Click here if you are having trouble logging into the forums
Privacy Policy |
Please read our Terms
of Service | Signup and
Troubleshooting FAQ | Problems? PM a Red or a Green Gu.... uh, User
btieman
sage
Reged: 07/24/08
Posts: 481
|
|
I had the oppertunity (due to clouds!) last night to try and measure the precision of my CCD measurements. I have an "artificial star"--basically an LED flashlight with 5 different sized pin points poked through a metal faceplate typically used for collimation. I decided to try and use the "artificial star" to generate a light curve without scintillation to try and measure how repeatable my CCD's response to light is.
What a journey! First, the light is *way* too bright. I wanted to get 60s exposures similar to a typical exoplanet run for me. I didn't want to destroy the PSF of the "stars" so external filters in front of the flashlight were out--big dew problem last night! I finally wound up with cutting a couple of layers of brown paper bag to put under the faceplate behind the pinholes. That was still too bright--but very red so I put the V filter on the CCD. That allowed me to go 60s exposures without saturating any of the "stars" using my typical optical path for measureing transits...challenge 1 solved!
Challenge 2: Fixing the light leaks around the faceplate! There's a little seam around the faceplate that leaked light that streaked across the faceplate obscuring the stars. I tried several rounds of trying to tape up the seams. It was getting rediculous, every new layer of tape blocked some light, but there was always enough leaking through to confuse the tracking of the stars for aperture placement--especially since the tape kept popping lose over time! Finally I remembered one of the kids posters was on the wall with that putty stuff. I took a piece of putty and rolled it out and tried to jamb it into the gaps all the way around the seam. That helped a lot! A sample image--stretch to breaking--is below. The white ring around everything is my postcard dewshield.
Then I left the CCD acquiring all night as I would when measureing a transit. The "star" was only about 80 feet away...on to challenge 3!
Challenge 3: The "stars" are on a flashlight that is 5 LEDs on top of 3 tripple A batteries. Over the course of about 6 hours the light curve from individual stars makes a gentle slop that dims by about 0.5mag over the course of the evening--linear for the most part with a vew slight bumps. Fine, the slope is probably the battery draining over time and the slight bumps may be due to temp change changing the battery output. I don't have data to support that so next time I may have to hook up my Hobo data logger--see what a project these things become! Anyway, slope = not a big deal, right? I've got exoplanet transits that have > 1 mag swings in star light curves due to passing clouds--this should really be no different. Wrong! The light output from individual stars don't track! When I plot all 5 star light curves on the same graph, the trends look similar, but when I adjust the offsets to get them all to overlap--the slopes are different! So I open the flashlight and look at how the LEDs are wired up. They are in parallel. They all will see the same voltage but they light outputs as a function of voltage must differ--man this setup is sensitive!
I did pick the two stars with the closest matching slopes and measured the differential photometry from them. I'll follow this post with a plot of that light curve.
Brian
--------------------
CPC 1100
Orion 120st
Milburn Wedge
QSI 520i
Meade DSI Pro
|
btieman
sage
Reged: 07/24/08
Posts: 481
|
|
Here's the differential photometry plot from the two closest matching "stars":
--------------------
CPC 1100
Orion 120st
Milburn Wedge
QSI 520i
Meade DSI Pro
|
btieman
sage
Reged: 07/24/08
Posts: 481
|
|
And here's the plot of the 5 stars and their raw light curves. Notice how ROI1 and 2 have different slopes--this is true for all 5 "stars" 
--------------------
CPC 1100
Orion 120st
Milburn Wedge
QSI 520i
Meade DSI Pro
|
groz
Pooh-Bah
Reged: 03/14/07
Posts: 1071
Loc: Duncan, BC
|
|
I've got the same little flashlight 'artificial star', and now you got me thinking. It uses 3 of the little AAA batteries, so, that's about 4.5 volts. Wondering now just how it'll behave if I feed it either 3.3 or 5.0 from a well regulated supply. That _should_ take battery decay out of the loop.
Where I'm set up, it's not possible to get the artificial star into a good location for the telescope, I have to move the scope out to our front yard to get the distances required, but, that'll introduce street lights into the equation. but this does have me thinking about a few things, the first being 'can I do stable power to the unit'? For that, I do have a cute little power supply that gives 3.3 or 5v regulated, have to check,I may have an option on another one for 4.0 or 4.5 V too. Then the problem becomes, how to power it out in the weeds where the star would have to sit, either have to find a couple hundred feet of extension cord, or, plug into a neighbors wall outlet. Guess I could use one of my 12V jumper batteries and the 35w inverter.
|
groz
Pooh-Bah
Reged: 03/14/07
Posts: 1071
Loc: Duncan, BC
|
|
Quote:
Here's the differential photometry plot from the two closest matching "stars":
On the tail end of this one, scatter seems to be from 1.61 to 1.66, for a total of 40. rough approximation of sd will be 1/5 of that, so around 8. On a quick look on the surface, that suggests 8mmag sd on this test, which is substantially more than we see on 'real stars'.
Is that the led, or the battery ? Whatever it is, I cant help but start to wonder, is it even possible to get the measurement you want from this little gadget? or does the gadget have more variation than the sensor you are trying to calibrate ? Comparing this data to 'real stars' suggests this gadget isn't stable enough to be used as a calibration tool.
|
StarmanDan
sage
Reged: 08/27/07
Posts: 391
Loc: China Spring, Texas
|
|
Have you tried imaging a Landolt field? I believe these are used to calibrate photometric cameras as the stars in these fields have very well known photometric data.
--------------------
"Starman" Dan Doyle
Texas Astronomical Society of Dallas
Central Texas Astronomical Society
8" LX200GPS w/ST80 guidescope, Canon 350D+DSI Pro
150mm f/8 Sky Watcher Refractor
10" f/4.5 Homemade Dob
RV-6 Criterion Dynascope
http://darcstar.wordpress.com
|
btieman
sage
Reged: 07/24/08
Posts: 481
|
|
Quote:
Quote:
Here's the differential photometry plot from the two closest matching "stars":
On the tail end of this one, scatter seems to be from 1.61 to 1.66, for a total of 40. rough approximation of sd will be 1/5 of that, so around 8. On a quick look on the surface, that suggests 8mmag sd on this test, which is substantially more than we see on 'real stars'.
Is that the led, or the battery ? Whatever it is, I cant help but start to wonder, is it even possible to get the measurement you want from this little gadget? or does the gadget have more variation than the sensor you are trying to calibrate ? Comparing this data to 'real stars' suggests this gadget isn't stable enough to be used as a calibration tool.
I, too, noticed scatter seemed to improve throughout the night. Not sure why that would be. The scope was outside severl hours prior to starting this run so it should have thermally equiliberated. The dew heater was at equilibrium. The CCD is regulated. The light is floating out there in the yard so it may have been at it's coolest temp by then--the end of the run was about 5:30AM. I don't know enough about outpt stability of LEDs vs temperature to hazard much more of a guess than that. MMag are small!
I thought of putting a regulated power supply on as well--but it shouldn't be needed if I take C1-C2--a decreasing light output shouldn't look much different than dimming from a cloud. I've seen 1mag shifts in star light curves over the same period and have still pulled out differential light curves better than this one. C1-C2 is my direct measure to a transit light curve anyway.
I think the first thing to do is kill 4 of the LEDs. Then there's only one light source for all 5 stars--should help with all the paper I've got wadded up in there too  I'll have to open it up again, but hopefully it's eassy to desolder the LEDs so that I can put them back--This device does make it trivial to collimate the scope and the light output is perfect then!
The LEDs aren't aligned behind the holes, but the light out of each hole will be very much dominated by at most 2 LEDs. Kill 4 of them and there's only one source of light and hopefully the problem goes away.
LOL...boy do we sound nerdy or what 
Brian
--------------------
CPC 1100
Orion 120st
Milburn Wedge
QSI 520i
Meade DSI Pro
|
groz
Pooh-Bah
Reged: 03/14/07
Posts: 1071
Loc: Duncan, BC
|
|
Quote:
LOL...boy do we sound nerdy or what
We are playing with exoplanet light curve measurements, right up there with the 'big boys', with one little difference. They spend more on a calibration light source, than the entire cost of our complete kig. ofc, they also spend a lot of time writing proposals to 'beg for access' to that big expensive telescope, whereas we just haul it out onto the back deck, and start shooting.
Just the fact that we are looking at our potential for better calibration, and how to 'fix' a 20 dollar light source so it can be used as a calibration tool speaks volumes for what we are doing.
I cant shoot at the artificial star on the deck, deck isn't big enough to get it in a spot I can focus. But, I've been wondering about shooting at a street light that's half a mile away, then stacking a whole mess of filters in front of the camera. If I stack the entire BVRI set in the camera nosepiece, then point at a streetlight that's half a mile away, wonder how much light will get thru, and how stable of a light source that would be? To create the 'muliple stars' take a completely different tack, use the 5 pinhole cover out of the artificial star somewhere in the filter sandwhich, and run the whole system far enough out of focus that it ends up with 5 little dots on the ccd, all coming from the same streetlight in the distance....
|
groz
Pooh-Bah
Reged: 03/14/07
Posts: 1071
Loc: Duncan, BC
|
|
Quote:
Have you tried imaging a Landolt field? I believe these are used to calibrate photometric cameras as the stars in these fields have very well known photometric data.
I think what we really want to achieve is a stable light source, close enough that atmosphere is not a player in the grand total anymore. That would give a 'direct measurement' that defines equipment limitations. Shooting a standard field is still an indirect measurement, because all of the atmospheric stuff comes into play. Extinction, scintillation, etc. Measuring a nearby, stable source, essentially takes all the atmospherics out of the equation, and leaves a direct measurement of the optical system.
|
Hrundi
Pooh-Bah
Reged: 02/06/08
Posts: 1228
Loc: Estonia
|
|
How stable would a monitor LCD be? I imagine these things aren't prone to too much fluctuation? I could entirely be wrong though.
Gotta use the same spot of the LCD though.
Actually, if you were to cover the LCD with something that doesn't leak light, and cut through holes in various places, you might get different types of light sources as well. It all depends on the quality of the backlight.
--------------------
|
cflrich
member
Reged: 08/12/08
Posts: 22
|
|
Quote:
Measuring a nearby, stable source, essentially takes all the atmospherics out of the equation, and leaves a direct measurement of the optical system.
Does it? I would posit that this is the reduction in scatter you see during the night. What type of surface was being shot over? The earth can retain a lot of heat into the night, and this can have quite an effect as it rises up into the night, if you're looking across over it.
--------------------
Rich
Rich's Astro Blog-- http://cflastro.blogspot.com
|
StupendousMan
member
Reged: 08/21/05
Posts: 95
|
|
Quote:
They spend more on a calibration light source, than the entire cost of our complete kig.
Regulated power supply -- maybe $200.
LED - $5.
So far, not so bad. But a good integrating sphere and the optics to get the light from the LED to the detector properly -- that could run three or four thousand dollars easily, it's true.
But to do the sort of tests you have been discussing really does require only a regulated power supply and a few LEDs. You can make pinholes and rig up simple ways to get around the expensive optics -- using a long hallway in a school building at night, for example.
Go for it.
|
btieman
sage
Reged: 07/24/08
Posts: 481
|
|
Quote:
...a good integrating sphere and the optics to get the light from the LED to the detector properly -- that could run three or four thousand dollars easily, it's true.
I have a an integrating sphere with a 1" diameter output port, a liquid light guide and a programable calibrated light source with numerous filters in a lab at work. I also have a lasing diode and the electronics to drive it at a few MHz rep rate. I'd never consider bringing that stuff home but I have considered bringing the camera into work and doing some proper calibrations/characterizations there.
Sorta takes all the fun out of it though...there's something almost theraputic to McGyvering this stuff up. Besides, I think taking advantage of that sort of equipment puts what I'm doing out of reach of the typical amature.
Groz is right, the goal right now is to come up with a simple measurement that puts an upper end on the limit of the equipment. A little heat off the backyard grass is nothing compared to an atmosphere or two so I'll live with it. The school hallway is a great idea! My wife's office has a hallway that's probably about as long as I'm using now and I can get in there at night! The emergency lights might be a problem though.
--------------------
CPC 1100
Orion 120st
Milburn Wedge
QSI 520i
Meade DSI Pro
|
brianb11213
Carpal Tunnel
Reged: 02/25/09
Posts: 2076
Loc: 55.215N 6.554W
|
|
If I remember my basic electronics right, there is a "golden" value for a zener diode - around 5.6V to 6.8V - where the voltage is pretty well independent of temperatures (in the normal working range).
So, 9V battery, 6.8V zener, load resistor & a voltage divider (using constantan resistance wire) should make a very stable LED source, until the battery voltage falls too low.
|
|
1 registered and 1 anonymous users are browsing this forum.
Moderator: RLTYS
Print Thread
|
Forum Permissions
You cannot start new topics
You cannot reply to topics
HTML is disabled
UBBCode is enabled
|
Thread views: 221
|
|
|
|
|
|
|
Previous
Index
Next
Threaded
Edit
Reply
Quote
[Re: 