---

Intro:

    During the day we can loose tension in the tiedowns because of the weight imbalance on the azimuth arm (dome=225 kips, ch =35Kips + counterweight). The elevation rails are shimmed to be above the paraxial surface as you go up in zenith angle. When shimming the elevation rails:

•  pretension the tiedowns so there is non zero tension in the tiedowns as the dome goes from 1 to 20 degrees za.
• compute the shims vs za so the receiver horn always sits on the paraxial surface.
• The forces at work are:
• The torque from the dome increases as the dome goes up in za (by the increase in the moment arm).
  
• The tension in the tiedown at the adjacent corner decreases as the dome end of the azimuth arm dips down
• The change in height of the horn is the torque increase from the dome motion minus the decrease in tension from the tiedown cable.
• If the platform is too low, then we can loose tension in a tiedown before the dome gets to 20 degrees.
• When this happens the corner moves down more than normal since the torque from the dome increases, but there is no decrease in tension from the loose tiedown
• (of course i've oversimplified since i'm ignoring the td on the other side of the azimuth arm).
• So  the horn ends up below the paraxial surface.

• compute the pointing error when moving the platform vertically, and when tilting beyond the normal tilt (more info).
• A  model has also been made (using the laser ranging data) to tell how much extra tilt,vertical offset the platform has (more info)
  

• All Saturday, Sunday observations
  
• Except:
• exclude 430 MHz observations since the tiedown tracking is normally turned off
• exclude any periods where we don't have all 6 laser ranger measurements
• This can occur when it rains,  or when the tilt is so large that the distomat beam not longer falls on the retroreflector  on the corner of the platform.

Pointing error plots

• Each page is a separate month (june,july,aug,sept)
• Each day is plotted in a separate color
• Top: pointing error from platform rotation.
• yaxis is great circle arcseconds.
• xaxis is hour of day (ast)
• I've clpped the vertical scale to + 100 arcseconds
  
• 2nd frame: pointing error from vertical height error
• Bottom frame: platform Height - platform reference height
• This shows how high or low the average height of the platform is (in cm).

Discussion:

• Most days have some rotation pointing errors 60 to 100 arcseconds
• This is probably when the dome is at higher zenith angles.
• the height of the platform can be 5 to 10 cm low
• How are the different receivers affected.
• The FullWithHalfMax (fwhm) is the beamwidth between the half power points of the beam.
• A pointing error of fwhm/2 will have a 3db gain loss. Larger errors will continue to drop down the gaussian.
• When moving the platform vertically i've seen a 3db loss when moving 1 lambda vertically (but correcting for poitning error).
• Ao receiver specs:
• I've interpolated the measured beamwidths to the edges of the bands
• sbw goes 1700 to 3000 with lots of notched filters
• I've used 2380 (sbn) freq and the upper edge
• cband: goes 4 to 6 MHz, but 5150 to 6000 is unusable because of the UNII band.
  

• Receiver	freq MHz	lambda cm	BeamWidth fwhm asecs	pntErr for 3db loss Asecs	focusMotion for 3db loss cm
  327	327	92	900	450	92
  430	430	70	660	330	70
  lbw	1120- 1750	27- 17	265 - 178	132 - 85	27- 17
  sb(n,w)	2380 3000	12.6- 10	123 - 95	61.5 - 47.5	12.6 10
  sbh	3000 4000	10- 7.5	98- 83	49 - 42	10- 7.5
  cband	4000 5000	7.5 6.0	75- 58	37.5 29	7.5 6.0

  
  

Summary

• Pointing errors from platform tilts and vertical offsets were inspected for weekends during june,july,aug,sept2019
• The pointing error  is often 60 to 100 arcseconds between 08:00 and 18:00
  
• this depends on the temperature, az,za, and probably whether it has rained or not (the water cools the main cables).
• Disclaimer:
• These errors are computed from the measured platform position, a model of the correct tilts, and an attempt at computing pointing errors from tilts
• They could easily be off by 10 arcseconds..
  
• The table shows the degradation in gain for these types of pointing errors.
• Observers should probably be informed of this problem.
• People probably next ask for the pointing errors versus az,, za (maybe one of these days :).
  

processing x101/190918/plottiltsum.pro

<- page up
home_~phil