Summary of the 1997 Holography Antenna Measurements at the HHT



Holography is a well-known technique in radio astronomy to study the surface accuracy of a dish surface. Some references that might interest the reader are:

  • A Holographic Surface Measurement of the Texas 4.9-m Antenna at 86 GHz by C. E. Mayer, J. H. Davis, W. L. Peters III, and W. J. Vogel (1983), IEEE Trans. Instrum. Meas., 32, 102-109.
  • Holographic Measurement and Setting of the Submillimeter Telescope Reflector Surface by W. L. Peters, J. G. Mangum, R. N. Martin, J. W. M. Baars, and G. Narayanan (1994), BAAS, 26, 1317.
  • Holographic measurement on Medicina radio telescope using artificial satellites at 11 GHZ by D. Tarchi, and G. Comoretto (1993), Astron. Astrophys., 275, 679.
  • Radio-holographic reflector measurement of the 30-m millimeter radio telescope at 22 GHz with a cosmic signal source by D. Morris, J. W. M. Baars, H. Hein, H. Steppe, C. Thum, and R. Wohlleben (1988), Astron. Astrophys., 203, 399-406.

Since the installation of the telescope surface, we have undertaken holography observations to measure the overall dish surface accuracy. Previous results, using the LES-8 satellite beacon at 38 GHz, are discussed in "Holographic Measurement and Setting of the Submillimeter Telescope Reflector Surface" by W. L. Peters, J. G. Magnum, R. N. Martin, J. W. M. Baars, and G. Narayanan, (1994) BAAS 26, 1317. The observations used a receiver on loan from NRAO Tucson, mounted at the prime focus of the HHT reflector. The Phase reference signal is obtained through a lens/horn combination at the rear side of the prime focus box. The two phase synchronous IF signals are fed into a digital signal processor box built specially for this experiment. Results obtained in March 1995 are also presented.

Unfortunately, the LES-8 satellite died, so that it was unavailable for further holography work. The LES-9 satellite was available, however, and the Air Force agreed to schedule our program among the many other projects that use the satellite.

From September 15th through October 9th, 1997, the HHT conducted holography observations to measure the dish surface accuracy. We used the LES-9 satellite, provided by the Air Force and MIT Lincoln Laboratory. This satellite has a 37 GHz transmitter, which the Air Force let us use for 6 hours a day for 24 days. The satellite normally is used for communications, most notably for communications with the research stations at the South Pole. The power capacity of the satellite, originally launched in 1976 has decreased enough that when the 37 GHz transmitter is turned on, the satellite cannot be used for anything else. To minimize the inconvenience to the other users of the satellite, we got the same six hours each night. This meant that the satellite was always at the same elevation (60 degrees). In 1995, using the LES-8 satellite, we were able to make measurements as low as 35 degrees, but that is when the satellite is above the horizon at the South Pole, and so was not possible during the 1997 season.

Initial Observations

In 1995, we believed that every panel was positioned to better than 40 microns, with an average of about 25 rms over the entire dish. When we made our first measurements, we found that two panels were 100-200 microns out of position. The location of the two worst panels was not surprising. They were the two nearest the two struts that were repaired in 1996 after two accidents.

As inspection of Scan 4488 below will show, the inner ring was still well aligned after 2.5 years (since the last measurement). Some of the outer panels, however, have moved. We are not sure as to the cause: whether they were disturbed during the accident or if they gradually crept out of alignment since the last holographic observation.


Adjusting the Panels

After convincing ourselves that the maps were repeatable and verifying the orientation of the maps relative to the antenna, we began a series of adjustments to the panels. We would adjust only the worst one or two panels and then make a new measurement. If one compares the first and last maps in the table above, the effect is dramatic! The surface is significantly better than the figure we had in 1995. Then we claimed about 25 microns rms overall, with some parts of the outer ring noticeably worse. We now are getting between 15 and 20 microns rms overall. The outer ring is still the worst, but is much better than in 1995. It is the hardest to measure, since the illumination of the antenna is least here, and therefore the signal to noise ratio is the smallest here too.


+/- 050 microns range +/- 100 microns range Scan number (rms)
First Map First Map Scan 4488
(rms = 47 microns)
The first scan
First Map First Map Scan 5490
(rms = 24 microns)
An intermediate scan
Some panels adjusted
First Map First Map Scan 5963
(rms = 19 microns)
A final scan after adjustments done
Shows a non-random pattern.
First Map First Map Scan 5998
(rms = 14 microns)
Taken right after 5963 on the same night
Much better consistency

The blue color corresponds to panels that are too far below the average, and red corresponds to panels that are above the desired average surface.

We are able to do much better than in 1995 because we solved a data contamination problem that we had back then. It turned out to be a hardware problem in the holography backend that caused some data from the dish channel to be put with data in the reference channel, and vice versa. It happened less than 1% of the time, but that was enough to cause problems. The problem was solved a few months after the holography measurements in 1995, but by then the transmitter in the LES-8 satellite that we had been using failed.


Reaching Our Repeatability Limit

By the end of our holography run, we had reached a repeatability limit. In order to better understand this, we stopped moving the panels a few days before the end of the run and simply repeated maps one after the other. Some maps have 15 micron rms deviations and others have 20 microns rms. Some of the repeatability is not random, and is undoubted partly artifacts in the data which might be removed by better analysis or editing out spurious data. In some cases, the antenna may actually be changing between measurements. We are in the process of examining the data more carefully, so that we can see how much of the non-random contribution can be attributed to artifacts.


Averaging Together Maps

One can attempt to minimize the effect of the non-repeatability by averaging together maps. Averaging the last three maps, the next to last three maps, and the last six maps, we find rms values of 19, 16, and 16 microns respectively. Based on this, we can say that the surface is set to between 16 and 19 microns rms. This is significantly better than in 1995 and is close to the 15 microns specified in the pre-construction error budget for panel alignment. (The panels themselves should be better than 10 microns rms). When time and weather permits, we will measure the antenna efficiency at high frequencies to independently estimate the accuracy of the antenna surface.


+/- 050 microns range +/- 100 microns range Scan number (rms)
Last 3 Maps Last 3 Maps Scans 6166-6283
(rms = 19 microns)
Average of last 3 maps
06-10-97 to 08-10-97
Next to Last 3 Maps Next to Last 3 Maps Scans 5963-6073
(rms = 16 microns)
Average of next to last 3 maps
04-10-97 to 05-10-97
Last 6 Maps Last 6 Maps Scans 5963-6283
(rms = 16 microns)
Average of last 6 maps
04-10-97 to 08-10-97

The non-repeatability can be quantified by differencing maps. The resulting rms of the difference varies from 8 microns to 19 microns with a mean of 14 microns. If the differences were random, then one can divide these differences by 1.4 to get the random measurement error in a single map: 10 microns.


Future Work

We want to repeat these measurements in 9 to 12 months from now to see whether the panels move with time. Lincoln Labs are not sure if the LES-9 satellite will have enough power to operate the 37 GHz transmitter in the future.

We are looking for another satellite that could be used for holography measurements but have not yet identified one. What is needed is a satellite that can transmit above 20 GHz (the higher the better), and which is in a high orbit so that it stays above the horizon for the 5 hours it takes to do a measurement.

 Copyright Arizona Radio Observatory.
For problems or questions regarding this web contact [tfolkers{at}email{dot}arizona{dot}edu].
Last updated: 11/08/11.