Transcript MARGE2005

An overview of the AEOS
Burst Camera
Heather Swan
University of Michigan
June 3, 2005
1
Outline
• Science Goals
• System design
• GRB Triggers/Response
• Grating analysis/Simulations
2
AEOS
Burst
Camera
Sensitivity
ROTSE-I/TAROT
Half of all GRBs have
no optical counterparts
 Could catch very fast
faders (short bursts?)
 High S/N for studying
variability or spectral
evolution
ROTSE-III
ABC w/ grating
ABC w/o grating
Keck
1 min. 10 min.
3
ABC field of view is well matched to the
Swift BAT error box
90% will be localized to
a 3 arc minute radius
(Can see them with the ABC)
50% will be localized within
12 seconds
(Can see them promptly)
(From Fenimore, et al)
4
How many GRBs do we expect to
observe in Maui?
• Swift promises 90/yr
• 1/3rd of the time it is dark in Maui
• 1/3rd of the time the GRB will have a high enough
elevation
• 3/5ths of the time the weather will be good
• 1/15th of the bursts can be observed, or 6 bursts/year
That’s approximately the same number we’re allowed to
observe per year (9)
5
The AEOS telescope is a large optical telescope
used by the Air Force
Advanced Electro-Optical
Systems Telescope (AEOS)
Largest ground based AF
optical telescope (3.67m)
Designed to track satellites,
can quickly (~20 sec) slew to
coordinates
ABC
Located in Haleakala, Hawaii,
at 10,000 ft
6
The AEOS Burst Camera (ABC) is attached
to the AEOS
• Optics designed by Carl
Akerlof
• Package designed by
Alan Schier
• Camera built by
Astronomical Research
Cameras
Field of view 6' x 6'
F/# 4.5
Focal length of 16m
7
AEOS telescope image reducer
design
8
ZEMAX point spread function
PSF gets bad near the edge of the FoV.
9
CCD camera specs
• E2V 2kx2k back-illuminated CCD
• Cooled to –40 F
• CCD readout time ~6 seconds
• Typical exposure length ~10 seconds
10
Improvements to the ABC image
quality
• The baffle
– Attached to the secondary, blocks stray light
• Improvements in alignment
11
The secondary baffle removes most of
the stray light
•Image of M1 without a baffle
•M1 with the baffle
12
13
The ABC took images of the Genesis probe,
just hours before it reached the earth!
Genesis during
separation
14
The ABC will try to observe GRBs within
minutes after they are localized
GRB
Swift
ABC Computers
(Modified ROTSE
Software)
User Interface
GCN
Burst
Filter
Fax
:::::::
:::::::
CD
15
We filter GRBs from the GCN to determine
if the ABC should go after them
We use the following criteria to determine if a GRB should be
observed in Maui:
1.
2.
3.
4.
5.
6.
7.
Less than 1 hour old
Localization should fit in our FoV
Sun should be far enough below horizon
Should be visible for at least an hour
Should be 20 degrees above horizon
Moon should not be too bright or too near
Should be 20 degrees away from galactic plane
If a GRB passes all these cuts, it is automatically faxed to the
AEOS control room (no humans required!)
16
ABC User Interface
For GRB fax alerts
Camera Status
Removes current images
from the queue, cancels current
set of exposures
Non-GRB observations
Thumbnail of last image
17
Observation Timeline
T=0 sec
T=15 sec
Receive GRB coordinates from fax and pager
AEOS operator terminates current task
T=30 sec
Operator moves telescope to GRB coordinates
T=45 sec
Operator moves trunnion mirror to ABC position
T=60 sec
Operator initiates data taking on ABC
T=10 min
Operator notifies team of action on GRB
T=5 hrs
T=12 hrs
Data taking concludes
Collected data transmitted to U of M
18
We can also send manual GRB alerts
• ROTSE-III finds an optical counterpart
– Average time between GRB localization and
ROTSE-III reporting afterglow, 30 minutes
• If It is still bright, and should be visible when Maui
can see it
– We send a fax w/coordinates and finding chart
• The ABC response won’t be prompt, but we
know there is an afterglow!
19
ABC analysis pipeline
Download images
from Maui
Correct images
with flat field and
dark.
Use SExtractor to
find objects
Match stars to
catalog
Calculate magnitudes
for each star
Find limiting
magnitudes
Create
lightcurves
Find spectra for
objects
And many other
things!
20
030329
GRB response
• 030329 – not prompt, but visible (t ~ 3 days)
• 030418 –not prompt, or visible (t ~ 8 days)
• 041006 – first image after 2.5 hours, bad
pointing
April 1, 2003
030418
No other “real” GRBs have been requested
•
(2 false HETE alarms, but the weather was
bad/mount was down)
April 25, 2003
21
041006: The pointing is off, how do we
fix it?
The problem was known
before this GRB
Here there be
Dragons!
It’s a problem that is not easy
to fix (hard to determine
what is wrong)
Now we send a finding chart,
and the operators use wiki
stars to get the correct
pointing offsets
We missed a GRB because
of this problem!
22
Test Burst response times
Date
Response Time
(min)
May 30
9
Good images!
May 30
1
Good images!
May 20
3
Very cloudy
May 17
7
Crowded/correct
pointing
May 16
7
Bad weather
When the weather is bad, and the operators are in
the room, sometimes the first images are taken
seconds after the fax arrives!
23
Grating analysis/simulations
• We installed a blazed transmission grating in Jan
2005
– 35 groves/mm, peak wavelength 640nm, blaze angle 2.2°
– ~5 cm from CCD
– R=/=8
• We’ve taken images of different types of objects
– Red/blue stars
– Quasars
• We can compare images to simulations
• Found limiting magnitudes for different exposure
lengths
24
Stars look like blackbodies
A star observed
with the ABC
Black body,
sun’s temperature
25
Can differentiate between red and blue stars
Ra
Dec
g-r
rmag
Red
126.00429
0.01915
1.4
18.1
Blue
126.01388
0.21709
0.2
15.2
Cooler temps
Hotter temps
26
We take a known spectrum of an object, and
cram it through our simulator
simulator
400
•
•
•
•
500
600
700
800
Wavelength(nm)
900
intensity
Flux density
SDSS spectrum
of a quasar
What we expect
to see with the
ABC
0
0th
order
50
100
150
200
Image offset (pixels)
1st
order
Start with spectrum from SDSS
Multiply by CCD and grating efficiencies
Use grating equations to see what happens to the light
Convolve with psf
27
Simulations look similar to actual data
intensity
What we expect
to see with the
ABC
0
0th
order
50
100
150
200
Image offset (pixels)
1st
order
• Quasars look “spiky”
What we actually see
with the ABC
This quasar has a z of 3.83, and an
Rmag of 18.53 (10 s image, taken
at twilight)
28
Higher orders look like what we would
expect
29
Limiting Magnitudes
Compare to SDSS images to find dimmest stars
(Gives a rough estimate of limiting mag)
Exposure Length
(s)
Limiting Magnitude
Oth order
1st order
10
17.4
19.1
15
17.7
19.3
Could be off- We have significant vignetting, and sparse
fields
Don’t have many fields to get limiting mags from
30
Conclusions
• ABC is running well, operators know what to do
• The pointing is off, but the operators know how to correct for that
• The ABC should be able to get images a few minutes after the
GRB is detected
• The spectral information is crucial for understanding the GRB
progenitor
Questions?
31