Staring at the Sun, 3

:: home ::
                Go to  1   2   3   4   5   6   7   8   9   10   11   12   13   14   15   16   17   18   19   20   21   22   23   24   25   >>  
                Solar Resources:  SDO | Sacramento Peak | X-ray Flux | Spaceweather | HF Radio | SRCH


Canon 50D (unmodified). Lunt LS60THa. Composite of double-stacked and single etalon data
to get best prominence and disk imagery. See details below. Click the image for wallpaper-sized view.


6/18/2010. I put the solar kit on the A-P mount (using the shaft extension as the only counterweight, which is too much, actually, but the robust servo motors have no problem with it) and took my time focussing and framing. Here's today's best workflow for H-a solar imaging: Use the 50D rather than the 20D, focus on limb detail in double-stacked mode. Shoot two images of the disk, each about 1/50 sec at ISO 400. Offset the Sun's image be several arc minutes between the two images. Remove the second etalon and shoot the edge detail. Start with 1/25 second at ISO 400.on A-P mount

Open the two disk photographs in Photoshop. Copy one into a layer atop the other. Align them. Select "lighten" as the blend mode. That eliminates dust, lint, and assorted other detritus that is otherwise troublesome despite multiple attempts to clean the sensor (the classic treatment would be to flat-field the image; I'll try that next). Flatten the layers in the image. Apply a strong unsharp mask to the central 95% of the disk (that is, avoid the limb): something like a radius of 6-8 pixels and 400-500%. I know those settings seem bizarre, but it works (see note below on oversampling the data; it's probably related to this point). [Note to self: get fanatical about focussing; there are results on the web with this kit that look vastly better than mine.] Additional smart sharpening may help. Season to taste. You might also try a dynamic mask to even out the surface brightness of the Sun. Although limb darkening and other shading owing to the narrowband optical train produces nice 3-D effects, they're a little distracting when compositing limb and disk detail.

Then process the limb image. That's much simpler: Drop out the disk (in the 50D, the disk today was 2,380 pixels across). Stretch, sharpen, apply a gentle Gaussian blur if needed to get rid of excess noise. Overlay the limb image on the disk image. You're pretty much done. Isolate the surrounding sky and apply a strong Gaussian blur to get rid of any lingering noise and render the sky a nice smooth gradient. [Look at later pages for more sophisticated and more successful workflows. dc]

Think about that image scale: 2,380 pixels to cover a 30 arc minute (= 1,800 arc seconds) subject. That works out to about 0.75 arc seconds per pixel. The aperture is only 60mm in single-stack mode and 50mm in double-stack mode. With such a limited aperture, the best possible resolution is no better than 2 arc seconds which means that I'm oversampling by a factor of 2.5 - 3. Daytime seeing may further compromise available resolution, although the atmosphere this morning seemed still and the solar image was very sharp, so that's less of a concern today than it might be. But then consider that only the red-filtered photosites are being used to sample the image because behind the H-a filter(s), the blue and green channels should see no photons at all and so can contribute only noise. The spacing of the red photosites is the square root of two greater than the spacing of all photosites on the image plane, so from a sampling standpoint, my pixels should be considered 1.4x larger than their physical dimensions. Uh, right?


Time out for notes for the unitiated:

I'm not at all sure about how all this shakes out and what it implies about appropriate unsharp mask settings and optimum file sizes. It may be that an early step in the processing of these files should be resampling down to half the original linear dimensions to avoid sharpening noise as much as signal. Or perhaps I should start by using Maxim DL to discard all but the red Bayer plane. Theory is tough. So here are some snaps of the hardware to make clear what I'm on about when I talk about prime focus adapters, snouts, and barlow lenses. That much I can explain with confidence.


snout   snout
Canon 20D attached at prime focus behind a B600 blocking filter; two-speed focus knob at right; electric remote release exits frame below camera.
1.25-inch adapter with barlow lens screwed to the nosepiece to gain backfocus and increase image scale.


Also, all this chatter about solar filters and etalons and stacking and .75A this and .55A that might need some explanation. There are two ways to render the Sun safe to view: reject almost all the photons with a filter made from a partially silvered mirror or reject all the photons except those of a specific wavelength. The first method gives a white light view of the Sun, "white" because the light that passes through the filter represents the full visible spectrum, only a million or so times dimmer. Sunspots, faculae, limb darkening — that's what you see in white light. The second method is far more interesting. Select the hydrogen-alpha wavelength of light and you see prominences, filaments, and flares. You do that with a filter (a Fabry-Perot interference filter, a so-called "etalon" made up of unbelievably precisely spaced reflecting and refracting elements). The idea is to admit only light with a wavelength of 6562.81A give or take 0.75A. (When two filters are "stacked" so that only light that passes through both reaches the eye, that becomes "give or take 0.55A"). An "A" is an Angstrom unit or one tenth of a nanometer, and a nanometer is one billionth of a meter. Light of that frequency is emitted by the hydrogen atom when its single electron falls from the third to the second available orbit.

This how narrow the bandpass of a 0.75A filter is: lay out the visible spectrum on a 100-yard football field with infrared light (7000A) at one goal line and ultraviolet light (3500A) at the other. A 0.75A bandpass would be represented by a narrow stripe 3/4 inches to either side of the 12.77 yard line on the red end of the field. The bandpass is so narrow that hydrogen on the Sun, accelerated by magnetic fields, is sometimes moving toward or away from the Earth fast enough that its light can be red- or blue-shifted outside the range admitted by the filter. Hydrogen-alpha filters are tunable: you turn a knob to shift the wavelength they admit by over an Angstrom either way, that is, you can move the center of their bandpass an inch or two toward either goal line. When using two filters, you can change the way their bandpasses overlap to reveal different details on the Sun.

As you were.



Five hours later, with the telescope sitting in direct sunlight and the heat of the day near 90F, I took a quick look to see what had changed. One prominence had morphed into a kind of acacia tree. And the filters had drifted substantially off band. A modest turn of the tuning knob on the front-mounted etalon brought all the detail back, as sharp and clean in the afternoon as it was this morning. Keep this in mind for when you try time-lapse: at least on a hot day, the filters can drift far enough to render disk detail very feeble and muted.

turningAbout 15:30 EDT, I took a sunfinder I'd turned from a short piece of aluminum stock out to give it a try. It works fine. [But it's unnecessary, see notes from next week: 6/25.]

At a glance, it was obvious that something energetic was going on on the Sun. The bright spot nearer the center of the disk in the frame at the top of this page had completely encircled a black, elliptical core, making a kind of bright, "6"-shaped ring around it. The center seemed pitch black. Tuning well off band did not lighten the core or dim the encircling shape. A narrow, dark, tapering tongue extended up from one end of the core at some etalon settings. Meanwhile, the "6" shape around it was brighter than anything else on the disk by a large margin. The immediate impression was of looking down into a whirlpool, or straight into a rifled barrel. If it weren't 93 million miles away, it would be scary. (Actually, it is a little scary: all that thermonuclear stuff going on right there in plain view. Every day. Right there. It's a wonder we aren't incinerated in a heartbeat every time we step outside.) reports that, "Tiny new sunspot 1082 is crackling with small solar flares. This one is 'only' about as powerful as ten million atomic bombs: [they include a movie clip here]. NASA's Solar Dynamics Observatory recorded the B4-category flare on June 17th. Each seemingly miniscule flash of light in the movie is a continent-sized blast dwarfing any explosion ever produced on Earth by many factors of ten... On the Richter scale of solar flares, B-flares are near the bottom, producing little or no effect on Earth. The really fearsome explosions are X-flares, more than a thousand times stronger than anything sunspot 1081 is likely to produce. X-flares will become more common as Solar Max approaches in 2013. Until then, relax. It's 'only' a B-flare."

Nice explanation. Very neat. Very tidy. But I think 1084 is wrapped up in the dimmer bright spot in the photo above. What I saw was at the other end of the long filament. By 18:00, all was as quiet as earlier in the day.


:: top ::


                   © 2010, David Cortner