|In Search of Solar Nirvana
(IR filter removal and other ill-guided stuff)
After additional sleuthing, I came to learn that CCD arrays in most consumer grade digital cameras have a hightened
sensitivity in the IR range. Believing that part of the spectrum to be disposable for general photography, some
manufacturers install what is known as a "Hot Mirror" filter to attenuate the signal beyond a specified cutoff frequency,
reducing the perceived "nusance". After several calls to NikonTech USA to determine the frequency response range
and if the CoolPix had a type of hot mirror filter, I was told that the information was proprietary and hence they would
not let me know if the camera could do what it was purchased to do! Being quite aggravated by this time (several other
Nikon SNAFUs had previously arisen), I politely told them that I was not competing to make a better camera... if that
was the case, I'd simply buy a Canon (Gasp!)
By now I suspected that such a filter lived inside my camera and might be part of the problems I was experiencing,
putting me into the position of examining my options. I wasn't ready to make the move to a dedicated astro-CCD (still
having so much to learn), and was unsure as to what other cameras have this same deficiency, leaving the
modification of this camera as the best option. Of course it might die on the operating table, but hey... good riddance
to some of the other aggravating Nikon foibles, AND perhaps just the excuse needed to justify a spiffy new Canon D60 !
Sooo... it went under the knife in a preliminary investigation, finding the offending filter and yanking it out like a bad
tooth! Measurements were taken so that an optical window could be fabricated and installed in its place if the removal
was to be permanent (otherwise it would be myopic... not coming to focus beyond a certain distance). For normal use,
an external Hot Filter and perhaps a color correcting filter could be used.
For the curious, this is what the filter looks like and its size (10mm X 11mm and ~3.0mm thick, earlier CP models are
~2.5mm) Please excuse the focus... myopia, remember? (taken right after surgery!)
Before ISO800 1/8sec After ISO800 1/8sec After ISO100 1/1000th
Using a TV remote control as an IR source (which is not really confined to a tight band), these panels depict the
camera's sensitivity at the indicated settings. The first panel, taken before the filter's removal shows a weak but
present signal with a reddish cast and is registered in all channels (distance of ~6 inches). The second and last
panels were taken after the filter was removed and at the same distance of ~6 inches. The second panel is at the
precise same settings as was the first panel, but shows a manyfold increase across the board. The final panel
still far exceeds the first, despite being lowered to ISO100 and the exposure bumped all the way up to 1/1000th
of a second. The hightened sensitivity across the board was an unanticipated result.
|Note: This is for informational purposes only!!!
Any such modifications are strictly unauthorized and will not only void
your warranty but have a good likelyhood of ruining the camera!!!
Please do not contact me for information or support concerning camera
modifications, as I strongly advise against doing them, and certainly do
not want to bear the onus of a ruined camera!
|(all images are uncropped and come directly from the original raw shots, save for being re-sized)
Having been frustrated with my attempts to capture the same level of Hydrogen-alpha detail in my solar images as was
present in the eyepiece, I undertook a journey to discover what was happening.
The most obvious of the anomolies was the inexplicable presence of "hot spots" (or "sweet spots", depending on the
exposure levels) having striking differences in demarcation with the remaining image. Generally circular in shape, they
follow the position of the solar disk in the image. As the disk is moved upwards in the image, the "spot" moves up a
proportional amount on the disk. Moving left, right, or down causes similar proportional movements of the "spot" on the disk.
|Note the "hot-spot" at the upper right
Splitting the images into their RGB components for further analysis revealed another anomoly that could not be
easily explained. The R channel had a bright, even intensity with no variations, while the best signal was located
in the G channel, and a reduced level mottled image was in the B channel. This should NOT have been possible,
as I used precision etalon and blocking filters to hold the frequency to 656.28nm with a bandpass of less than
0.07nm... RED(!!!), which means nothing should have been picked up by any channels other than the R channel!
Original Raw Image R Channel G Channel B Channel
The problems persisted and were traced to signal leakage inherent in the design of the CCD chip used by Nikon.
With a target frequency of 656nm, the leakage is 15% to the G and 10% to the B channels. The CoolPix 995design
allocates 50% of the pixels as G, and 25% each for R and B channels, thus the strongest visible H-a signals at the
LCD display occur when the G and B channels (with 75% of the total composite signal) have highest leakage and
the R channel (which contains the true H-a signal) is saturated.
Reducing the R channel below saturation lessens the effects of G and B leakage, but with the R channel contributing
only 25% to the image, it is very dark on the camera's LCD monitor (muddied by the now darkened G and B channel
components), and the resulting images are also very grainy (again, due to only 1/4 of the pixels are being used).
Soooo... this was my "ticket to ride" to a Canon D60, with the initial results looking much more promising. One step
closer to Solar H-a Nirvana!
And this my friends, is what the
Canon D60 has been delivering!
The Green and Blue channels now remain inky-dark until the
Red channel approaches gross saturation. Gone is an extreme
"hot spot" and (when properly exposed) the "sweet spot" is
now the entire disk!
Note the crisp definition and subtle shading of the filaments
and surface detail that had previously eluded me. The
prominences along the rim benefit from longer exposure, but
they too show a level of detail that was previously hidden.
For more information on this particular image, and H-a imaging
in general visit the 27-Dec-02 page, or for the entire Solar
section visit the Solar Imaging section.
The journey "In Search of Solar Nirvana" continues with Part 2 illustrating typical processing steps for an h-a image.