Photographing the impact plume
0 views
Skip to first unread message
Arnold Ashcraft
Dec 2, 2008, 4:08:16 PM12/2/08
to lcross_ob...@googlegroups.com
Folks,
I would like to image the impact plume from the LCROSS mission and I would like to use the correct exposure and focal ratio to maximize my chances of getting something. In slide 24 of the talk presented by Dr. Jennifer L. Heldmann at the May 2007 AAS conference is a plot of the expected impact plume irradiance as a function of wavelength and time. The units given for the irradiance is watts per meter squared per micron per steradian. I did a rough integration of the plot over the range of wavelengths my camera CCDs are sensitive to and got an area of about 20 watts per square meter per steradian.
From the aperture and focal length of my telescope and the area of a pixel on my cameras, I calculated that if I placed the CCD of my DMK31 camera at the un-amplified focus of my 12.5" f/6 Newtonian, I will get 0.012 microwatts of power deposited in each 4.65 micron pixel. Doing the same calculation for my SBIG ST-8 and assuming 2x2 binning of its 9 micron pixels, I get 0.18 microwatts in a 2x2 bin.
That's where my calculations ran out of gas. I would like to relate this power level to the exposure I would need to get a perceptible image (say, 10x the noise level from dark current) with either of the cameras. If possible, I would like to take either a video or a time-lapse sequence of images from the DMK31, or if it is not sensitive enough, a time exposure using the ST-8.
Does know know how to combine the power levels I calculated from the data on slide 24 with the kind of information tabulated in the spec sheets for CCDs to estimate the exposure needed to get an image?
Clif
Jim Mosher
Dec 3, 2008, 11:27:28 AM12/3/08
to LCROSS_Observation
That's a really great question, Clif! Some special planning will
certainly be required if the predicted impact plume is too faint to
register with the combination of f/numbers and exposure times normally
used for lunar observations.
In addition, I would guess that all estimates of the plume size and
brightness must be extremely rough, since the LCROSS scientists
probably have only a vague idea of how the surface will respond to the
impact (how much lunar dust will it kick up?) and even less idea of
whether it will contain a (more reflective?) water/ice component
(since deciding that is, I think, the point of the experiment).
That said, if they want their predictions to be accessible to non-
scientists (and probably to other scientists as well) it would seem
helpful if they could supplement the raw photometric predictions with
some kind of comparison to more familiar terms of reference. For
example, a curve showing how bright the predicted plume is in
comparison the normal lunar surface would be useful. It would seem
much easier to plan observations if one knew, for example, that (at a
particular wavelength) the plume after 60 seconds is expected to have
(over an area of 10 km ?) twice the brightness of the nearby normal
lunar surface, or one-hundredth of it, or whatever.
Given such information you could easily make a test run with your
telescope adjusted to simulate those conditions (for example, by
stopping it down or using a neutral density filter if the brightness
is expected to be some fraction of the normal lunar surface
brightness), and seeing if you could detect anything under those
conditions. That would completely avoid any need for complex
calculations of the sort you are attempting, and the many additional
assumptions that are needed to complete them (e.g., is the curve
presented in the slide the intensity before or after attenuation by
the Earth's atmosphere?, how great are the additional losses in your
reflector?, etc.).
Hopefully at some point before the impact, the LCROSS science team
will give a more clear indication of how their predictions for the
size and brightness of the plume relate to some easily accessible
reference (such as normal lunar surface features); although based on
their participation in this forum to date I wouldn't hold my breath
for an answer.
As a rough guess, I believe the surface brightness of the Sun (at its
peak in the visible and as observed above the Earth's atmosphere) is
said to be about 26,000,000 W/m^2/micron/sr. The "Full Moon" (which
is not really the real Full Moon, but rather something obtained by
extrapolating the brightness from other phases) is said to be about
1/400,000th of that, giving an average Full Moon lunar surface
brightness of something like 66 W/m^2/micron/sr. At the First and
Last Quarters the surface brightness is said to be only about 1/5th of
this value, which would work out to something like 13 W/m^2/micron/sr
(the total light from a Quarter Moon is only about 1/10th of that from
a Full Moon because only half the surface shines with this reduced
brightness).
I haven't had a chance to download the AAS presentation you are
referring to:
http://lcross.arc.nasa.gov/observation.htm
but it sounds like the slide you mention is the one that shows a peak
plume brightness in the visible (10-60 seconds after impact) of 35-40
W/m^2/micron/sr. If my numbers given above for the "normal" Moon are
correct, this would seem to imply that the LCROSS scientists think the
initial impact plume will be considerably brighter than a typical
lunar surface feature at Quarter phase, and probably even brighter
than the maria (but not as bright as the brightest surface features)
in a nearly Full Moon. Of course, I have no idea if this is actually
what they are predicting or not.
I'm also not sure how large a diameter the peak surface brightness
quoted at 10-60 seconds after impact is supposed to extend over; and I
would assume that whatever the initial surface brightness is, the
plume's surface will, at some point, start to fade as it grows.
-- Jim
On Dec 2, 1:08 pm, Arnold Ashcraft <wa2...@optonline.net> wrote:
> Folks,
> I would like to image the impact plume from the LCROSS mission and I
> would like to use the correct exposure and focal ratio to maximize my
> chances of getting something. In slide 24 of the talk presented by
> Dr. Jennifer L. Heldmann at the May 2007 AAS conference is a plot of
> the expected impact plume irradiance as a function of wavelength and
> time. The units given for the irradiance is watts per meter squared
> per micron per steradian. I did a rough integration of the plot over
> the range of wavelengths my camera CCDs are sensitive to and got an
> area of about 20 watts per square meter per steradian.
>
> From the aperture and focal length of my telescope and the area of a
> pixel on my cameras, I calculated that if I placed the CCD of my
> DMK31 camera at the un-amplified focus of my 12.5" f/6 Newtonian, I
> will get 0.012 microwatts of power deposited in each 4.65 micron
> pixel. Doing the same calculation for my SBIG ST-8 and assuming 2x2
> binning of its 9 micron pixels, I get 0.18 microwatts in a 2x2 bin.
>
> That's where my calculations ran out of gas. I would like to relate
> this power level to the exposure I would need to get a perceptible
> image (say, 10x the noise level from dark current) with either of the
> cameras. If possible, I would like to take either a video or a time-
> lapse sequence of images from the DMK31, or if it is not sensitive
certainly be required if the predicted impact plume is too faint to
register with the combination of f/numbers and exposure times normally
used for lunar observations.
In addition, I would guess that all estimates of the plume size and
brightness must be extremely rough, since the LCROSS scientists
probably have only a vague idea of how the surface will respond to the
impact (how much lunar dust will it kick up?) and even less idea of
whether it will contain a (more reflective?) water/ice component
(since deciding that is, I think, the point of the experiment).
That said, if they want their predictions to be accessible to non-
scientists (and probably to other scientists as well) it would seem
helpful if they could supplement the raw photometric predictions with
some kind of comparison to more familiar terms of reference. For
example, a curve showing how bright the predicted plume is in
comparison the normal lunar surface would be useful. It would seem
much easier to plan observations if one knew, for example, that (at a
particular wavelength) the plume after 60 seconds is expected to have
(over an area of 10 km ?) twice the brightness of the nearby normal
lunar surface, or one-hundredth of it, or whatever.
Given such information you could easily make a test run with your
telescope adjusted to simulate those conditions (for example, by
stopping it down or using a neutral density filter if the brightness
is expected to be some fraction of the normal lunar surface
brightness), and seeing if you could detect anything under those
conditions. That would completely avoid any need for complex
calculations of the sort you are attempting, and the many additional
assumptions that are needed to complete them (e.g., is the curve
presented in the slide the intensity before or after attenuation by
the Earth's atmosphere?, how great are the additional losses in your
reflector?, etc.).
Hopefully at some point before the impact, the LCROSS science team
will give a more clear indication of how their predictions for the
size and brightness of the plume relate to some easily accessible
reference (such as normal lunar surface features); although based on
their participation in this forum to date I wouldn't hold my breath
for an answer.
As a rough guess, I believe the surface brightness of the Sun (at its
peak in the visible and as observed above the Earth's atmosphere) is
said to be about 26,000,000 W/m^2/micron/sr. The "Full Moon" (which
is not really the real Full Moon, but rather something obtained by
extrapolating the brightness from other phases) is said to be about
1/400,000th of that, giving an average Full Moon lunar surface
brightness of something like 66 W/m^2/micron/sr. At the First and
Last Quarters the surface brightness is said to be only about 1/5th of
this value, which would work out to something like 13 W/m^2/micron/sr
(the total light from a Quarter Moon is only about 1/10th of that from
a Full Moon because only half the surface shines with this reduced
brightness).
I haven't had a chance to download the AAS presentation you are
referring to:
http://lcross.arc.nasa.gov/observation.htm
but it sounds like the slide you mention is the one that shows a peak
plume brightness in the visible (10-60 seconds after impact) of 35-40
W/m^2/micron/sr. If my numbers given above for the "normal" Moon are
correct, this would seem to imply that the LCROSS scientists think the
initial impact plume will be considerably brighter than a typical
lunar surface feature at Quarter phase, and probably even brighter
than the maria (but not as bright as the brightest surface features)
in a nearly Full Moon. Of course, I have no idea if this is actually
what they are predicting or not.
I'm also not sure how large a diameter the peak surface brightness
quoted at 10-60 seconds after impact is supposed to extend over; and I
would assume that whatever the initial surface brightness is, the
plume's surface will, at some point, start to fade as it grows.
-- Jim
On Dec 2, 1:08 pm, Arnold Ashcraft <wa2...@optonline.net> wrote:
> Folks,
> I would like to image the impact plume from the LCROSS mission and I
> would like to use the correct exposure and focal ratio to maximize my
> chances of getting something. In slide 24 of the talk presented by
> Dr. Jennifer L. Heldmann at the May 2007 AAS conference is a plot of
> the expected impact plume irradiance as a function of wavelength and
> time. The units given for the irradiance is watts per meter squared
> per micron per steradian. I did a rough integration of the plot over
> the range of wavelengths my camera CCDs are sensitive to and got an
> area of about 20 watts per square meter per steradian.
>
> From the aperture and focal length of my telescope and the area of a
> pixel on my cameras, I calculated that if I placed the CCD of my
> DMK31 camera at the un-amplified focus of my 12.5" f/6 Newtonian, I
> will get 0.012 microwatts of power deposited in each 4.65 micron
> pixel. Doing the same calculation for my SBIG ST-8 and assuming 2x2
> binning of its 9 micron pixels, I get 0.18 microwatts in a 2x2 bin.
>
> That's where my calculations ran out of gas. I would like to relate
> this power level to the exposure I would need to get a perceptible
> image (say, 10x the noise level from dark current) with either of the
> cameras. If possible, I would like to take either a video or a time-
Arnold Ashcraft
Dec 3, 2008, 11:58:33 AM12/3/08
to lcross_ob...@googlegroups.com
Jim:
I did some more calculations, this time in a spreadsheet, not the back of an envelope, and found that I had dropped a few powers of ten in my earlier calculations, but more importantly, I was able to go from watts deposited in a pixel of my camera to electrons accumulated in the wells of the chip. This time I got 4.6e-12 watts per pixel for the DMK31 and 17.1e-12 watts per pixel for the 9 micron pixels of the ST-8 and 68.6e-12 for the 2x2 binned pixels using the low res mode of the camera.
I then looked up the Planck constant and figured out the energy of a 0.5 micron green photon. It turns out to be about 4e-19 joules. Since a watt is a joule per second I can now convert the watts per pixel into photons per second per pixel, and with some assumptions about quantum efficiency, to electrons per second per pixel and get some idea aboutexposure time needed to get some good signal levels.
My ST-8 pixels will hold at least 100000 electrons without saturating, according to the spec sheet Kodak supplies for the detector SBIG uses in the camera. I don't know how many will fit in the wells of the DMK31 pixels, however, assuming the same amount, I figure that at f/6 with the 12.5" reflector it takes only 0.17 sec to get that many electrons in the little 4.65 micron pixels of the DMK31, 0.05 sec for the ST-8 unbinned and only 0.01 sec for the 2x2 binned mode of the ST-8.
This suggests that at f/6 I could actually take videos with exposures in the 1/8th to 1/4 sec per frame range using the DMK31, and that the ST-8 would probably be overexposed at the shortest exposure time I can set for
that camera of 1/8th second. Putting on a 2x Barlow actually gives me f/14 because of the added length behind the Barlow compared with an eyepiece. At this focal ratio I calculate an exposure time of 0.25 sec will be about right in the high resolution (single pixel) mode of the ST-8.
This is an encouraging result which says I might actually be able to image the plume, assuming of course that the estimate of the plume brightness in the PowerPoint presentation is a good one. At the moment, there is not much else to go on, that and my probably flaky calculations. I may go back and do the watts to electrons per second calculation before integrating over the wavelength range. That would let me include the change in quantum efficiency with wavelength in a more satisfactory way.
Right now I am trying to decide which camera to put on what telescope. My ST-8 is more or less permanently mounted on the 12.5" Newtonian working at a focal ratio of f/14 for my double star observation program, and the above calculation suggests that an exposure of a 1/4 sec or so should record the plume. I think I will leave that setup intact. But I would like to get a time lapse movie as well. I have a 10" Meade SCT with a focal ratio of f/11. The DMK31 might give me a usable movie on this scope. I calculate 0.58 second per frame might work. If I put the camera on my 7.25" Schupmann, the smaller aperture requires an exposure of about a second per frame. I could set the webcam recording a minute before the impact event and run over to the Newt and see if I could manually get images with the ST-8.
Clif
Arnold Ashcraft
Dec 4, 2008, 4:55:27 PM12/4/08
to lcross_ob...@googlegroups.com
I got some more information on the well size of the DMK31 video camera.
Apparently its well is not nearly as deep as that of the ST-8, only holding
about 12000 electrons, with that amount being dependent upon the gain
setting as well, more capacity at lower gain. I worked this into my calculations
and found that if I use the 7.25" Schupmann at f/14, the pixels are saturated
by the anticipated impact plume at an exposure of only 1/88th of a second.
If I use some Barlow lens amplification, say to f/32, the exposure needed to
fill the well is 1/17th of a second. I would normally use an exposure between
1/100th of a second and 1/200th of a second to photograph the Moon itself
at this focal ratio (the image I just uploaded to the files section of the group
which I took on November 8 this year was taken at 1/137th of a second). At
this exposure (assuming the plume is as bright as predicted in slide 24 of Dr.
Heldmann's talk) the area of the plume would result in an image having
about 1500 electrons in the well. The noise level at low gain is about 25
electrons, so I would end up with a s/n ratio of about 60. This suggests that
I should just set my camera exposures to normal values for getting a well
exposed video of the moon limb and if there is a plume, I will record it.
With any luck, the plume will be snow and ice, not just dark moon dirt,
and will be even easier to see...
Now for a question: what are the chances that the moon will even be above
the horizon as seen from the east coast of the US when the impact occurs?
Clif Ashcraft
Jim Mosher
Dec 7, 2008, 11:09:11 AM12/7/08
to LCROSS_Observation
On Dec 4, 1:55 pm, Arnold Ashcraft <wa2...@optonline.net> wrote:
> Now for a question: what are the chances that the moon will even
> be above the horizon as seen from the east coast of the US when
> the impact occurs?
Clif,
> Now for a question: what are the chances that the moon will even
> be above the horizon as seen from the east coast of the US when
> the impact occurs?
Since the impact is supposed to be timed so it will be simultaneously
visible from Hawaii and Chile(?), I would think the chances of its
being visible from the US East Coast are fairly high, although the
chance it will be near the horizon are also high.
It would not be visible from the US East coast if the Moon is at too
southerly a latitude, which (I think) would mostly likely happen if
the impact occurs during the Fall or Spring, since the Earth's axis is
tipped most strongly towards and away from the First and Last Quarter
Moons at that time.
I have no idea when the impact is now expected to occur or whether it
will be at a First or Last Quarter phase, although apparently it will
be near one or the other.
-- Jim
Reply all
Reply to author
Forward
0 new messages