Thermal issues (hot stuff...)

Last updated Sept 1st, 2000


See drawing CH2917.C001

On February 9th, we have replaced the old temp4m PC with the new Labview-based system. The data is accessed by typing 'temp4m' from any machine. That will show you the important fields logged in /ut02/seeing/temper/newtemp.log which are described below:


1 2 3=A 4=B 5=C 6=D 7=E 8=F 9=G 10=H
UT time UT date Lo tru NW Lo tru NE Lo tru E Lo tru W  Up tru N Up tru E Up tru S Up tru W
11=I 12=J 13=K 14=L       15=M 16=N 17=O 18=P 19=Q 20=R
unused Mirror glycol Mc air out Oil pump Low dome Oil bef pad Oil aft dome High dome PF cage Cass cage out
21=S 22=T 23=U 24=V 25=W 26=X 27=Y 28=Z 29=a 30=b
Mirror W Chimney Mirror N Mirror E Mirror S outside Dome floor Oil glycol Dome RH Oil volt read
31=c 32=d 33=e 34=f 35=g 36=h 37 38 39 40
MC volt read Oil pump status Cass cafe in air M1 cell air above M1 guider air dew point oil volt calc mc volt calc  



  •  the letters are the addresses of the DGHs
  • Mc stands for "mirror cooler"
  • 35 and 36 added on 31Aug00
  • Since 04/nov/99, both temperature logs (the old one 'temp.log' and the new one) read UT time (instead of local time before).



Typing "temp4m" on a Sun machine gives a list of temperature:

See drawing CH2890-A001 for mirror cooler details.



  • In day time: "mc glycol" should be negative typically between -5° and 0°(if the cooler works)
  • At night time: "mc glycol" gently comes back to ambient temperature
  • In day time: "Valve voltage" should read between 0 and 5 volts.
  • At night time (cooling off): a voltage is still calculated by the PC and logged but there is a pressure gauge in the air duct (just after the cooler) which, after the sucker mode is being selected at the start of the night, detects 0 pressure in the lower air duct and will override the PC voltage and send 0 to the Mamac so that the valve is closed. The glycol stops flowing and warms up slowly to ambient during the night (#5 in temp.log).



  • Day and night: "Glycol" should be negative typically between -5° and 0°. If this is not the case, think about an eventual problem in the glycol room (tank + pumps) located in the building next to Las Tacas. When glycol doesn't flow to cool the oil, "Before Pad" and "Glycol" will say typically +30° (oil warms up to ambient temperature and a bit higher at night when the telescope moves). During day time the oil valve is closed and the "oil glycol" sensor will be a good indicator of the glycol temperature sent by the plant and getting to the mirror cooler system (which comes second in the system after the oil loop).
  • At night time: "Valve voltage" should be >0 and "Pump On/Off" should say 1 (if telescope is active)
  • In day time: "Valve voltage" should be =0 and "Pump On/Off" should 0



  • Day and night, "Low dome", "High dome", "Serr truss", "Primary North", South" and "West" should always be within 1-2° from each other (and closer at night time), the mirror temperatures being hopefully slightly lower than the dome ones.
  • Cass cage is always a few degrees hotter than the previous set of sensors.



  • There are 2 separate circuits starting from the Glycol room: one for the building air conditioning and the other for the 4m telescope of : oil + mirror air cooler + M/C floors.


Notes about the old 'temp.log':

a- On 04/Nov/99, we found out errors in field labels in temp.log:

* #3: previously saying south, was actually east
* #4: previously saying west, was actually south
* #6: previously saying north, was actually west

b- On 04/nov/99, we modified the labels on the old temp4m PC display and swap connectors at the multiplexer input in order to have the PC algorithm use the correct sensors and log a trustful temp.log as indicated in the table above.

c- the old grey relay box in the Cass cage has the following inputs for the mirror sensors: #1 with cable labeled north, #2 with east, #3 with south and #4 with west. DO NOT change that!

d- the values displayed by the multiplexer in Tolnet show small variations (+/- 1°) from the reality (due to sensor calibration, effects of the cables,...) and serves only for debugging. The old PC algorithm uses a correction table for each sensor (file "tcoef").