wiki/ news/ 2003-11-23 - One more v2 CPA test with VNA; smush testing, etc

Met at the PSU Microwave Lab in the basement of FAB. It's back by the Sun station labs, the door closest to the giant chip tester behind glass. See attached map at bottom of page.

Here's the agenda:

  1. Remeasure the cylindrical patches: resonant frequency, SWR, bandwidth -- done
  2. Vary the radius of the cylindrical antennas while observing resonant frequency. -- done
  3. Squish a flat antenna with thick polyethylene sheet) and see how that affects things. -- done
  4. Try baking an antenna, and see if that makes a difference. (How?) -- partly done
  5. Carefully measure the antennae geometry, particularly thickness. -- partly done

  6. Anything else we should do? (Maybe, see the conclusions below)

Here's what we need to bring:

  1. All three CPAs
  2. All three flat CPAs
  3. 3/4 polyethylene sheet (Is this really 3/4"? I think it's 1/2".)
  4. Kapton tape, xacto knife, blades, scissors, foam tape, etc
  5. Two calipers (Good to check consistency.)

During the experiments the room temperature was approximately 20 degC

Re-measure Flat ATV Antenna

f0 = 1.277 GHz, SWR = 1.3, BW = 14 MHz

This is consistent with previous measurements of: 1.2787 GHz, 1.34, 13 MHz

(Why is the Smith chart so wacked?)

SCRN00.JPG: Flat ATV antenna Smith Chart
SCRN01.JPG: Flat ATV antenna SWR

Re-measure cylindrical ATV antenna

New: 1.242 GHz, 1.48

These are consistent with the old measurements: 1.243 GHz, 1.36

SCRN02.JPG: Cylindrical ATV antenna Smith Chart
SCRN03.JPG: Cylindrical ATV antenna SWR

Re-measure cylindrical GPS antenna

New: 1.523 GHz, 1.16, 89 MHz

These are partially consistent with the old measurements: 1.524 GHz, 1.09, 38 MHz

It's unclear why the bandwidth measurements differ. 38 MHz seems more believable.

SCRN04.JPG: Cylindrical GPS antenna Smith Chart
SCRN05.JPG: Cylindrical GPS antenna SWR

Re-measure Flat WiFi antenna

New: 2.422 GHz, 1.31, 53 MHz

These are consistent with the old measurements: 2.422 GHz, 1.28, 54 MHz

SCRN06.JPG: Flat WiFi antenna Smith Chart

SCRN07 is similar to SCRN06 but SCRN07 is taken over a larger frequency span.

SCRN07.JPG: Flat WiFi antenna Smith Chart
SCRN08.JPG: Flat WiFi antenna SWR

During this lab session we began by re-measuring the same old antennas. This gave us confidence that their properties were fairly consistent over time.

We then began pressing on the flat WiFi antenna with the flat edge of the polyethylene sheet. Here are the main observations:

  1. SWR, and to a lesser extent resonant frequency, are sensitive to the presence of dielectric over the feed structure.
  2. Frequency and SWR are very sensitive to the presence of dielectric over either radiating slot.
  3. There is very little sensitivity to the presence of dielectric over the resonant patch.
  4. Resonant frequency is quite sensitive to deformation of the patch, though the dependence is not very simple.
  5. Resonant frequency is only weakly sensitive to heating of the antenna.

The first 3 points were completely expected. In exploring the temperature dependence, we heated the ground plane (that is the interior) of the cylindrical WiFi antenna with a hot air gun. The temperature measured by contact thermocouple on the outside of the antenna reached 52 deg.C. At this temperature the resonant frequency was down by about 10MHz.

Most of our experiments concentrated on deforming the foam. Locally deforming the patch by applying an estimated 10 pounds of force with the flat edge of the polyethylene sheet resulted in resonance shifts up to about 20 MHz.

Where along the patch the deformation was concentrated was important to the effect the deformation had on the resonant frequency. Pressing at point approximately 1/4 to 1/3 of the way from either edge had little effect on resonance. Pressing closer to the edge lowered the resonant frequency, while pressing nearer the center raised it. Can someone post the actual data on this??

We also measured the cylindrical WiFi antenna in four states. Wrapped on the aluminum module, free on the bench with the vertical seam taped shut, on the bench with the vertical seam naturally sprung (about a 1 cm gap), and on the bench holding the seam apart under tension to a gap of about 5 cm. All these states produced very similar resonant frequencies and SWRs.

Finally, we measured the thickness of two cylindrical antennae. The WiFi antenna had an overall thickness of 0.080 +/- 0.001 inches, and the ATV antenna was 0.077 +/- 0.001 inches.


We assume that shifts in resonant frequency equal to or below 10 MHz are tolerable in our application. Shifts greater than 20 MHz may be intolerable.

  1. The antennae seem to be fairly stable over time
  2. Temperature stability is reasonably good
  3. Mechanical stability to deformations is marginal
  4. The resonant frequency of the cylindrical antennas is not sensitive to the exact radius.

The fairly good time-stability means that the foam is reasonably inert, and that the natural abuse the antennas have received hasn't greatly affected their properties.

The temperature stability measurement would have been more interesting if we could also have gotten a thickness measurement. The mechanical stability is described as marginal because the magnitude shift is large enough to be of concern and we have a known mechanism that might cause the antennae to deform in flight, namely low pressure expansion (see Andrew Sucks Foam).

The relatively low sensitivity to thermal expansion may be due to the genuinely small change in thickness resulting from the heating, or it may be due to a greater sensitivity to non-uniform patch deformation vs uniform deformation.

I'm not sure extra sensitivity to non-uniform deformation is plausible, but we didn't do the experiment simply because at the time we lacked a dielectric of the right width. Perhaps we should have tried harder.

These points are relevant in that they address the suitability of the foam core antennae. The concerns are great enough that i think two things are warranted:

Finally, the independence of the cylindrical antenna properties with respect to wrap-radius is interesting. Theoretically the resonant frequency should be mostly independent, but the effect of the vertical seam was unclear. Since it now seems that the exact radius has only a weak effect, we need to re-examine the analysis of the patch antennas for mistakes rather than complete omissions as we try to explain the observed differences between the flat and cylindrical test antennae (see the RedBlueGraph). One interesting thing we learned here is that the cylindrical antennae seem to be thicker than the flat ones, maybe this will help.