Subject: minutes of the October Userr's Committee meeting

From: Michael Strauss

Submitted: Tue, 10 Oct 1995 11:21:22 -0400

Message number: 19 (previous: 18, next: 20 up: Index)

	Minutes to APO 3.5m User's Committee
	 Phone Conference, October 9, 1995

  Attending: Julie Lutz, Rene Walterbos, Bruce Gillespie, Ed Turner,
Michael Strauss, Don York, Chris Stubbs, Lew Hobbs, Rich Kron

  1. Update from Bruce Gillespie. 
  2. Recent engineering activities and results
  3. Scheduling issues.
  4. Preparation of budget for Board of Governors in November (we will
get copies of this, distributed by Aspasia). Let Bruce know of any
specific suggestions.

Bruce Gillespie: 
  Things are working so well, there is not much to say.
  Highlights of this past month:
    The amount of uptime has been much better than previous months due
to good weather and fewer hardware problems. 80% of scheduled time was used
for science or scheduled engineering. 17% of the time was lost to
weather, only 3% to instrument problems. Quite respectable.
   The seeing has also been better. r0 as measured by the seeing
monitor has gone up from 6 cm to 10 cm over the last month. Impression
has been that the telescope has been delivering better seeing. Close
to 1.3" average over the past few weeks. (0.9" last night!). Does this
just reflect that the weather patterns have changed, or is it due to
improvements at the telescope? 

  70-75% of all runs are remote observing. The networks have not been
as noise-free as we would like. A test has been run every hour to test
how well the network is working. A series of packets is sent to each
ARC site, and number of packets returned and time for return are
monitored. Success rate is 70-100%; round-trip time 50-300
seconds. Looking about speeding up transmission using
compression. Also looking into alternative routing (APO to UW requires
23 hops!). Chris Stubbs also looking into alternative routing which
can reduce this number to 7, using an NASA backbone.  We definitely
need to monitor and improve the network connections, as well as backup
(telephone) connections if the Internet is down.

  An alternative is to develop specific ftp-like transfer software
optimized for the problem at hand; potential speed up of transfer time
of images by a factor of 10.

  Average rate of data transfer during observing is ~10% of T1. We pay
$14K for T1. T3 would probably cost around $100K.

  Stubbs: tasks to interrogate RA and Dec from telescope through DSC
brings telescope to its knees. Information transfer is somehow screwed
up; the protocols that DSC uses are different from that of the other
  Kron: Fermilab group is out there right now working on the DA for DSC,
Monitor telescope; presumably working on this problem. 
  Stubbs: DSC has some electronic problems with CCD that need some
clean-up, UW wants to bring the instrument to Seattle to work on this
and have TIm McKay come out as well. 
  Instruments all behaving themselves reasonably well. John Brinkman
is on-board. Will be deciding what to do with instruments during
down-time; let him know any suggestions you might have, 

Engineering (Gillespie): 
  Telescope collimation: A problem with secondary actuators appeared;
mostly repaired, but we're not convinced that we can return reliably
to collimation over day-long timescales. Are actuators missing steps?
The secondary is rezeroed every day in an attempt to get around this.

  20 Hz resonance: By moving frequency of servo motors to 25 Hz much
diminishes this, but image motion at other frequencies goes up. Total
power unchanged, unfortunately.
  Tension of diagonal struts holding secondary cage is not to spec
(too low by factor of 4), putting resonances at too low
frequency. Tightening these rods is non-trivial (getting close to
failure mode).
  The altitude drive servo has limited dynamic range; if we reduce the
gain by a factor of several, the least significant pulse will be much
smaller, cutting down on induced vibration in the telescope. Will slow
down altitude slew speed by the same factor, but this is rarely the
limiting factor in speed of slew. 

  Stubbs: Plan to trail images across DSC with power off during a
night of good seeing to see how much of the seeing is due to the
optical figure, and how much is due to these various resonances.

  CHAOS measured 0.4" images in  1-2 sec exposures at 6000A. Wow!
Corrections he has to put in are in agreement with known figure errors
in secondary.

Full-blown Hartmann screen with 44 holes will be done tonight (unless
the weather doesn't cooperate; a possibility to take science time for
this project in the next few nights if it can't be done tonight).
This will characterize overall optical performance of the system,
although it doesn't tell you unambiguously which piece of optics is
responsible for the errors that are seen. Strong suspicion that the
secondary is the culprit. One test would be to redo the test with the
secondary rotated 90 or 180 degrees; it might be doable reasonably
straightforwardly with minor work on the secondary support
structure. How about the tertiary? Very difficult to rotate this?
Stubbs: what about a prime focus Hartman test? This would be
absolutely unambiguous. Not easy however; you have to get a camera up
there which could take this very fast beam. Could this be done during
this fall engineering run? Probably not.  Could the Rodier technique
be used to measure the wave front at prime focus more easily since it
utilizes out-of-focus images?

  Enclosure wheels: redesigned wheel is on order; expected in 2 weeks
from L&F. Cracks are continuing to spread. A real problem if this
wheel is delayed; we may have to delay the start of the month-long
engineering shut-down by one week, by shuffling science
programs. We'll make a decision in 7-10 days. There is a reserve fund
of $60K for catastrophic failures, such as this problem with the
wheel. The whole project cost of replacing all four wheels is $25K.
Exact weight of the enclosure is not known, thus complicating the
wheel problem.

  Guider: lots of engineering being done. Guider definitely works to 0.2"
closed loop on bright (8th mag) guide star. To do:
  Characterize guiding ability as function of brightness of star.
  Characterize guiding ability close to zenith. 
  Implement a calculation of rotator angle which gives you a bright
star in the guider.
  Hope to get something out for use by scientists in the next week. 
  Remark should be able to operate the guider soon, but not yet; right
now, you have to have the night assistant do it. 

  For longer term, there are plans to put a 512x512 chip into the
camera, while the current 1024x1024 chip is sent back to Spectrasource
to reduce read noise. Another improvement will be to put in a field
reducer to get a larger field of view (pixel size is 0.1"). Another
possibility is starting over with a new camera.

  Throughput testing: DSC data by Chris; not yet reduced.
             A long integration PSF to get wings would be very useful,
	     best to use DSC. 
Scattered light is probably from the secondary? dirtiness of the
optics; baffling might also be an issue. 

  Planning for upcoming shutdown going well. A list of 40 tasks lies
ahead. Enough to keep us busy for six months. The 4-week shutdown will
concentrate on mechanical aspects. Bruce will distribute this list of
tasks soon.
  This task list includes a number of goals, such as, at the end of
six months: 
        1. Have the amount of uptime available to do science >= 90%
        2. Telescope points, guides, tracks to within a factor of 2 of spec
	3. Images of 1.2" median, and a best seeing of 0.6". 
  (and so on).  

  A new model, following ROSAT: Each institution decides on some
fraction of the proposals making up 50% (70%?) of its time allocation
to be of high priority, to be scheduled exactly has requested. An
additional 100% (60%?) of the time would be proposals (perhaps in a
rank-ordered list), which can be scheduled according to what fits best
into the remaining schedule. This requires a new level of commitment
for insitutional schedulers to make sure that people don't get
rejected many quarters in a row. Either people will get closer to what
they requested, or they won't get anything at all. Also requires that
people be very specific as to what the minimum time is that they
require to do their science.

  Next year's budget will be discussed at the Board of Governor's
meeting next month. Possible major new items include:
  1. Money for a secondary mirror ($250K?  A blank exists, polishing
is straightforward; the testing is the most involved process. May take
up to two years).  
  2. Add a software engineer or systems administrator for APO (a goal
of software self-sufficiency on the site in 1.5 years).
  3. A new camera for the guider (see above). 
  4. Infrastructure for people on the site: test equiptment,
basic electronic components, spare chips, and test devices

Gillespie: observing support from site staff. We do not do service
observing, whereby observatory does all the observing for you.
However, staff would like opportunity to be involved in scientific
collaborations; they could actually carry out some of the observations
of some such collaborative project. Of course, Bruce has to be kept
informed, and this takes lower priority than their regular
duties. Turner et al.'s QSO gravitational lensing program has been
done in collaboration with Long and Bergeron. People should be aware
of this, and take advantage of it where it is appropriate. Ideally the
site staff would be involved in scientific process (i.e., not just
carrying out flat fields in the afternoon). We have to guard against
it being abused, which would cause much ill-will. It is not part of
the duty of the site staff to take part in these collaborations;
participation in projects would be up to the individuals involved,
assuming that it doesn't take up too much of their time. 

  Next meeting Monday, November 13, 1995 at noon. 
				-Michael Strauss

------- end -------
APO APO APO APO APO  Apache Point Observatory 3.5m  APO APO APO
APO  This is message 19 in the apo35-general 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