Subject: ECHELLE Review Report

From: elt@astro.Princeton.EDU

Submitted: Thu, 22 Oct 1998 14:46:17 -0400 (EDT)

Message number: 2 (previous: 1, next: 3 up: Index)

Attached please find a copy of the report of the 17 September 1998
review of the 3.5-meter Echelle spectrograph.  Please excuse the
archaic troff typesetting inserts; I think it is still quite readable
in this form.

I am pleased to report that progress on the "aggressive and ambitious
plan for delivery and installation" described in this report is good
at this writing and that installation in November still appears likely.

Ed Turner


A group consisting of Bruce Gillespie (APO), Ed Jenkins (PU), Chris
Stubbs (UW) and Ed Turner (PU, chair) visited the Echelle spectrograph
group at the University of Chicago Department of Astronomy 
on 17 September 1998 in order to
assess the state of the instrument and to make plans for its installation
and commissioning at the APO 3.5-meter.  The APO group spent a full day
receiving presentations from and in discussions with the Echelle team,
which is led by Roger Hildebrand, and with other members of the University
of Chicago administration and Astronomy Department.  These included
D. Oxtoby (Dean), M. Turner (Dept. Chair), D. Duncan, L. Hobbs, D. York,
J. Gee, G. Kelderhouse, L. Fiscelli, S. Heimsath, B. Lynch, R. Lowenstein,
D. Sandford, J. Sundwall and J. Thorburn.  S. Wang, who will lead the 
Echelle APO
installation team, could not attend but has discussed related
matters with some of us on various other occasions.
This report gives a brief summary of the findings and conclusions resulting
from the visit and review.  Copies of transparencies and other material
provided to the group are appended to this report and provide many more
technical details.
We, the APO group, found this exercise to be very informative and helpful;
we thank our University of Chicago hosts for both their hospitality and their
thorough and clear presentations.
The Echelle was in stable laboratory operation and could routinely produce
good quality spectra of both lamps (line and continua) and solar light.
With two exceptions (see below), instrument performance was nominal and
entirely satisfactory.  The wide wavelength coverage at high dispersion
and with uniform resolution and line profiles was particularly impressive.
In several respects (notably wavelength coverage per exposure), performance
exceeds the original design goals, and it can be expected that the 
instrument will
be very powerful scientifically when it is fully operational at APO.
As of 17 September, the Echelle was being operated with an engineering grade
CCD chip (high readout noise) and via a temporary lab setup of camera
electronics and data system.  Completion of the Echelle requires installation
of the science grade chip and replacement of the test set electronics and
computers with the "flight hardware" versions.  Installation of the Echelle
on the telescope also requires completion of a new support structure,
fabrication of which is well underway.
In addition, one major functional capability had not yet been implemented,
namely guiding.  A guide camera and optics had been installed, but no
supporting electronics or software was yet available.
Echelle performance in the lab fell significantly below design specs 
in one important
respect.  The unresolved line shapes produced by the instrument have a FWHM
of 2.6 pixels, as opposed to the expected 2.0, and are moderately asymmetric.
Extensive attempts to locate and correct the problem during the past few
months have been unsuccessful.  Although a few possible correctable causes
remained to be investigated, the Echelle team felt that it might not be
possible to achieve any better performance.  The line widths and shapes
are quite stable over the entire spectrum, thus making it practical to
take account of the line PSF correctly during data reduction.
[Approximately one month after the 17 September visit, the line asymmetry
is reported to be substantially improved due to further work on internal
spectrograph optical alignments.]
In addition to the known problem with optical performance, the instrument 
throughput remains to be determined.  This quantity is very difficult to
measure in the laboratory but will be relatively easy to measure "on the
sky" via observations of standard stars once the instrument is installed.
Although there is no particular reason to expect throughput problems, it
is always a topic of concern for devices with complex optics involving
many individual optical elements.  The design goal is a total spectrograph
throughput of 9% or better at all wavelengths above 4100 A.  This would
be excellent performance for an Echelle spectrograph.
An aggressive and ambitious plan for delivery and installation was adopted.
It calls for immediate installation of the science grade CCD and replacement
of the engineering electronics and data systems with the final "flight
hardware" versions, followed by fairly quick (perhaps 10 days or so)
laboratory evaluation of the reconstituted instrument.  Assuming no unexpected
problems are uncovered, the shipping procedure would then commence with
the goal of carrying out installation of the instrument on the telescope
during the early November 1998 bright time.  This timeline also calls for
completion of the new support structure and its shipment to APO in roughly
mid-October.  Finally, there would also be a parallel and fast paced software
development effort aimed at allowing instrument engineering and commissioning
operations at the completion of installation.
It is clear that this plan is optimistic and even somewhat risky (in the
sense that subtle problems might not be uncovered during the brief
laboratory evaluation), but it was felt to be justified by several factors,
including a "window of opportunity" in the schedules of key personnel needed
to carry out the work, the oncoming winter weather at APO (making work in
the unheated dome difficult) and, of course, the desire to achieve science
operations of the long delayed instrument at the earliest possible date.
Nevertheless, it was generally agreed that the Echelle should not be
shipped to APO until the following five goals have been achieved:
.IP -
All "flight hardware", including the science grade CCD and the operational
camera electronics and data systems installed and tested in the lab.
.IP -
Instrument performance re-characterized, at the same level of detail as now
available for the engineering incarnation, and found to be essentially as
good or better.
.IP -
Development of a software system and interface sufficient to allow
engineering and science operation of the instrument on the telescope
(though this is unlikely to be the "final" version of the instrument
.IP -
Preparation of user and engineering documentation, or commitment to
preparation of a specific list of such documents, sufficient to allow
APO maintenance and operation of the instrument.
.IP -
Implementation of at least a rudimentary guiding capability (none appears
to exist currently) which will allow commissioning and early science
A specific commissioning plan (describing tasks to be accomplished,
quantities to be measured, acceptable performance, etc.) yet needs 
to be developed.  
At present, there is a general plan to carry out
a relatively quick first phase of commissioning via a few nights of
observations (probably of standard objects of various sorts and of
calibrations) immediately following installation.  This should reveal
any gross malfunctions or problems immediately.  The "rule of thumb"
criterion for 
acceptable performance is that the instrument should 
perform as well as the KPNO 4-meter Echelle, which has been scientifically
productive for many years.  This notion was suggested by Don York and
met general approval but is not intended to replace a more definite
and quantitative characterization of instrument performance and success.
In fact, substantially better performance than the KPNO 4-meter is
Following this initial commissioning phase, a more elaborate and extended
commissioning phase is expected to blend into early science operations.
This work will be carried out to a large extent by University of Chicago
builders and users of the Echelle but could also profitably involve other
ARC users on a "shared risk" basis.  If the best projected schedule is
achieved, such shared risk science operations could begin during 1Q1999.
The discussions on 17 September revealed three possible enhancements/additions
to the Echelle which will be sufficiently useful and easily accomplished that
they are likely to be very early upgrade priorities.  To be explicit, these
are not items which are felt to be necessary before the instrument is accepted
by APO as delivered, but it does seem likely that they will be among the
next priorities for work on the instrument.  They are enumerated below:
.IP -
The fundamental design of the instrument projects the spectral dispersion
direction onto the sky in a direction perpendicular to the horizon.  This
means that differential refraction will either reduce spectral resolution
or coverage (with a wide or narrow, respectively, aperture in that direction).
For example, according to calculations by Chris Stubbs, spectral coverage
(a primary strength of this Echelle) will be degraded by about 30% at an
air mass of 1.5 and by about 50% at 1.7 with the standard 1.6" slit.  This
problem can be alleviated by introduction of an ADC (Atmospheric Dispersion
Corrector) into the pre-slit optics.
.IP -
Initial guiding capability may be as primitive as manual "paddle guiding"
based on watching for the target to "peep out" of the slit.  This will be
inefficient in terms of human resources and may be difficult for remote
observers (due to net speed problems).  APO OS's will not carry out this
function routinely.  Thus, autoguiding capability will be an early priority,
and capability for on-slit guiding via a pellicle beam-splitter would be
highly desirable as it would avoid unnecessary loss of spectral resolution
during good image quality conditions.
.IP -
Laboratory tests show that the unexplained image (line shape) degradation,
discussed above under "Performance Issues", disappears when the slit is
replaced by a pin hole source.  This implies that a pre-slit field stop
device could be used to attain better spectrograph optical performance
(up to design specs) for sources sufficiently bright (or integrations
sufficiently long).  This will surely be desirable for some applications.
Doug Duncan and Don York, both expecting to be frequent users of the
Echelle, will work jointly as its co-instrument-scientists and will thus
take leadership roles in post-commissioning development and maintenance.
The following points were noted during the review and are recorded here
in no particular order.
.IP -
Surface errors in the Echelle grating itself are one possible source of the
poor optical performance discussed above.  Ed Jenkins pointed out that this
possibility could be checked via fairly simple and straightforward measurements
(by Ball Brothers, perhaps) which could be made while the instrument was
dis-assembled for shipping, among other possibilities.  If this were confirmed
as the problem, much pointless and time consuming work on optical alignment,
etc. could be avoided.  Also, modification of the grating support structure
might correct or reduce the problem.
.IP -
Previous experience with other instruments suggests that it would be useful
to retain capability to operate the spectrograph via "stand alone" software
independent of the REMARK interface, although this mode might not be
available to general users.
.IP -
The APO installation team, to be led by Shu-I Wang, is yet to be definitely
defined.  Who will be there and when, etc.
.IP -
Details of the post-commissioning maintenance arrangements between APO and
Chicago need to be worked out.
.IP -
A specific/formal administrative process by which the instrument is accepted,
with ownership and responsibility transferred from Chicago to APO, needs to
be devised.  (Tom Lutz proposed a general one for APO use many years ago, but
it has never been used.)
The exercised described in this report was considered to be a valuable one
by all concerned and may well be considered a useful precedent for future
APO 3.5-meter instrument projects.
.sp 0.2i
Submitted by
B. Gillespie
E. Jenkins
C. Stubbs
E. Turner (chair)

APO APO APO APO APO  Apache Point Observatory 3.5m  APO APO APO
APO  This is message 2 in the apo35-echelle archive. You can find
APO  the archive on
APO  To join/leave the list, send mail to
APO  To post a message, mail it to