Scintillation Detection via Digital Camera... experimental results indicate otherwise.

[Chris Trent]

I did a quick experiment tonight involving a sealed shoebox, digital camera, and chunk of scintillation plastic.

Equipment:

1 shoebox,
A Canon S5iS digital camera.
Black tape, and some black felt.
1 Block BC-408 scintillation plastic.
1 thorium mantle
1 Am241 source
1 Vaseline glass source.
1 spinthariscope

I took a shoebox, similar to ones that I have used for pinhole cameras in the past, and cut a hole in the top that fit the lens tube of the camera snugly. I made a lightproof seal between the lens tube and box with the black felt.

The source and block of scintillation plastic was placed in the box, sealed with black tape, and a 15 second exposure was taken at ISO 1600, in macro mode with manual focus set at about 10cm.

A series of images were taken with scintillation plastic placed directly atop each source, as well as the empty box for control and spinthariscope for reference.

Contrast was set for the entire group of images using the empty box for reference.

The images came out quite grainy from background noise, but with no discernible image, or difference in brightness between the control and images of the scintillation plastic with the various sources.

The spot in the spinthariscope images was readily apparent, as expected.

To further rule out the possibility of inadvertent light contamination an additional photograph was taken with the camera lens covered for comparison. It appeared identical to the other images once processed.

I'm not bothering to post the images here since they are little more than noise.
[Edit, Uploaded images for posterity's sake]

Conclusion: At the very least, the CCD on the camera I have is insufficient to register light from scintillation plastic over the maximum exposure time available.
Presuming that my camera has at least average low light performance for a modern digital camera then I must conclude that this would not be an effective method of radiation detection, let alone measurement.

-Chris

[Carl Willis]

I also find this general approach to be simply unworkable.

The below example photos are made with a Canon S3IS consumer-grade digital camera. All exposures are made for 15 seconds at ISO 800 and all are processed by the "auto white balance" tool in GIMP and subsequent scaling for use on the forum. When looking at scintillation plastic, the camera is set in its super macro mode. I also include some images made with a Precision Optics PS61 x-ray image intensifier tube with a CsI scintillation screen for comparison; these are taken with the camera focused through the PS61's optics at about f/3.5.

First image: a BC-400 plastic scintillator disk 2" diameter, 1/4" thick has been taped point-blank on the camera lens and a 5-mCi Am-241 gamma source is placed next to the disk.

Second image: the same plastic disk with no source next to it.

Third image: The Am-241 source as viewed by the camera via the x-ray imaging tube.

Last image: A 3-1/8 oz. shaker of Morton Salt Substitute (close to 100% KCl) as viewed by the camera via the x-ray imaging tube.

Conclusions: This was a fast and dirty experiment rather than an exhaustive study of the possibilities, but I cannot discern the presence of an extremely hot Am-241 gamma source using a camera and BC-400 plastic, which pretty much precludes this method's viability for much weaker sources like natural K-40 or from regional contamination from the Fukushima accident. I will also point out that the glow of the plastic near the Am-241 source is faintly visible to the naked eye and is easily detected by the PMT to which the plastic is usually coupled, as is the KCl shaker, but as I have learned, that does not make it visible to a pretty average digital camera. On the other hand, a standard medical x-ray intensifier tube that probably runs about $15,000 off the shelf can obviously pick up the Am-241 source (well enough to discern its geometry and even make radiographs with it), but it's debatable whether the KCl shows up (I'd say probably not). Certainly the bottle's shape cannot be recognized and noise dominates the image. The combination of consumer digital cameras and BC-400 plastic appears to have ZERO chance of functioning as a scintillation dosimeter for low levels of gamma radiation, despite a certain YouTube claim that got posted on our forum by someone who has been since banned for rules violations. I additionally suspect that claim of not only being inaccurate, but of being fraudulent self-promotional material.

-Carl

[Richard Hull]

I bit my tongue on this one too, feeling, "if you can't say anyhting nice"....... So, I left it as an exercise for the student. Thanks for pokin' at this doughboy, Chris and Carl.

Digital cameras just aren't up to the 10e6 photon multiplication factor of a PMT.

Richard Hull

[Chris Trent]

I added some images to my original post. The camera is plainly not up to the task.

It is quite clear now, after two independent tests that the original post on this technique is nothing more than promotion for another YouTube hoax video. I don't know what he was thinking putting dropping this into the domain of people willing and able to test it, but to quote Mythbusters, "This one is Busted."

[Dan Tibbets]

I'm not so sure that the blanket statement about the inadequacy of CCD's is valid.
It does look like this particular point and shoot camera fails. The first set of images did not have a dark subtraction (except the dark frame itself. This is obvious from the amplifier glow in one corner.
The comparison by CW is interesting. That the glow of the scintillator is dimly visible, means that a CCD should easily pick it up. Perhaps between all of the coated lenses, RGB filters, antialiasing filters, infared cutoff filter and ultraviolet cutoff filter(?) the frequency of light from the scintillator is considerably suppressed even before it reaches the CCD.

The small ~ 1.5-2 micrometer sized pixels in point and shoot cameras do not collect much light, and tend to have small signal to noise ratios without bright lighted scenes.

Even with a macro setting, the focal plane is generally up to ~ 1 or more feet away from the front of the lens. Having a light source up against the lens would be badly out of focus, and much of the light may actually be falling on areas outside of the CCD.
Proper lens characteristics could focus the entire scintillation surface to a central few pixels, instead of diffusing it so that only a portion of the light falls on all of the pixels. This alone could magnify the light signal by perhaps several million.

White balance is not very important in processing,. What is important is playing with gamma, and the histogram to maximize the shadow contrast.

A much better candidate CCD would be one in a digital SLR camera. They have bigger and thus more sensitive pixels, with better signal to noise characteristics. You can remove the regular lens and use a cheap magnifying lens (even a magnifying glass would work) or wide angle lens(?). There would be much less light lost in the lens assembly, The light could be focused to a small area of the chip- a few pixels for maximum sensitivity, several pixels to perhaps a few thousand pixels if you are trying to obtain some spacial resolution. This focusing to a small area is very important, as you could get magnification gains similar to PMT, except the concentration is done optically, rather electronically. Of course a PMT could also do this optical concentration.

Finally, the filters may be greatly dampening the wavelength of light that the scintillator emits, and can vary from different cameras.

The best solution is to use astronomical B&W CCD (or CMOS) cameras, with the appropiate lens, cooling, and software. No cutoff filters, no RGB filters, no antialiasing filters, etc. In another thread here, the author used an ancient CCD B&W camera (designed for astronomy?) and did detect scintillator glow without optimizing the optical concentration. This camera probably only had quantum efficiencies of 20-40%, while modern astronomical CCD cameras may be ~ 80-90%. Also, the readout noise tends to be much less than the older cameras. An old camera my be doing good to have readout noise of ~ 15-25 electrons RMS. Modern ones may be closer 3-5 electrons (varies depending on pixel size). Cooling is important as the background random dark noise doubles for each 10 degree rise in temperature.
Finally, consumer point and shoot cameras generally have maximum exposure times of 3-15 seconds. DSLR cameras ~ 30 seconds, and have a Bulb setting. Practical exposures of up to several minutes may be obtained. Also, images can be stacked.
Cooled astronomical cameras can take exposures of many minutes if desired.

Rough guesstimates of the improvements over the camera setups mentioned in this thread follows.
Proper focus (probably minimum focal length setting with the light source in focus (definatly not right in front of the lens , but at a distance determined by manually focusing on a substitute target. If the camera does not have a manual focus, who knows what the camera is doing as it cannot focus on a very dim object.. A wide angle (or is that a magnifying lens(scratching head), or collimating lens in front of the source may also help.

Better focus, lens arrangement- 2 to 1000 X improvement (eg: shining the light on 12,000 pixels instead of 12,000,000 pixels)

Larger pixels - ~ 5-15 X improvement ( related in part to the first item)

Eliminating various filters, and reducing the lens elements- ~ 2 to 100 x improvement (depends on the light wavelength vs filters)

Cooling- ~ 10 to 20 X improvement

Proper post processing- ~ 1 to10 or more.

Increased exposure times- ~1 to 30 X improvement, or more.

Total improvement possible- ~ 200 X to 10,000,000,000 X

The top number may be rather optimistic, but it illustrates the range of results possible depending on your setup.

This is essentially the same inputs that determines how dim an object can be detected with a telescope and CCD camera. Limiting visual magnitude is ~ 6. I have taken pictures with a 27 inch telescope and an sBig CCD camera that has exceeded 21.5 magnitude. Each one increase in magnitude represents a ~ 2.5 dimmer object Each 5 magnitudes is ~ 100X dimmer. So, 15 magnitudes difference means ~ 1/1,000,000 as bright objects has been detected compared to the human eye. The objects were generally stars or asteroids, and the light was generally spread over 4-8 pixels. The light pollution at the site is modest, but much worse than at an ideal dark sky site. Probably 1-2 magnitude deeper exposures could probably be taken at an excellent dark sky site (equivalent to a very good light tight box). The light ( a star or asteroid) is focused onto only a few pixels. If the focus is poor or there is a lot of air turbulence or if the telescope is bouncing in the wind, the light is unfocussed and detection can suffer by as much as 2-3 magnitudes. If there is poor transparency (thin clouds, high humidity, etc), the detection suffers further (equivalent to multiple lenses and filters).
All of this has direct correlation to my above arguments. Results can vary widely with the same equipment depending on how you set it up.

Dan Tibbets

[Dan Tibbets]

PS: I'm uncertain what 'automatic White balance' would do. It might help a little or hurt a little. I'm not sure what alogrhythms is used in the auto mode. Set modes generally adds or subtracts from the RGB values based on the set light source (Sun, incandescent, fluorescent, etc.). In an image with only one color RGB channel showing a weak or very weak signal (assuming the scintillator fluoresces at only one wavelength) who knows what wild contortions the software would go through in trying to produce a pleasing image tome. If the scintillater light is blueish, setting the white balence to incandesent, or a cooler temperature, may result in the blue channel being magnified more over the other channels. Better would to split the image into RGB channels, and work with only the blue channel.

Also, if you wondering why I used such a broad range for the possible color filter effects, it is because in these Cannon cameras, the green and blue filters may not have sharp cutoffs, The light transmission is more curved, and if the narrow spectrum light source falls on the minimum between filter ranges, the transmission may be very low. Also, the extremely complex RGB demosaicing (sp?) software may be confused and throw the data out. True B&W (not post processing synthesized B&W), or at least LRGB (if they made them) is much better for light sensitivity in general (3 -4 times better) and especially if a large portion of the light is a wavelengths in the border regions between the RGB filter transmission.

Dan Tibbets

[richnormand]

Reminds me of this thread a few months ago:
viewtopic.php?f=13&t=6179#p34623

Posting #12 is with a regular CCD camera with zero results! Regular CCD (not scientific grade), colour (RGB filters absorb available light by about three unless you are using three CCDs with dichroic filters), optics, etc....

A few #s later a positive results was with a Cs test source with NaI scintillator with a -40C cooled scientific grade CCD, so not much of a chance for a run-of-the-mill cheap consumer camera I think.

[John Futter]

Well done guys
If you go back and look at the Fe55 activation that I put in the files section and view the video you will notice bright pixels. That was a brand new serveillance camera that I used and it now worthless due to the CCD damage.
That damage was caused by 12MeV 2-3uA proton beam in air spallation products. Neuron rate was a bit over 2Rem/Hr for the 30seconds that we had the beam on.
somone else can put up the products from N, O, Ar, C.
So for high energy type irradiation CCD damage is of some use.
But why? when dosimeters do it more accurately and are cheaper.

And please note that the dose to damage the camera is much less than a lethal dose to humans. If I had been present with the camera my dosimeter badge would certainly have shown it with a big please explain how this happened. My guess as to total flux Neutrons /Gammas would have been around 50mrem at 3 meters from the beam

Edit
Just remember that the flux that did this damage to my camera is many giga times a check source.

[Chris Trent]

Dan,

I agree with you on all points, however this was specifically a test of the assertion that off the shelf consumer grade digital cameras could be used with literally any scintillation material to detect and measure radiation with no special tools, modification, software, or user training.

In that respect it failed miserably.

Off the shelf, 15 seconds is the maximum exposure for my camera. It does not natively support RAW format, and the processed JPEGs are not well suited for dark frame subtraction.
Obviously there's not much I can do about the lens or filter arrangement of a point and shoot, but it is in fact able to focus at 6mm from the front lens, somewhere in the middle of the scintillation material I am using.


It's funny that you should mention astrophotography though... Tonight's experiment involves using some of the tricks of the trade. I'm in the process of taking a series of 30 minute exposures, in raw format, with dark frame subtraction, that I will be overlaying to hopefully generate an image of the scintillation pattern from the check source. Alas, I don't have a high grade CCD so a hacked camera will have to do. I'll let you know how it goes.

[Dan Tibbets]

Why do it, when almost anything is better for ionizing radiation? Because with a scintillator, the detection of neutrons might be possible at substantially reduced costs. I am not surprised that the CCD was destroyed by a few REMS of direct radiation. That is alot of energy. Any CCD will degrade as it is bombarded by radiation- cosmic rays will damage a CCD, but the rate is obvously acceptable. Even attempts to detect neutrons directly on the CCD by proton recoil would be expected to damage the CCD. The dose rate is of course the critical factor. If this proton beam destroyed your CCD in a few seconds, a fusor producing a few micro REMS per hr, might destroy the camera in a few million seconds of exposure (assuming a lineer effect). Then there is shielding and standoff considerations. Also, X- rays cannot be ignored.
Of course a proton beam in air does not penetrate far so the camera would need to be very close. If you were shooting through a lens, I'm not sure any protons would reach the CCD. The damage may have been due to secondary X-rays produced by the proton collisions in air and the surface of the camera. Even if it was a bare CCD, there would be a thin glass cover on it, so again I don't know how many protons were reaching into the CCD substrate. With a few REMS beam there may have been some significant penitration, but again the X-rays or just induced heat may have done the damage. I believe ~1 milliwatt of heating power is available from ~ 100 mcro REM(?) of neutrons (~ 500,000,000 D-D fusion neutrons) . The conversion to high energy protons would throw in a fudge factor, but ~ 1 REM of ionizing protons would possibly impart ~ 10 Watts of heat, even if the high energy protons was not getting through the cover glass. A few or modest number of seconds could be enough to heat the CCD to destructive temperatures, and that is without any direct high energy particle effects on the CCD itself.

PS: Cosmic rate hits (mostly protons, at least in space) are very easy to detect on CCD's. The Hubble space telescope images are severely compromised by this noise, and they have to go through multiple contortions to eliminate most of the noise.

Dan Tibbets

[Dan Tibbets]

I stand corrected on the minimum focal distance of the test camera, but hedge my admission. The depth of field (what is in focus) is very shallow in this range - perhaps a few millimeters, especially if your F-stop was wide open. So, part of the scintillator would be in focus, other parts less so. The amount of the effect would certainly be less, perhaps much less than my first assesment.

I think a minimum setup would be a DSLR (or newer interchangeable lens mirro less camera (like an Olympus EPL1), a wide angle, low F number camera lens, or adapted other lens. and some manipulation software- The GIMP would work.

Though... Hmm...
Adustuing the focus (say to infinity and/ or moving the scintillator further away may bring the total surface area of the glowing scintillater sheet to focus on a small portion of the CCD. This would provide the light magnification I described befor. This may be significant.
I'm not sure of the best arrangement, expermantation would be needed. The best arrangement if you have a spare simple lens or cammera lens, would be to place this lens close to the scintillater (or equivalent size brigher source (like a carboard cutout in front of a light), then adjusting the focus and or distance from the light, so that you get a columated beam of light (like in an ols slide projector). Then shine this light at the camera, adust the distance until you get a spot of light in the center of the CCD image.

I think I will try this myself, except in stead of a block of irradiated scintillator I will use a old film slide viewer.

Dan Tibbets

[Chris Bradley]

Just got back from a trip to find these attempts and couldn't help pick the challenge up, seeing as it is now midnight here, all is dark, and it is an 'easy-to-do'!!

I have what I consider to be the plus-ultra of 'consumable' cameras for low light - an Olympus E-600 DSLR (same sensor as the 620) which [for a digital] has a staggering max of 30 minutes bulb!.. and also very little noise for a 'pro-sumer'. (I bought it specifically for imaging my very dim plasmas, as per another thread. It took me a long while to find a digital slr that could do this, as it is not info normally listed. I recommend it as a 'scientific' camera, if anyone is looking for such a thing.)

The image below was taken in what I think you would ordinarily consider total darkness. A piece of uraninite, in a poly sample bag (which reads ~50,000cpm from outside the bag on my mini-900) sits directly atop a piece of the Saint Gobain 408 scint material that Geo currently sells (which I believe is listed by SG as the optimum choice for beta sensitivity, in that particular family of material/fluor).

The exposure was 60 seconds at ISO 200, then pushed digitally by ~15 stops and then smoothed with Noiseware. The light that does illuminate the picture is either starlight (!) or a light left on in a room downstairs getting underneath two closed doors!

I could take a longer exposure, but it hardly seems any point, seeing as you can already begin to see the piece on top of the scint block before any 'glow'.

Maybe I'll do a 10 min, or longer (how long would I need!?!?), exposure tomorrow night and try to get the room even darker, just to try to get *some* glow detected!

[richnormand]

@Chris:

Using the CH250 camera in my previous posting with a few minute exposure in my dark basement around one in the morning night and the windows covered with black felt I was surprised I got a very nice photo of the room! Found the illuminating light was from the three TEK oscilloscopes, the transistor curve tracer and the logic analyser screens glowing even after almost 5 to 10 minutes..... Can see them immediately after shutting off the room lights, but certainally not after 10 minutes, but the camera can! You may have to hunt down unexpected sources of photons!

PS: I have an Olympus E500, but looks like your E600 is much better and might be worth the upgrade. Comments, same lens system (4/3)?

[Chris Bradley]

[[*please excuse the quick topic drift* I think it has the same lens system, yes; the Olympus digital type. I was looking at the 400/500 also, but went for the 600 because the 13 million pixel sensor was given some extremely fawning reviews at the time, plus the image stabilizer (only on 600/620) is a fantastic boon and effectively boosts your ability to take high quality low light exposures (but only to 2 sec, I think). I have only praise for it, perhaps excepting for the menu system which is a real beast!]]

[Chris Trent]

Here's a much more effectively processed set of images.

These are four minute exposures, dark frame subtracted, adjusted to the same values.

The first is 1uCi Am241 behind a ZnS(Ag) screen. It actually produced enough light to make out the outline of the screen in complete darkness.

The second is the same source, with plastic scintillation material instead. Processed the same way.


I have also tried exposures of over 30 minutes with the scintillator, and the results were no better. I'm calling this a wrap, for my equipment at least.

[Frank Sanns]

Some of my posts on a previous thread showed nice alpha glow from an Am-241 source but nothing else. I tried again with three different scintillator plastics and with two different high speed x-ray intensifiers. Then I tried Registax of 10 images. Nada. Nothing at all, just darkness. There just is not enough photons coming out of a hot piece of uranium to be seen by any consumer grade camera including one of the finest and most sensitive DSLR camera. My Canon 5D MkII set at iso 25,000 with a 50 mm f 1.4 lens at 30 seconds saw nothing even with dark frame subtraction.

Frank Sanns

[Dan Tibbets]

What is the difference between the ZnS(Ag) screen and the plastic scintillator? Obvously the phosporesent output is far different. If the ZnS(Ag) material is on the surface of the screen it would be exposed to nearly the full radiaion dose if it is Alpha or Beta. I suspect the plastic scintillator has a dillute distribution of scintillaor active compound dispersed throughout the substrate, so only the small percentage of of scintillator at the surface would respond, and most of the radiation would be absorbed by the plastic as undetectable heat.
What is the concentration of the scintillator in a plastic, 1 PP10, 100,1000?
It seams reasonable (everything else being equal) that the Signal/ noise ratio between the two images would give a minimum ratio for the sensitivity (to non penetrating radiation) of the two systems.

Dan Tibbets

[Richard Hull]

After all of this, we can conclude........

No digital camera had for under $1000.00 can see any nuclear event or collect them enmass from scintillating plastic.

A cheap camera with suitable time exposure and image processing can record long term, collected alpha scintillations on a good ZnS:Ag screen. Even the eye can see these.

In short, if you can see a nuclear event with a youthful, naked eye, it can probably be collected over time with most any decent digital camera in a cleverly designed setup using a little software manipulation.

Richard Hull

[Dan Tibbets]

I tried taking some pictures with a D100 Nikon DSLR camera of a dimly illuminated screen. I thought going from a telephoto setting to a wide angle setting (shrinking the image on the CCD would increase the S/N ratio (brighter image)).
But. the images were identical in brightness. Doh! ... of course they are, the F number was the same. So my comment about going to wide angle setting is worthless if you have a constant F number lens. Now, if the wide angle sitting allows for a lower F number, then the brightness will increase accordingly. An F 1.4 fixed lens was impressively brighter than my F 5.6 setting on the Zoom lens. The F number improvement would be ~ 16X. Also the fixed lens probably had fewer lens elements , so the light transmission would be even greater, perhaps by a factor of as much as 2X. ( depending on the quality of the anti reflective coatings on all the elements of the two lenses.

Trying to shrink the illumiated area on the CCD by reducing the magnification of the lens does not help, with a dispersive light source. The question remains if a collimating lens that is very close to the source and covers all of it's surface would help. I may need to dissect my old slide projector to get at the collimating lens. I think this would would help.

It is not looking good for my contention that even DSLR consumer cameras would work with plastic scintillators. But, would it work with an image intensifier screen, or a night intensifier screen? A low light video camera, even with only a few stacked ~ 1/ 20 sec exposures can detect ~ 14th magnitude stars with a good intensifier.

That leaves direct detection of neutrons on the CCD throug proton recoil. I think the possibility is dependant on the expected neutron flux and how likely it is to interact in comparison with cosmic ray background levels/ reaction rate that CCD's do easily detect. A good, preferably large CCD with constant temperature cooling would probably be needed, if it works at all.

The cost for this stuff is approaching or exceeding what a used neutron, or certainly what a bubble detector would cost, but if you have surplus or multi - use stuff .....

Dan Tibbets

[Richard Hull]

Again, given enough cash, one can image anything in the nuclear world. The start of this thing was that a cheap nuclear detector solution for fukushima sufferers was at hand. It morphed into a general hunt for a cheap nuclear imager, then maybe a bit more expensive one and we find.... No dice....As it has been demonstrated over and over, zero imaging of scintillation plastic is done with any camera.

If one starts running in auxilliary gear, enhancers, intensifiers, super coolers, astro-cam systems, etc., spending the long buck and going to grand complications then, probably yes, an image may be had.

Got the money? It appears a lot of folks have the time. A few probably have the gear here, but where does that leave the budding fusioneer wanting to image or detect neutrons who can spend only $50.00 as might have been hoped in the great original posted URL?

Richard Hull

[Frank Sanns]

I had access to a forth generation night vision optic on a .338 Lapua so I thought I would give it a go for detecting scintillation light. The night vision is optimized for near IR and not for the blue of the scintillator but I was curious. The night vision device could see more than my camera could at ISO 25,000 and 30 seconds of exposure. It could see the glow of an old radium dial like it was a brand new phosphorescent one that had just been exposed to the light. There was nothing however when it came to seeing any kind of light from any of several different scintillators that I tried. Like Richard said, it is very difficult to beat a PMT that is optimized for the proper wavelength and can detect single photons. All for 50 bucks!

Frank Sanns

[richnormand]

OK, with all this talk about scintillators I dusted off the old CH250 (TEK512 back thinned CCD, temp -45C, 16bit A/D, dark current 1.2e/sec, well capacity 308.8 kilo electron per well at max A/D, estimated quantum efficiency in the deep blue must be around or less than 30% I guess) and removed it from the telescope. Then cut a 1" by 1" by about 2" of plastic scintillator that I got years ago from Don Orie (I think it is similar to Bicron BC-400, same type I used for my muon telescope). Polishing took most of the experiment time!

Source is an alpha point source.

First exposure was 1 min with the source alone. Nothing. Just to make sure the source does not emit any light by itself:
viewtopic.php?f=13&t=6179#p34623

Second exposure with the source taped on the plastic scintillator as shown. Nice signal, you can make out the outline of the square 1" by 1" scintillator exit surface as well as the reflection of the inside wall of the 2" PVC pipe from all the light emitted (well not that much really).

Tried same thing with my Olympus E500 with 1 min and max everything: nothing at all.

EDIT: Added two pix. Used same setup with a 5uCi Cs137 source sitting on top of the scintillator block. Put a small piece of paper taped on the back to help in focussing and inserted the lot in the tube. Photo is with the back removed. Source is black rectangle on top, paper is in focus and front face edges are slightly out of the depth of field.

Last photo is a 15min exposure with CH250 (and the end cap back on!!)

Again tried 15 1min exposure with the DSLR camera and post processing with no results. Just not enough light for a consumer camera.

[richnormand]

Last try today.

Used the same setup as previous measurements above and did a 4 hours exposure with a full 311g container of KCl 100% salt substitute. You can almost imagine the scintillator outline but image processing did not enhance that feeling!

Did not even try the DSLR this time!

As Chris said and as far as I am concerned "This one is busted" unless someone comes up with a magical scintillator material.

[Chris Bradley]

I also felt there was one last attempt due just to totally close it out and have a laugh at taking images of nothing!

So I repeated, noting [as my eyes adjusted on a longer run] that the extra light, as my first attempt, was actually a street light in another street causing what I thought at first to be opaque curtains to glow.

These were then repeated but with the additional mitigation of being done inside a black silk-lined luggage case with black-out material then draped over it.

I also tried a smaller pixel count, on the theory that maybe it then averages the result from more sensor pixels contributing to each jpg pixel.

First up is a 10 min run at ISO3200 on my E-600.

Next is the same but done for 30 min.

haaaahaa!!!....

[[Nonetheless, it's a great advert for the E-600. It (the camera hardware) takes the same time to process the image than to take it - so te pic took an hour to take - and presumably it does something clever with the noise. Strage bit of noise at the bottom, though? I must get on to trying out some atronomical photos, in preparation for YU55 on Nov 8th?!]]

[richnormand]

@Chris,
"(the camera hardware) takes the same time to process the image than to take it - so te pic took an hour to take - and presumably it does something clever with the noise"

The CH250 has a "flat" mode that takes an exposure the same amount of time that the photo was taken. The flat mode does not open the shutter and only records the dark noise acumulation in each CCD well. You can then do a photo/flat to normalise the two. This can then be substracted with a "ref" image (taken with the shutter open but no outside signal. This procedure reduces the effect of CCD defects and site dependent response. In addition taking the ref pix substracts out backgroud signal from your setup leaving only the desired signal. All three of these have to be taken at the same temperature and length of time of course.

Corrected Image = ([New image - Reference] x Mean Pixel Value)/([Flat field - Reference])

In your case if the exposure is 30min and the camera then acumulates a flat image for another 30 min (without opening the shutter) to remove CCD effects. This would explain the one hour required to take it and assume there is on-board memory for two full frames and a place to put the intermediate result.

Just a guess but I would think a room temperature CCD would accumulate lots of dark counts thus the need to biased out at each individual CCD site.