- On-chip differential photometry is well-suited to obtaining
lightcurves which record the time-varying brightness of an object,
in 12.410 usually a variable star or an asteroid.
It can also be used
to ``bootstrap'' the relative brightnesses of non-varying objects,
such as adjacent stars in an open cluster.
Note that for on-chip differential photometry,
observing can continue through thin clouds
at the cost of additional noise in the data.
-
For lightcurve projects
our likely observing conditions, available time,
and available instrumentation constrain which
amplitudes and periods are most suitable.
For starters you'll need to choose a target object whose
lightcurve variation is large enough to detect at Wallace
(0.1 magnitude or larger),
and whose
lightcurve period is short enough that you can get enough lightcurve
to meet the needs of your project
during one or two 4-hour observing sessions.
Watch out for longer periods commensurate
with the 24-hour Earth day,
or with the 7-day interval between weekly observing lab sessions.
-
For short-period variable stars such as dwarf cepheids,
choose a target whose
period is less than or equal to approximately 4.0 hr
so that you can record an entire cycle of lightcurve
in one observing session.
-
The observing program needed to record an asteroid lightcurve is similar
to that needed for a variable star, except for the additional
complication that the asteroid moves substantially each night.
As a result it takes more preparation to acquire the target with the telescope,
the availability of good on-chip comparison stars will vary,
and splicing together lightcurve fragments recorded on different nights
(if necessary) is more involved.
The apparent brightness of an asteroid depends in part on the cross-sectional
area it presents to an observer,
so as an irregularly-shaped asteroid rotates
it exhibits doubly-periodic brightness variations
(i.e. 2 possibly-unequal maxima and 2 possibly-unequal minima per period).
Recommended range of target asteroid rotation periods:
(P <= 5.7 hr) OR (6.3 hr <= P <= 7.4 hr)
-
In order to be able to perform on-chip differential photometry on
an image of your target,
there must also be at least 1 (more are better)
reference comparison stars ``on-chip'' in the field.
It's a good idea to confirm the availability of on-chip reference stars
of comparable brightness for the dates and times you plan to observe
before you go to the telescope.
Owing to CCD fields of view being small
you'll need to use
a catalog of relatively faint stars with known positions,
such as the ``HST Guide Star Catalog'' (GSC).
You'll need to know the field size of the instrument you're using.
Whenever possible it's best to try and get more than 1 comparison star
``on-chip'' with your target object,
so that you can intercompare them and check for variability.
Sometimes you can sneak in an extra comp star or two by
rotating your field.
On the other hand,
if your field of view is small enough that it contains no comp
stars brighter than your object,
then on-chip differential photometry by itself can still give
you a lightcurve but it's likely to be noisy.
because it'll incorporate the greater intrinsic noise of the fainter comp star.
In this case you'll want to identify a suitable brighter comp
star as nearby as possible, and spend the extra effort to also use it for
``near-chip'' differential photometry:
- If the night is ultimately fairly photmetric and you can get
a reasonable fit to the extinction as a function of time, then the
lightcurve with respect to the ``near-chip'' star will be your best
data.
- If the night is non-photometric, then you'll
have to go with the on-chip curve, extra noise and all.
Of course, if there is *no* on-chip comp star,
then you have no choice in the matter and must use
a ``near-chip'' strategy and hope for well-behaved sky conditions!
-
Another thing worth checking for when planning observations of
a moving object is whether
it'll have moved too close to
a background star at the time you want to observe,
a situation called an ``appulse''.
Appulses are bad because
when a star's image overlaps
and interferes with that of your target object
you won't get a good brightness measurement.
Workaround: if possible,
wait for your object to move away from the star
and then observe.
(Tinkering with a more
sophisticated image measurement scheme is
possible but beyond the scope of our class!)
-
When planning your observing program
you'll need to think about both
how frequently you'll need to sample the lightcurve
to carry out your project,
as well as
the total amount of lightcurve coverage you need.
The overall brightness of the asteroid,
the lightcurve amplitude, and
lightcurve period are all factors here.
Conflicting needs may require compromises.
-
``Keep your telescope out of the trees'' -
Try to observe your target objects at altitudes of at least 30 degrees
if possible,
but definitely don't bother observing lower than 20 degrees
altitude.
-
Keep objects of interest away from the edges of the frame,
because if they're too close to the edge you won't be able to get
brightness measurements.
-
If you're shooting ``discrete'' data points
(rather than continuous lightcurve)
shoot in triplets,
so that you can have some confidence that your measurements are
repeatable and have a way to estimate the experimental uncertainty.
-
Sometimes observations of a relatively faint object can be contaminated
by light from a nearby much brighter object,
both by scattered light and
by CCD ``blooming''
(charge-transfer streaks along columns resulting from overexposed pixels).
To work around this problem you'll need to try one or more
of the following:
-
excluding the bright object from the field of view
while still including your target faint object and reference stars,
-
rotating the instrument so that blooming on the image is directed away from
objects of interest,
- shooting a larger number of shorter exposures of the same field,
and then summing the images later.
(works if your target objects aren't moving too rapidly)