Previous Article
Next Article
Table of Contents

Active Optics on the CTIO 4-m (1Dec94)
(from CTIO, NOAO Newsletter No. 40, 1 December 1994)

The new active optics system on the CTIO 4-m telescope came into routine
use on 21 October. The active optics system includes both a computer-
controlled collimation unit for the f/8 secondary mirror, and an active axial
support system for the primary mirror. As in other active optics telescopes
(cf. WIYN, NTT), coma is compensated for by moving the secondary mirror
while the other low-order aberrations (spherical, astigmatism, trefoil, etc.)
are removed by changing the shape of the primary mirror. The new
collimation system for the f/8 secondary has been in use since January of
this year, but the active primary support system is new, and appears to work
quite well.

                  The Active Primary Mirror Support System

As has been described in previous Newsletters, the axial support system for
the 4-m primary mirror lent itself very well to an upgrade to active control,
because it is based on 33 force actuators which are arranged in two rings
below the back of the mirror (Figure 1). The actuators are pneumatic pistons
("air bags"). They had previously been controlled by supplying all of the air
bags within each ring with the same air pressure from a single controller,
adjusting that air pressure to compensate for the changing axial component
of gravity as the telescope pointed away from zenith.

                          [Figure not included]

To convert this to an "active" support system, we changed the plumbing in
the mirror cell so that each air bag now has its own separate air pressure
controller. These are controlled by a personal computer that calculates the
force patterns (i.e. the correct air pressures for each air bag) to bend the
primary mirror to compensate for astigmatism, spherical, trefoil, and
quatrefoil aberrations. The desired corrections to the mirror shape can be
obtained either from a lookup table (the normal operating mode) or can be
explicitly entered (for instance, if new values have just been measured with
an image analyzer). For each air bag, the PC adds the forces needed to
correct each of the individual aberrations to the nominal force needed to
support the mirror against gravity, converts the total force to a desired air
pressure, and in turn converts this to a control voltage, which is sent to the
air pressure controller. The PC receives commands from the telescope
operator and obtains the telescope position from the Telescope Control
System (TCS) computer. The active optics can be turned off, in which case
the PC exactly emulates the old air-pressure control system.

The air-pressure controllers are inexpensive ($200 each) eloctropneumatic
transducer units manufactured by Mamac Systems. With a slight
modification, these work with ample precision to control the shape of our
rather stiff mirror to well within the measurement limits. The PC talks to
these controllers through ADCs and DACs connected to an RS-485 network.
Also connected into the system are various interlocks to shut down the
mirror support system in case of excessive air pressure, or if the mirror lifts
off any of its three defining hard points, or if the TCS or other vital
components should drop out of communication. These software safeguards
are backed up by a parallel series of electromechanical safety shutoff
switches, consistent with our great desire not to be the people who launch
our 4-m mirror into orbit.

Our active optics system is patterned directly after the one pioneered on the
ESO New Technology Telescope (NTT). Our main innovations are to apply
the technique to an old-fashioned, thick-mirrored telescope, and to use the
very low-cost approach based on the existing air-bag force actuators. While
planning and carrying out this project, we have received tremendously
valuable help from Ray Wilson and Lothar Noethe, who were principal
figures in the NTT project. We are deeply grateful.

Our in-house team was led by German Schumacher (software), Gabriel
Perez (mechanical hardware) and Eduardo Mondaca (electronics hardware),
with Jack Baldwin and Brooke Gregory functioning as scientific
rubberneckers. These people were backed up by a considerable number of
other engineers, machinists and electronics technicians who put in a huge
amount of work during the installation phase. We would especially like to
thank Gale Brehmer, Jorge Briones and Oscar Saa in this regard. Finally, we
must give full credit to Bill Weller, who became the original champion of
this project after we had all visited the NTT one day.

                              Why We Need It

The primary mirror support system has worked reasonably well for
approximately twenty years now, but as a result of an intensive series of
measurements with various image analyzers over the past few years, we
have come to realize that low-order optical aberrations do cause significant
degradation of the image quality. This has been especially true over the past
year or so, during which the telescope was plagued by serious amounts of
astigmatism when it was looking to the north. Figure 2 shows a map of
astigmatism over the sky, with the zenith at the center. Astigmatism is
expressed as a vector, since there is both a magnitude and a direction
associated with it. There are obviously large and systematic effects for
northern declinations.

This problem appeared sometime during 1993. For most of the time since
then we thought we had concrete evidence that it was due to a problem in
the primary mirror radial support system, but our attempt to fix it during the
4-m shutdown this August did not work out (see "CTIO Instrumentation
News" article in this issue). In late October (just as this article was being
written) we found indications that the astigmatism might instead be due to a
problem with the f/8 secondary mirror's support system, so we are re-
evaluating our various measurements over the past year. In any case, the
active primary mirror support is capable of compensating for this
astigmatism.

                       Results with the Active Optics

The active primary mirror support system was installed during the month-
long shutdown this August. It basically worked the first time we turned it
on. After a couple of nights of testing using the old air-pressure regulators to
make sure that the telescope was working at some basic level, we first tested
the new system in a mode that emulated the original support system and
showed that it produced results identical to the old system at telescope
positions all over the sky. Then we calibrated the active modes by forcing
various amounts of each aberration, using a Hartmann screen to measure the
amount and direction of the resulting aberrations.

Once this calibration was in hand, we went to the area at hour angle = 0,
declination = +30^o, where the astigmatism problem was at its worst. The
Hartmann screen showed that with the active optics we were able to
decrease the astigmatism by at least a factor of 6, which brought it down to
zero to within the measurement errors.

The following night we used the telescope without the Hartmann screen. The
seeing was good, with 0.7" FWHM images at zenith. We looked at a star at
0h, +20^o and measured FWHM = 1.22" without the active correction.
Turning on the active correction reduced this to FWHM = 0.97". We were
consistently able to obtain a 0.2"-0.3" improvement in the image FWHM at
various points in the north, and focus sequences showed that the astigmatism
clearly went away when the active optics corrections were turned on.

Since that time, we have had two subsequent engineering runs which
concentrated on mapping the aberrations over the sky and on developing and
testing the software necessary to use those maps as lookup tables for
correcting the image problems. As of this writing, we have installed a
system that uses a still-preliminary map, but which as far as we can tell is
able to reduce the astigmatism to acceptable levels at all points in the sky.
Figures 2 and 3 compare astigmatism maps with the active optics corrections
turned OFF and ON, respectively. We have left the new support system
running in its "active" mode while the f/8 focus is in use, as a test, because
we feel that it will be of benefit to all users of the telescope.

                           [Figures not included]

                                Future Plans

Our main goal over the next 2-3 months is to continue improving the maps
used to generate the lookup tables and to continue enhancing and debugging
the active primary mirror support system. To do this, we have 1-2
engineering nights each month through January.

We have now gotten most of the way through a three-year plan for
upgrading the optics on the 4-m telescope. We have completed the
conversion of the telescope control program (TCP) to the VxWorks
operating system (thus gaining comparability with KPNO), refiguring the f/8
secondary mirror (which improved our best f/8 images from 1" to 0.7"
FWHM), the computer-controlled collimation system for f/8 (another
important component of our active optics system), and now we are almost
done with the primary mirror support system.

The above has all been done within our initial three-year schedule. The one
thing that has slipped is our image analyzer project. Full use of the active-
optics system requires having an image analyzer permanently mounted on
the telescope in a way that it can be used with any instrument. This will
permit astronomers to tune up the optics in real time for limiting
observations. Our plan is to install a Shack-Hartmann image analyzer on the
existing Cassegrain offset guider. This is aimed mainly for the f/8 focus,
although it may be possible also to use it with the f/14 secondary, which is
now being figured. All of the mechanical work and optics for this system
have been finished for a year, but the CCD detector system has been
delayed because of manpower shortages. We now hope to have the image
analyzer in operation sometime during the first half of 1995.

Beyond these projects, the continuing image improvement program for the
4-m will consist of a lower-level effort of trying to consolidate the
improvements in the optics and thermal environment to push closer to the
site's seeing limit. While we find it gratifying to see 0.7" images frequently,
we still need to find out why that is our best case rather than our median.

The main effort at improving the 4-m telescope (as opposed to its
instrumentation) over the next few years will be directed towards
a) implementation of an f/14 tip-tilt secondary system, and b) a three-year
program of improvements to the telescope's control systems and drive
servos.

German Schumacher, Gabriel Perez,
Eduardo Mondaca, Jack Baldwin

Previous Article
Next Article
Table of Contents