• scan19 fitting a sphere to the data
  
• remeasuring a wedge 9 times to improve the rms measurement error

Topics:

• Gain loss as we change the radius of curvature of the primary.
• Current errors as we vary the radius of curvature
• Gain improvement from fixing tiedown blocks
  

• the radial errors are in cm. the x,y axis is in feet :).
• Red means the dish is below the design surface (radius too long)
  
• blue has the dish above the design surface (radius too short)

Gain loss as we vary the radius of curvature.

• These plots assume the correct radius of curvature is 265.176 meters. It then looks at the radial error if you use a sphere of smaller curvature.
• the smaller sphere has its center of curvature moved down by the same amount as the radial shortening (so the curves all overlap at the center of the dish)
• Page 1: assume the radius of curvature is 3cm short (2002)
• top frame: the radial error vs xy distance from the center of the dish.
• Near the center, changing the radius does not affect the radial errors (since you also moved the center).
• Toward the edge of the dish (150m) the radial error if -.55 cm (it's too short)
• The red and blue lines show  where a 225m beam is placed for the gain loss computation
• Bottom frame: The gain loss vs frequency for a radius 3cm short.
• The complex electric field was integrated over a 225m circular beam.
• The phase  was radialErr/Wavelength*2*pi.
• The red line is a beam centered at za=0 degrees (center of the dish)
• the blue lines is a beam centered at 8.7deg za (40meter offset). This beam just touches the edge of the 305meter diamter.
• At 5 Ghz the loss is -.02 and -.1db for the 2 curves
• at 10 Ghz (3cm) the loss is -.13db and -.34 db
• so  the 2002 sphere did not add appreciable loss
• Page 2: assume the radius of curvature is 7 cm short (2020 4 parameter fit).
• The two frames are similar to page 1
• At the edge of the dish, the radial error is -1.3cm.
• The gainloss is
• 6cm: -.2db and -.5db
• 3cm: -.7db and -1.9db

    The 7cm  radial error  gets up to -2db at 10ghz. The measured gain loss at xband can be up to 8db, so this is only part of our problem

processing: x101/p50/200311/analyzerrs/lossradiusshort.pro

Current radial errors and radius of curvature

This section looks at how the distribution of errors across the dish depend on the radius of curvature.

The image shows the radial errors (Measured - fit) for the 3 parameter fit (.png)

• The colors show the radial errors.
• Red has a radius 5cm too long (below the reference surface)
• blue has a radius 5cm too short (above the reference surface)
  
• You'll notice a blue area around the edge of the dish. Which means it is above the reference surface.
• Part of this can be explained by having a radius of curvature that is too short.

•  Start with the radius, center of the reference image (265.176 meters)
• Recompute the errors using a shorter radius
• make the radius shorter, and lower the z0 center of the sphere by the same amount
• Clip to show only the errors above 2 cm (red points) and below 2cm (blue points)

• The red ovals are problems with main cable tiedowns pulling too much on the portion of the dish above them.
  
• As the offset moves from 0 to -7 cm, the blue outer rim decreases.
• Changing the radius does not affect the central portion of the dish as much (at least not outside of 2cm error).
• The 4 parameter fit (center and radius) ended up with a radius 7 cm short
• It pushed the center of curvature down until the errors at the center started to cancel,

processing: x101/p50/20311/analyzerrs/errsabovecliplev.pro

Gain loss from the surface errors.

• Use the 2000x2000 grid of radial surface errors. Any errors > 5 cm were ignored (to remove things like hf dipoles)
• generate are 2k x 2k reference grid of random errors with an rms of .22 cm. This was the goal for the 2000 survey results.
• Pick an x,y spot on the dish
• find all points with measured rms errors within 225/2. meters of this point (beam radius)
• generate the phase errors for a particular wavelength for these points: radialErrcm/lambdaCm * 2*pi.
• generate the complex E field for all of these points using unity amplitude and the measured phase.
• Sum the E field for the reference and sum the Efield for the measured points.
• Take the ratio of the intensities as the gain loss.

• Each frame  shows beams moving from -400 to +400 x position. (- is west)
• The top frame is y = +400 (north), the bottom frame is -400 ft (south)
• The gain loss was computed for 21, 12.3,6, and 3 cmd (1420,2380,5000, 10000 Mhz)
• the colors show the different wavelengths
• The large errors at +x and -Y  are mapping into gain losses (especially at 10 ghz)
• The losses at x and cband are similar to what we see on the telescope calibration runs.
  
• The sband losses are smaller than the ones we've measured on the telescope.
• What these plots don't measure:
• I only took points on the dish.. for the measured and reference beams. When the beam spilled over, i ignored those points (since there were no measured errors). The plots do not show the normal gain falloff because of the spillover.
• There may be points > 5cm that were excluded (by my thresholding)

The image gives a rough idea of what the gain loss should be at 6cm across the dish (.png)

• the image was made by creating a grid  via interpolation of the points in the line plot above.

These errors only use the radial errors on the dish. Errors from collimation/positioning of the secondary and tertiary would add to these errors.

processing: x101/p50/200311/gainloss.pro

Gain improvement  by fixing the large errors.

    The large  errors  centered on the main cables are caused by tiedown blocks that have moved downhill, pulling too hard on the main cables.
The other large error is from the east-west cable that was broken and then repaired.
    An attempt was made to compute the gain improvement we would get  if we just fixed the large errors. The steps were:

• Start with the 2000x2000 gridded image of radial errors (clipped to +/- 5cm)
• compute the gain loss for these errors using the same steps as "gain loss from surface errors" above.
• make a copy of the 2000x2000 image and set all points with |errors| > 1.2 cm to no data.
  
• this would get rid of most of the large errors.
• recompute the gain loss for this dataset.

The image shows the radial errors after clipping all points > 1.2 to nodata (.png)

• The dark blue areas were > 1.2 cm error and have been excluded.

The plots

• generate hase
  

<- page up
home_~phil