Makee: Keck ESI Observing Strategies
Last update for this page is March 2001.
Approximate Throughput Curve:
This is a rough estimate of the total fractional throughput
for Keck plus ESI assuming a 10 meter diameter aperture:
and in ASCII x,y format:
, as a function of wavelength in Angstroms.
Note that overlapping orders have been added together.
This was derived using the G191B2B standard star and the
atmospheric absorption features have not been corrected.
Note that this does not account for losses due to light falling
out of the slit (1.0 arcsec slit, so a 5% to 30% loss) or
for possible cloud cover.
Binning, Gain, and Dual Amp Readout:
With ESI, binning 1x1 and a dual amp readout is usually used.
The advantages to a larger binning would be:
(1) lower readout noise per binned pixel
(but the sky Poisson noise dominates the background noise
for most of the orders);
(2) faster readout (but it is already fairly quick, about 30 seconds);
(3) smaller files (but disk space is getting cheaper and faster).
So, I recommend you stick with 1x1 binning.
Note that the November 2000 version of
MAKEE only works with 1x1 binning for ESI,
however, a newer version of MAKEE now works with 1x2 binning (binned twice in
the dispersion direction).
For faint objects, use the high gain setting (fewer electrons per
digital number). This will sample the readout noise better,
although it limits your dynamic range somewhat. Low gain is
sometimes used for very bright objects. However, you should probably
stick to the same gain during your run.
For ESI, the electrons per digital number (eperdn) is set
according to ./makee/eperdnESI_higain.dat (eperdn=0.50)
or ./makee/eperdnESI_logain.dat (eperdn=1.29). The readout noise
is set according to ./makee/ronoiseESI.dat (ronoise=2.7 electrons).
You can override these values by editting these files, or
giving "eperdn=" or "ronoise=" on the MAKEE command line.
Flat field (quartz lamp):
Ideally the total quartz lamp counts at any position in your spectral
range should be much higher than the count rates in your object spectrum.
This may require several exposures to be taken for each setup.
Usually you will take exposures of 2-6 seconds (depending on slit width)
each with the internal flat in echellete mode.
Unfortunately, you get relatively few counts in the bluest echelle orders.
You may want to saturate the reddest orders on a few additional exposures
to try to get enough counts in the blue, however, as of now (November 2000),
there are no MAKEE programs (yet) to mask out the saturated orders.
Fringing is significant with ESI, but fortunately the fringes do not
move much with elevation thanks to a compensating system. One set of
flats for each slit width per observing run should be adequate, although
you may want to take these at a telescope elevation similar to most of your
observations, rather than in the telescope stowed (elevation zero) position.
HgNe+Xe and CuAr (arc lamps):
You must take at least one HgNe+Xe (about 4s) or CuAr (about 300s)
exposure during your observing run. It is recommended that you take
arclamps a few times during the night, perhaps for each object.
An arclamp for each slit width may also be a good idea, although probably
MAKEE does the wavelength calibration by using a template calibration
and determining a pixel shift as a function of column number. This
pixel shift function may depend upon telescope elevation (and Position
Angle?) with amplitudes on the order of half a pixel.
Note that a Xe exposure will work in place of a HgNe+Xe, or if
you take separate Xe and HgNe exposures you can add the raw exposure
files together using opim .
The night sky lines are used for an overall pixel shift for each
object exposure, but the column depedency requires an arclamp
CuAr has more lines than Hg+Xe, but the most of the lines occur in
different orders. Taking both types is optimal, although one of
either is probably adequate. The accuracy of the wavelength calibration
for ESI is about +/-0.2 pixels (1x1 binning).
You can take Hg and Xe individually, although this is probably not
necessary. In any case, MAKEE can handle only two ESI arclamps (total)
at a time.
Bias (short dark):
It is probably not necessary to take a short dark during your run.
By default, MAKEE uses an old template pattern for the short dark or
bias pattern. However, it is still a good idea to take a couple
short darks during your run, even if you don't plan to use it in your
reductions (just in case the bias pattern changes).
The long dark pattern is believed to be negligible.
You need at least one bright point source exposure in order to define
the echelle order pattern on the CCD. If you have a bright object
with no large absorption lines, you can use the object itself to
define the trace function. Otherwise, you should take a bright
star exposure for each setup at some time during the run. This need
not be next to the object exposure.
You can also use a pinhole quartz exposure to define the trace function.
This avoids having to take a star exposure and should may work just as well
as a star.
However, a bright star is still recommended-- bright blue stars are
best in order to define to trace in the bluest orders. See the
recommended standard star list:
Standard Star List .
Accurate flux calibration is generally difficult with echelle
spectrographs. With HIRES, it appears to be very difficult.
With ESI, I do not have enough information at this time (November 2000)
to know how difficult flux calibration might be. Hopefully, there
will be a MAKEE routine for flux ESI flux calibration in the not too
Here is a list of "good" ESI standard stars:
standard star list .
A 100 to 200 second exposure is usually adequate for these stars.
For more on flux calibration, see:
Back to makee home page.