HYDRA ALERT- February 8, 2005
Greetings, Hydra user -
As you are aware, Hydra has been suffering for the past 2 years with fibre
interaction issues. We have been exploring all potential sources of the problems
during the recent upgrades, but we are still not confident that it is behaving
reliably and predictably in the deployment of its fibres. In particular, we are
concerned about fibres deployed with significant angles out of the pivot points,
and about buttons placed near the predicted arc of such fibres. We have seen
multiple cases of crossed fibres occurring when simulation suggests the fibre
deployments and button placements are valid assignments.
Until we can uncover the underlying causes of the problems and correct them, we
cannot allow Hydra to continue to be used without imposing
further restrictions on deployments in an attempt to safeguard those fibres
which remain in service.
For the next few runs, we will publish a revised version of "whydra" in which
the ANGLES of fibre deployments are restricted to values which we understand to
be "safe". That is, when a fibre is deployed it will be along a path that has
been predicted and is measured within small, acceptable errors. That deployment
angle is currently estimated at 4 degrees.
What does this mean for forthcoming runs? It is going to force only those
deployments where the fibre "bend" is very small to be allowed. We will be
preparing an interim version of "whydra" and will ask you to use this to obtain
your fibre assignments for runs in February (and March). The resulting fields
may be examined on the Hydra Simulator whereupon additional valid deployments
could be identified on a version which matches the limits set in "whydra".
We request that each Hydra configuration file be mailed as a separate text file
to diharmer@noao.edu and it will then be checked out on a revised version of the
simulator here at NOAO. It would be helpful if the majority of "unassigned"
objects could be edited from the file, leaving only potential FOPs and the "next
level" priority targets which could be assigned if a valid assignment could be
identified.
It has been suggested that we might need to remove Hydra from service altogether
until our problems can be uncovered and fixed. Since we cannot reasonably
predict such a time, we are offering this process so we may keep Hydra
operational but with minimal risk to fibre damage.
We recognize that some fields will yield fewer - and possibly different -
targets than originally planned, but hope this will not have too great an impact
on survey tasks or similar studies where large numbers of targets are ultimately
to be observed. Please bear with us as we try to keep Hydra operational while
striving to identify and fix these long-standing problems.
You will be advised where/how to obtain an interim version of "whydra" as soon
as it is ready for use. Configuration files may be e-mailed to diharmer@noao.edu
for review. Thank you for your cooperation and understanding.