Having learned from these
experiences, we are building the GMOS-North Hamamatsu CCD system with
the new video controller
boards to avoid the banding problem entirely.
available. After significant lab testing, we
verified that these boards would solve the
problem, but required that we modify our
software to be compatible with them. As of
August 25th, we have fully installed and integrated these new boards and eliminated
the banding effect.
The screenshots shown here (Figure 3) present examples of standard star fields with
bright stars using 2 x 2 binning before and
after the board replacement. In the left image, the banding effect is visible as dark
lanes on the saturated stars. In the right
image, taken after the fix, the effect is completely eliminated and saturated stars do
not produce the dark lanes.
Since the controller has changed, we are
now re-characterizing the entire detector
system, including re-measuring gains, read
noise, and full well for all modes. Meanwhile,
GMOS-S has resumed normal operation. We
found the primary science mode read noise
to be unchanged at ~ 4e-. A nice additional
benefit of this upgrade is that the full well
increased by ~10% with respect to the previous value.
Before (left) and after
(right) data from
GMOS-S showing
the banding present
before the fix (left)
and a different field
with bright stars after
the fix (right) with no
banding present.
July 2015
GRACES:
More Efficient than Ever
The Gemini Remote Access to CFHT’s ESPaDOnS Spectrograph (GRACES) began its
first scheduled observing runs in 2015B.
(See page 58 for a more complete overview
of the GRACES story.) The team made several improvements that make GRACES both
more efficient and easier to operate. To improve efficiency, the team recoated some
key optics and redesigned one element that
was vignetting the optical path). As a result,
GRACES’s efficiency is improved by an average of roughly 20% towards the red part of
the spectrum (see Figures 4 and 5).
A second problem involving a charge transfer smearing effect, most noticeable during
long nod and shuffle sequences, has gone
away on its own. We have made several efforts to reproduce the problem, but the effects have remained elusive, so we continue
to monitor. However, the bottom-line result
is that the Hamamatsu CCDs in GMOS-S are
now operating at full capacity and are ready
to deliver high-quality data to our users.
January 2016
Figure 3.
2015 Year in Review
Figure 4.
Comparison of the
GRACES sensitivity as
measured for the 2-fiber
spectroscopic mode in
May 2014 (dashed line)
and in June 2015 (solid
line). The sensitivity is
defined as the magnitude
of an object that would
provide a signal-to-noise
ratio of 1 after an hour
of integration time.
This figure illustrates
GRACES performances
improvements in the red
part of the spectrum.
Figure 5.
Greg Barrick (CanadaFrance-Hawaii Telescope)
installing the GRACES
receiver module in the
ESPaDOnS spectrograph,
incorporating some of
the recently improved
optics.
GeminiFocus
33