the true exposure time is in general different than the value displayed by the camera ! For the cameras I know it's in fact a geometric progression with a constant factor instead the traditional exposure time displayed by the camera. Yours truly, Roger

Using the Star Image Tool and the ADU values from Star-Sky, I obtained the attached results, with the regression line fitted only to Series 1. So, I am assuming that so long as I stay under 10,000,000 ADUs for my target and comps, I should be able to obtain good data. Is this correct? Or do I need to use another way to measure. I seem to remember that Richard Berry emailed me some time ago about adjusting the raw ADU. Thanks for the feedback, Ed (WEY)

Hi Ed, Very good ! This is the result of the integration for all pixels of the star image in the photometry aperture.By the way the progressive saturation above 10 M ADUs comes from the fact only the brightest pixels areclipped by the ADC (analog to digital converter), more and more of them being clipped when increasing the exposure. In fact atISO 200 for a 450D you do not reach the imager saturation but the max range of the ADC (14 bits = 16384 - 1024 of system offset) This is due to the x2 additional gain of the analog amplifierthat drives the ADC when it isset at ISO 200.To test the imager saturation you shoulduse ISO 100. In fact you can't rely on that 10 M ADUs level as in practical situation you will not know the number of pixels being involved in the star image and its profil. If for exemple you have less pixels in that image the 10M ADUs sum would correspond to a clipping situation ! In fact you need to test the level of the BRIGHTESTPIXEL, either against ~14000 ADUs if at ISO >100 or against ~12000 ADUs if at ISO100 (after the 1024 system offset subtraction) In normal operation with sky images you would need to read the ADUs AT PIXELLEVEL that to determine the value of the brightest pixels and check ifthere is nosaturation either by the ADC or by the imager depending the ISO. Under IRISthis isvery easy to determine by using the "Statistics" function that provides the max value of the pixels at the CFAlevel (raw image) of aselected area.I have no idea if such function exists in AIP4Win... Yours truly, Roger

Thanks for all your comments, Roger. It is now coming back to me--Richard Berry relayed much of what you said some time ago in an email--I think I can find it and make the necessary corrections in AIP4WIN. In the meantime I will check out IRIS. Slowly making progress! Ed

OK, I found Richard's email and did the conversions using Christian Buil's data for the 40D (0.593e-/adu at ISO 200) and 1257 pixels in the aperture and came up with the break of 5000 electrons/pixel, perhaps a bit more (eyeing the graph). This seems fairly reasonable as I think the well well capacity is about 12000 electrons/pixel. I will do running more tests at ISO 400 and ISO 800. Any feedback greatly appreciated. Now, I assume I need to get under the sky and try different exposures for various magnitude stars to get a handle on exposure times for any particular target. Clear skies, Ed

Ed, My own characterization of the 450D shows 1.14 e- per ADU at ISO 200. The well capacity is about 30.000 e- per pixel (but onlyreadableat ISO 100 ) At ISO 200 the ADC range being limited at 16384 - 1024 = 15360, only 15360 x 1.14 = 17510 e- per pixel are readable, any above this level is clipped by the ADC. Tonight we got a very good sky after many days of poor conditions, and the result shows that the egress seems engaged ! Clear Skies too, Roger

Hi Roger: I assume you calculated gain using flat and bias frames. I have done this for CCDs. I will give it a try with the 450D. Many thanks for all your comments. Ed