FEIC Beam Scanning meeting -- 31 July 2008

Attendees

  • Geoff, Sri, Todd, Josh, Antonio (NTC)
  • Richard, Fred (phone)

Agenda

  1. Status of compressor and beam-scanning system.
  2. Next datasets to obtain?
    • It was noted that scans with Andrey's horn produce narrower beams than with the NTC horn, even when both are corrected with the right cos^N values. However, scans 105/106 were done with the feed not at 45 degrees. Should we repeat these in order to test the NSI correction process?
    • Richard noted that we look forward to obtaining scans with a larger (and offset) scan window in order to avoid truncation effects. How large? In which receiver bands and polarizations?
    • We only have a theoretical cross-pol pattern of the transmitter horn. How important is it to measure the relative contributions to the -20dB of cross-pol we see at ~6 deg off-axis arising from the FE vs the probe (e.g. using a polarizing grid). What are the concerns in the experimental setup?
    • Can we aim to measure something specific from the list above on the days that Richard may be visiting CV? (Aug 14-15?)
  3. Probe correction
    • Any news from NSI about reading in the actual pattern of a horn rather than using N values?
    • Shall we proceed to try to do the probe correction offline? What mathematical processing is needed? Who will do it?
  4. Status of plans to measure the pattern of a spare reference feed from the ALMA holography receiver.
  5. The calculation of XYZ offsets and beamwidths in the NSI software is similar to Richard's spreadsheet for scans 103 and 104. What are the remaining discrepancies to investigate? both in scans with a probe vs. an OEW?

Minutes

'> 1. Status of compressor and beam-scanning system.

Geoff expects to be running scans again by Friday.

'> 2. Next datasets to obtain?
'> * It was noted that scans with Andrey's horn produce narrower beams than with the NTC horn, even when both are corrected with the right cos^N values. However, scans 105/106 were done with the feed not at 45 degrees. Should we repeat these in order to test the NSI correction process?

We have evidence that the probe correction does not have the effect we expect in terms of beam widening. Richard suggested a simple test of the probe correction. Take the ratio of the farfield corrected and uncorrected patterns of a scan. The should give the transform of probe pattern. Does it make sense? NSI uses cos^N(Az)*cos^N(El)

ACTION: Josh will output the data and make it available. Fred or Todd or Darrel will analyze.

  • Hi, I have done this with one of the recent band 9 scans (scan 106), and put the results here.
  • Ok Richard, I did it again using another scan (band 9, scan 94), which had larger N numbers.

Of the two probes, the one with the smaller horn implies a smaller correction and is likely to be closer to the truth. Richard asked if we could just use Andrey's band 9 horn? Geoff: Yes, we can.

But we are also trying to obtain an open-ended waveguide. Ideally, one would to attach to the final multiplier output a piece of ~1 inch WR0.8 waveguide with the end tapered and absorber around it. But we apparently don't have a piece of this small waveguide. After the meeting, Geoff and Todd sketched a drawing for how to build a home-made version using a split block groove tapered on the output and soldered to a standard flange on the input. For Band 3, we do have an open waveguide probe in house. ACTION: Antonio to order open-ended waveguide probes for bands 6,7,9. Geoff to complete the drawing for a home-made band 9 probe and submit to shop.

ACTION: Geoff will start scanning at Band 9 initially using the small horn probe to test the experimental setup. He will send around the initial results. If the results make sense, he will swap to an open-ended waveguide when available. He will compare the S/N obtained with the open-ended waveguide and see if it is sufficient for our purposes.

  • Richard noted that we look forward to obtaining scans with a larger (and offset) scan window in order to avoid truncation effects. How large? In which receiver bands and polarizations?

We need ~15% larger coverage than present. Need 10 deg square for band 9. Could use the same center, but note that the beam propages toward the principal axis, so a shift might make sense. Suggest using the same sampling spacing as before. May be able to decrease it later if the results do not noticeably change. To be complete, there are five scans to be made per frequency setting: each polarization channel with the transmitting horn at 90 deg, each polarization channel with the transmitter at 0 deg, plus one scan with the transmitter at 45 deg.

(Todd: Why don't we need two scans at 45 deg? one for each channel)

'> c. We only have a theoretical cross-pol pattern of the transmitter horn. How important is it to measure the relative contributions to the -20dB of cross-pol we see at ~6 deg off-axis arising from the FE vs the probe (e.g. using a polarizing grid). What are the concerns in the experimental setup?

After the tests of the open-ended waveguide described above, Geoff will experiment by adding a polarizing grid (tilted by ~10 deg) with absorber placed at the termination of the reflected direction(s). Take scans to re-measure the cross-pol with Band 9 and Band 3 (both?)

'> d. Can we aim to measure something specific from the list above on the days that Richard may be visiting CV? (Aug 14-15?)

Richard is attending URSI in Chicago and may choose to visit CV in place of the final days of the meeting pending airline schedules. We expect to still be doing scans then, but we should not defer ourselves from other FE priorities if the need arises.

'> 3. Probe correction

'> a. Any news from NSI about reading in the actual pattern of a horn rather than using N values?

Josh reports that NSI wants to charge ~$10k to produce a pattern file into an (apparently) proprietary binary format. It's not clear if this charge is per file (i.e. per band). We decided to pursue the open-ended waveguide approach before pursuing this option or related options (such as deconstructing the file format). But we will inquire on the cost to get the software to create a pattern file.

'> b. Shall we proceed to try to do the probe correction offline? What mathematical processing is needed? Who will do it?

This was viewed as a fall-back position of the open-ended waveguide approach does not work out due to S/N or other isues.

'> 4. Status of plans to measure the pattern of a spare reference feed from the ALMA holography receiver.

Antonio reported that the current plan is to make a full room-temperature 3mm receiver (using the one spare feedhorn) that could be mounted into the place of the ALMA cryostat and be scan-able with the same (undisturbed) setup. This way, it could be loaned to different FEICs for testing purposes. As far as timescale for construction of this unit, the effort required must be considered with respect to other FEIC priorities.

'> 5. The calculation of XYZ offsets and beamwidths in the NSI software is similar to Richard's spreadsheet for scans 103 and 104. What are the remaining discrepancies to investigate? both in scans with a probe vs. an OEW?

One item that Richard noted is that the polarization calculation should compute the ratio of crossPol/coPol over the subreflector area, rather than the area inside the 10dB contour (or some other region). Also, for computing offsets, use the first moment of the power over the subreflector area. He has this in his spreadsheet. Josh intends to use Richard's spreadsheet for analysis in the near future, as it is a tractable path. Consider writing standalone code using the same equations in the long term (C, python, etc.)

-- ToddHunter - 20 Aug 2008
Topic revision: r4 - 2008-08-21, ToddHunter
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding NRAO Public Wiki? Send feedback