• The Gregorian dome and trolley weigh over 200 kips.
• The ch weighs about  35 kips.
  
• There is a 47 kips counter weight at the end of ch side of the azimuth arm to try and balance the weight (when the ch is at 8.8 and the dome is around 8 degrees, the azimuth arm is balanced... the gregorian za is the optical axis position, not center of mass za).
• During normal operations the ch sits at za=8.8 deg while the dome moves 1.06  to 19.6 deg za.
• This causes the entire platform to tilt as the dome moves.
• We compensate for this tilting by shimming the elevation rails higher than normal, so the dome tilts down onto the focal surface.
• The entire platform also moves vertically as the main cables heat or cool.
• We correct for this by pretensioning the tiedown cables (that pull down on the platform).
• As the temperature goes down and the platform rises, we pull harder on the td cables.
• When the temperature goes up  and the platform drops, we release tension on the tiedown cables.
• If the temperature gets too high, we lose the pretension in the td cables, and we can no longer push the platform back up.
• As the dome moves in front of a td cable at mid to high za, the dome weight causes the td cable in front of it to lose tension.

    So

• during noon, on a hot day, when the dome is at a high za, and the azimuth points at a tower, we lose tension in the tiedown.
  
• This gives an added tilt to the platform.
• The higher the temp, and the higher the za, the more the tilt.
• This extra tilt causes a pointing error.
  
• This error is exasperated by the tiedown tracking mode. This consists of:
• tiedowns are in tracking mode (tracking the average platform height).
• every 2 minutes the distomats measure the average height of the platform.
  
• If the platform is too high, we pull down on the tiedown cables.
• If the platform is too low we release tension in the tiedown cables.
• The problem occurs when we run out of tension in 1 of the td cables.
• The program does not check to see if there is any tension in the td's (since the load cell reading fail so often).
• It pushes all 3 td's up (releasing tension).
  
• two of the td's have tension so 2 corners of the platform goes up,
• the td with no tension does not move its platform corner when it moves up (since there is no tension to release).
• This causes the platform tilt to be worse.
• --> So when we lose tension in a td, we should stop td auto tracking (or make it so the load cell tensions are more reliable).
  

    If we new how much extra the platform was tilted, we could use the above links to compute the pointing error.

Note: The the platform also translates...

•  If you work out the forces from the tilt, gravity, and the main and td cables:
• with the dome at high za pointing at a td cable
• The corner rotates down
  
• We lose the tension in the td cable
• the length of the main cable has to increase (since corner is moving down) so the tension in the cable increases
  
• This causes the horizontal component (toward the tower) to increase
• This increase is not canceled by the horizontal components of the 2 other sides..
  
• 
  

Measuring the platform tilt when we lose tension.

• From the laser ranger data we compute:
• The average platform height
• the x,y,z translation of the platform
• the rotations about the x,y,z axis.
• Note:
• x axis is west
• y axis is north
• z axis in down
• take all the laser ranger data for a given epoch with:
• tension in all of the tiedowns
• the average height is within .1 feet of the correct height.
• make a model of the platform tilt vs az, za.
• When we lose tension in a tiedown:
• get the platform tilt from the laser rangers.
• subtract off the platform tilt from the model for the given az,za..
• Compute the pointing error from this residual tilt.

Platform tilt models

• 2017: 01jan17 -> 18sep18 (Pre-maria)
• 2018: 01jan18 ->29jun18 and 30jun18->31aug18
• post maria. included distomat change (and jump) on 29jun18