Measured a CP antenna and linear good, CP pattern bad.

[groundhog1]

I built up a plain rectangular patch antenna with the probe feed on the diagonal to set up circular polarization.

Using my network analyzer, and putting itentical antennas on both ports, everything looks fine to me. I can rotate the antenna and at center freq, the insertion loss between the antennas (ant gain and space loss) looks fine and predictable.

The gain is slightly lower than I expect at around 1 dbi.

Then I sent one of the antennas to my buddy in another state. He has an antenna chamber. I also sent him with the antenna, a cavity backed wideband spiral RHCP antenna to use as a pickup.

He sends back data that doesn't make sense.
When he uses his horn pickup, and takes two cuts 90 degrees apart, the pattern looks ok. But when he uses the RHCP spiral as a pickup, the pattern is weird and very lossy. He also has a LHCP antenna and used that too to plot the opposite polarization. His plots of RHCP and LHCP gain are about the same in magnitude.

Any hints as to what could be causing this?

thanks,
gh

 

[VEBill]

"...a plain rectangular patch antenna with the probe feed on the diagonal to set up circular polarization."

While waiting for the others to join in, let me ask a quick question:

Does this antenna create LHCP or RHCP?

 

[Higgler]

With the same plot in RHCP as LHCP, that automatically means you have a linear antenna (or nearly linear).

Did he measure the correct frequency? Patches for circular can be very narrow band.

I assume that when you had both identical antennas connected to the analyzer, and rotated one, that the amplitude S21 meausured between them didn't change much with rotation. Did you rotate one at least 90 degrees? Explain that measurement a bit.

What size ground plane do you have. If your tests at the analyzer don't use a ground plane, and your friend mounted the antenna on a ground plane for pattern tests, that can change the frequency of the antenna alot and you'd have gone from a circular analyzer measurement to a linear antenna pattern measurement.


kch

 

[groundhog1]

The antenna radiates RHCP.
The ground plane is not much bigger than the patch. Substrate er=10.

At the network analyser in my lab, I used two bare antennas.
I rotated 360 degrees or more.
What you see in the S21 display is like a wave in the ocean when you rotate it, with the center fo point staying still.
No other ground close by except the floor and other reflectors normally found in a lab. With all that, it looked ok to me.

At my friend's chamber, the two linear cuts are very similar in shape and are what I would expect to see.
The two CP cuts LHCP/RHCP are not identical. They are both lumpy and have gains that are maybe 20dB lower than the linear ones that are just below 0dbi.

Did I answer everything?
gh

 

[VEBill]

My point was perhaps too subtle.

You antenna needs to have a fundamental circular asymmetry (with 'screw-thread' like physical or electrical characteristics) so that it 'knows to how to choose' (excuse the expression) RHCP vice LHCP.

"...a plain rectangular patch antenna with the probe feed on the diagonal..." doesn't seem to have that essential characteristic fundamental to making one CP versus the other. Based on the description, your antenna sounds 'slant linear'.

Apologies in advance if I'm reading too much into the description you provided. But it also just happens to explain the apparently non-CP results, which is why I raised the question.

 

[groundhog1]

Thanks Bill.

The rectangular patch with the probe feed on the diagonal (almost) is a very common CP design. They radiate on all four edges rather than the two for regular patch antennas.

Also, in my lab with the network analyzer, using two of these identical antennas, I am getting good RHCP (seemingly). I can rotate the antennas and the S21 at fo just sits there still at the predicted separation space loss value.

It's just in my friend's chamber that the bad results are happening.

Am I answering your question or maybe I am missing something?
gh

 

[VEBill]

The "(almost)" helps with my understanding. Thanks.

 

[Higgler]

your friend measured the RHCP and LHCP wrong based on your info. You rotation tests prove it's a good Circular antenna, you can't rotate one antenna fully 360 degrees and have the coupling stay steady if it isn't circular.

Normally, you measure antenna patterns Linear (say V pole), then rotate your transmit horn to H pole and measure the same antenna pattern, then calculate RHCP and LHCP, which typically goes up 3 dBish in the gain of your antenna.

It's impossible for the RHCP and LHCP gains to be lower than your Linear Gains by nearly 20 dB, unless your friend made an OOPS, just simple math.

Ask your friend how he measured the V, H, RHCP and LHCP.

kch

 

[Higgler]

One other note regarding patch antenna gains. Pozar, in his "Microsstrip Antennas" book writes that one wavelength groundplanes will provide 1 dB of patch antenna gain, and as you increase your groundplane size you get approximately one dB increased gain for each additional groundplane length you increase. It's a bit of an S shaped curve, not linear, but linear is easier to remember.

Buy the book if you want alot of good info on Microstrip Antennas, isbn 0780310780.

i.e. 3 wavelength square groundplane will make 3 dB gain increase from a small (say half wave groundplane) and 5 wavelength square groundplane will have 5 dBi gain for a standard patch. Anything above 5 groundplanes and you don't get anymore gain.

This was for a linear patch, probably applies somewhat to circular too. If someone says your gain's too low, tell them it's due to your small groundplane, not a poor design. I'm certain very few people will know that, and you'll look brilliant.

kch

 

[groundhog1]

Thanks for the tips on seeming brilliant!

I often hear wonderful complements like " oh yeah... you're a freakin Genius"

So your suggestion should only enhance the effect I have on people.

groundhog