Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.stsci.edu/ftp/instrument_news/FOC/Foc_tips/tips_22aug96.ps Äàòà èçìåíåíèÿ: Wed Apr 2 02:37:23 1997 Äàòà èíäåêñèðîâàíèÿ: Sun Dec 23 06:55:58 2007 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: emission nebula

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
Summary
of
FOC
Activities
for
the
period
18
July
-
22
August
1996
Operations The
FOC
was
successfully
recovered
from
the
HST
safe
mode
The
F/96
relay
has
been
operating
nominally.
The
F/48
relay
was
again
successfully
used
for
external
observations.
Proposals
Executed
There
were
4
GTO/GO
proposals
and
1
CAL
proposal
executed
using
the
FOC:
6746
Dynamics
of
the
Circumnuclear
Disk
of
M87
(GTO)
6349
Snap
Shot
Search
for
High
Redshift
QSOs
with
Far­UV
Flux
(GO/SNAP)
6775
Secular
Changes
in
the
Jet
of
M87
(GO)
6161
FOC
Cycle
5
Aperture
Location
Monitor
(CAL)
6304
Long­Slit
Spectroscopy
of
the
Nuclear
Regions
of
a
Sample
of
Radio
Galaxies
(GTO)

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
Summary
of
FOC
Activities
for
the
period
18
July
-
22
August
1996
Exposure
breakdown:
F/48
F/96
SMS
Proposal
Internal
External
Internal
External
Total
96.204
6746
6
7
13
6349
8
8
96.211
6775
10
14
24
96.218
6775
5
8
13
6161
1
3
4
96.225
6304
6
4
10
Total
12
11
16
33
72

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
FOC
Special
Topic:
Geometric
Distortion
in
the
F/48
Camera
Since
the
Faint
Object
Camera
uses
electromagnetically
focussed
electron
beams
in
the
cameras,
it
suffers
from
geometric
distortion.
The
distortion
for
the
F/48
camera
looks
like
this:
200
400
600
800
1000
200
400
600
800
1000
PIXEL
NUMBERS
LINE
NUMBERS
diff
file
:
dist75.tab
run
entry
:
day835
,
dated
Tue
13:50:59
05­Apr­94
,
history*
V
*
ref
file
:
optref.tab
ref
entry
:
optref
,
dated
Tue
13:50:59
05­Apr­94
,
history*
V
*
magnif.
:
2.
plotted
:
Tue
16:34:47
05­Apr­94

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
FOC
Special
Topic:
Geometric
Distortion
in
the
F/48
Camera
Geometric
Distortion
affects
F/48
observations
in
two
fundamental
ways:
.
Since
interactive
target
acquisitions
use
the
raw
(distorted)
data,
this
introduces
a
position­dependent
error
into
the
calculation
of
the
slew
required
to
move
the
target
to
the
slit.
Since
the
slit
is
only
0.062
arcsec
wide
(2.25
pix
els),
even
small
errors
of
a
few
percent
over
a
100
pixel
(2.8'')
slew
can
make
the
difference
between
the
object
being
successfully
acquired
and
missing
the
slit
completely .
.
The
output
data
from
the
long­slit
spectrograph
is
distorted
in
such
a
way
that
the
relationship
between
(wavelength,
position)
and
(sample,
line)
is
complex,
diffi­
cult
to
invert,
difficult
to
calibrate,
and
time­dependent.
This
makes
data
reduc­
tion
a
pain
in
the
neck
complex
procedure.

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
FOC
Special
Topic:
Geometric
Distortion
in
the
F/48
Camera
Geometric
Distortion
in
Target
Acquisition
Here
is
an
(exaggerated)
example
of
how
geometric
distortion
affects
target
acquisi­
tion: While
the
best
OPUS
solution
attempts
to
put
the
target
at
the
fiducial
point
(506,
512),
the
target
is
actually
moved
to
(522,
509),
an
error
of
16
pix
els
(0.46
arcsec).
And
you
can't
get
around
this
by
just
improving
the
OPUS
solution;
the
nonlinear

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
FOC
Special
Topic:
Geometric
Distortion
in
the
F/48
Camera
nature
of
the
distortion
means
that
what
works
at
one
position
does
not
work
else­
where. To
do
this
properly,
we
need
a
geometric
distortion
model.
This
can
be
obtained
in
two
ways:
.
Using
the
grid
of
reseau
marks
etched
onto
the
FOC
photocathode
(they
are
exactly
1.5mm
apart)
.
Using
the
method
developed
by
Perry
Greenfield,
whereby
several
images
of
a
rich
star
field
are
taken,
each
offset
by
a
known
amount.
Using
the
fact
that
the
position
offset
between
the
fields
is
the
same
f
or
all
the
stars
in
the
field
allows
reconstruction
of
the
geometric
distortion
Note
that
in
the
first
method,
it' s
necessary
to
combine
the
camera
distortion
with
the
optical
distortion
that
arises
from
the
off­axis
position
of
the
FOC
apertures
in
the
OTA
field
of
view.
In
the
second
method,
the
total
(camera+optical)
distortion
is
calcu­
lated.

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
FOC
Special
Topic:
Geometric
Distortion
in
the
F/48
Camera
This
is
the
formalism
for
method
2
(see
FOC
ISR
#85):
The
solution
to
the
least
square
problem
becomes
the
solution
to
the
follo
wing
matrix
equation:
d
1q
r
(
)
0
.
.
.
0
S
--
1
r
(
)d
1q
r
(
)
.
.
.
S
--
L
r
(
)d
1q
r
(
)
0
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
0
.
.
0
.
.
.
0
d
Qq
r
(
)
S
--
1
r
(
)d
Qq
r
(
)
.
.
.
S
--
L
r
(
)d
Qq
r
(
)
S
--
1
r
(
)d
1
.
.
.
.
S
--
1
r
(
)d
Q
S
1
r
(
)S
1
r
(
)
.
.
.
S
1
r
(
)S
L
r
(
)
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
S
--
L
r
(
)d
1
.
.
.
.
S
--
L
r
(
)d
Q
S
L
r
(
)S
1
r
(
)
.
.
.
S
L
r
(
)S
L
r
(
)
r
1
=
R Å
Õ
Ü
Ã
Â
Ã
Â
Ã
Â
Ã
Â
Ã
Â
Ã
Â
Ã
Â
Ã
Â
Ã
Â
Ã
Â
Ã
Â
Ã
Â
Ã
Â
Ã
Â
Ã
Â
Ã
Â
Ã
Â
Ô
Æ
u
1 . . .
u
Q
c
1 . . .
c
L
Du
r
d
1q
r
(
)
. . .
Du
r
d
Qq
r
(
)
--Du
r
S
1
. . .
--Du
r
S
L
r
1
=
R Å
=

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
FOC
Special
Topic:
Geometric
Distortion
in
the
F/48
Camera
To
calculate
the
distortion,
we
used
method
2.
Images
were
taken
of
the
center
of
the
globular
cluster
47
T
ucanae
using
the
256x1024(z)
format:
and
a
full­format
image
of
the
field
rounded
off
the
dataset.
By
matching
up
stars
that
appear
in
more
than
1
image,
it
is
possible
to
calculate
the
geometric
distortion
model
for
the
256x1024(z)
format
that
makes
the
star
position
offsets
invariant.

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
FOC
Special
Topic:
Geometric
Distortion
in
the
F/48
Camera
The
geometric
distortion
model
is
calculated
as
a
bicubic
spline ,
which
can
then
be
used
to
calculate
the
actual
slew
required
to
move
a
target
from
1
pixel
location
to
another
in
the
raw,
distorted
images.
This
method
is
still
new
and
not
completely
tested
yet,
but
we
plan
to
use
it
for
subse­
quent
F/48
target
acquisitions.
One
other
thing
to
remember:
Sometimes
images
are
shifted
due
to
presence
of
a
filter
(i.e .
if
it's
not
quite
square­
on
in
the
beam).
It's
VERY
important
to
know
what
these
filter­induced
shifts
are.
For
example,
the
F/96
F486N
filter
shifts
the
image
b y
17
pixels!
What
little
F/48
data
we
have
shows
that
any
filter
shifts
are
small,
<1
pix
el.

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
FOC
Special
Topic:
Geometric
Distortion
in
the
F/48
Camera
Geometric
Correction
of
F/48
Spectr
oscopic
Data
What
we
have:
What
we
want:
Wavelength
is
a
complex
function
of
Wavelength
is
proportional
to
y
pixel
[i,
j]
Objects
trace
a
curved
path
as
a
Objects
all
parallel
vertical
lines,
with
function
of
wavelength
separation
proportional
to
Dx
3600å
5400å
12.5''
Hb
Hb
5400å 3600å

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
FOC
Special
Topic:
Geometric
Distortion
in
the
F/48
Camera
Calibration
involves
calculating
the
functions
l(i,j)
and
z(i,j),
and
rebinning
the
data
so
that
the
resultant
data
conforms
to
``What
we
want''.
Since
the
FOC
has
no
internal
line
source,
determining
the
wavelength
calibration
involves
looking
at
celestial
emission­line
sources.
An
ideal
source
.
is
large
enough
to
fill
the
slit
(>12.5'')
.
has
high
surface
brightness
.
has
no
velocity
structure
at
the
few
tens
of
km/s
level
(FOC
pixels
are
100­
150km/s)

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
FOC
Special
Topic:
Geometric
Distortion
in
the
F/48
Camera
Such
a
source
is
the
Orion
nebula
(away
from
the
Trapezium),
which
was
observed
during
Science
Verification:

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
FOC
Special
Topic:
Geometric
Distortion
in
the
F/48
Camera
We
can
use
an
internal
flatfield
image
(with
well­exposed
reseau
marks)
to
measure
and
remove
the
camera
distortion:

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
FOC
Special
Topic:
Geometric
Distortion
in
the
F/48
Camera
We
can
then
use
the
taks
in
the
noao.twodspec.longslit
package
to
measure
and
remove
the
distortion
of
the
spectral
lines,
and
to
perform
the
wavelength
calibra­
tion: This
gives
us
l(i,j).

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
FOC
Special
Topic:
Geometric
Distortion
in
the
F/48
Camera
But
we
still
need
z(i,j).
For
this,
we
need
a
spectrum
of
a
point
source.
We
have
data
for
M87
taken
for
a
GO
science
program:
Using
tasks
in
the
noao.twodspec.apextract
package,
we
can
trace
the
point
source
and
calculate
z(i,j)
This
diagram
is
temporarily unavailable

SPACE
TELESCOPE
SCIENCE
INSTITUTE
TIPS
Meeting
22
August
1996
R
Jedrzejewski
FOC
Special
Topic:
Geometric
Distortion
in
the
F/48
Camera
Since
the
F/48
camera
distortion
varies,
correction
of
spectra
to
the
camera
system
is
important,
since
we
know
the
transformation
from
the
camera
system
to
(l,z).
So
for
any
data
set,
once
we
correct
the
camera
distortion,
we
can
subsequently
convert
to
``what
we
want''.
So
far
we
have
used
imperfect
data
(pre­COSTAR,
faint
point
sources).
The
Cycle
5
F/48
calibration
program
will
provide
new
spectra
of
an
extended
high­surface
bright­
ness
emission­line
source
(the
planetary
nebula
NGC
6543)
and
a
spectrophotomet­
ric
standard
star
LDS749B
(the
FAINTEST
IUE
standard).
These
data
will
allow
us
to
refine
our
geometric
correction
files
to
the
goal
of
0.3
pix
els
r.m.s.
error.