Temperature difference Centerbeams - Conebars
Temperature differences between Centerbeams and Conebars
February 2005
Measurements with temperature sensors located in different parts of the
telescope showed that large systematic temperature differences exist between
the two structures which connect the telescope and the counterweight: the
centerbeams (located within the receiver cabin) and the conebars
(located outside the cabin). These temperature differences vary significantly
between day and night. This is shown for June 2004 in the plot below (the red
line). The amplitude of the temperature difference is 6-7 degrees. The blue
line indicates when roof and doors were open. The dotted lines are the times
of sunrise and sunset on Mauna Kea (not the moment when the Sun appears
above the ridge). The grey lines show the
measurements of the individual 12 conebar and 16 centerbeam sensors on one
day.
The variation of this temperature difference is due to two effects: daily
ambient temperature variations and heating of the cabin by the instruments
located there. Smaller effects are probably due to elevation changes of the
telescope. The warm air from the cabin is removed via a hole in the cabin
ceiling near RxB. There is a fan and a system of tubes which leads air to
the plinth area. Also warm air is escaping through the hole in the top of
the cabin towards the SMU.
Long ago it was found that this was not sufficient and insulation material
was placed around the four centerbeams. See e.g. this picture of the
cabin.
Air in between this material and the
beams was removed by fans located near the cabin floor with colder air entering
from above the roof of the cabin. I don't know whether this system ever worked
well. There were different rumours about its limited effect, but this was
never properly investigated. Since many years the fans were not running.
On 8 December 2004 this insulation was repaired. One fan was found
to be absent, and one did not work well because the space behind the fan was
blocked by large pieces of insulation material. This was removed, the broken
and absent fans were replaced and holes in the insulation were closed. The
fans were tested but not turned on because of an E&C program to study the
variation of the SCUBA FCF at the beginning and end of the night had to be
completed first.
The fans were turned on on February 13 at 8h HST.
The figure below shows the same as the first figure, but for the month
February 2005. It is seen that the amplitude of the red line decreases
significantly after February 13. Unfortunately in the beginning of the
month the weather was very bad and on several days roof and doors could not
be opened.
The next figure shows this for two typical days before and after February 13.
It appears that still four centerbeam sensors show somewhat larger temperatures
than the rest of them, although they show the same variation. These are the
sensors 207, 208, 227, and 230 which are located in the bottom left (northeast)
centerbeam (the one between RxH3 and RxB3). It is unclear why these
temperatures are higher. Within each centerbeam there are no large temperature
gradients such as before.
The last figure shows for a number of days before (black) and after (red)
February 13 the temperature difference variations. On these days roof and doors
were open during a typical period (i.e. not much extended observing and no
closure because of bad weather) from 17h-17h30 - 9h30-10h. The amplitude of
the variation decreased on February 13 from about 4.5 degrees to about 1.5
degrees.
It is hoped that this decrease in the temperature variation between centerbeams
and conebars is reflected in an improved surface accuracy at the beginning
and end of the night.
An investigation of surface changes between
June and November 2004 showed that surface accuracy was typically
33 micron at the beginning of the night, compared to 26 micron in the middle
of the night.
New holographic measurements were done during first shift on February 13.
An 80 GHz map obtained directly after opening roof and doors at 17h38 had an
aperture rms of 28.4 micron. A second 80 GHz map at 20h53 had an rms of
26.9 micron, and a 160 GHz map at 21h44 had an rms of 25.6 micron
(corresponding to an rms at 80 GHz of 26.9 micron). The rms at the beginning
of the night was only 1.5 micron higher than at 22h HST, compared to typical
values of 6-7 micron in 2004. This suggests that the surface now indeed is
better at the beginning and end of the night. However it should be confirmed
by more observations and FCF measurements with SCUBA.
It should be possible to decrease the temperature variation even more if
additional measures are taken (e.g. allowing warm air from the cabin to
escape more easily).
|
