Author Topic: Antenna Polarization indoors: doesn't matter much  (Read 4249 times)

joelucid

  • Hero Member
  • *****
  • Posts: 867
Antenna Polarization indoors: doesn't matter much
« on: May 17, 2016, 07:18:22 AM »
Given how well the dipole (or how badly the wire) antenna worked in my setup I'm currently looking for alternative and less intrusive antenna's for my tiny motes. I figure with a dipole on the GW I have enough link budget to spare to run a tiny antenna on the mote at 300kbit - and probably at -18dBm tx power.

The one I like best is the semi-loop as described in this app note: http://www.hoperf.de/upload/rf/ANTENNAS_MODULE.pdf. It's basically a loop trace antenna an the border of the main PCB.

However I was always concerned about the polarization mismatch - horizontal small magnetic loops radiate with horizontal polarization. Good news on that front: I can't really measure much degradation with a simple TH mote using a wire antenna at different angles to the gw dipole indoors. And http://wireless.stanford.edu/papers/Kyritsivtc01_fall.pdf comes to the same conclusion - they measure -3db cross coupling across polarization in an indoors lab.

It seems that reflections, refraction etc mess with the polarization enough that it doesn't matter much at the receiver. One step closer to the ultra-thin disk life-time th mote  :)

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: Antenna Polarization indoors: doesn't matter much
« Reply #1 on: May 17, 2016, 03:30:34 PM »
Given how well the dipole (or how badly the wire) antenna worked in my setup I'm currently looking for alternative and less intrusive antenna's for my tiny motes. I figure with a dipole on the GW I have enough link budget to spare to run a tiny antenna on the mote at 300kbit - and probably at -18dBm tx power.

The one I like best is the semi-loop as described in this app note: http://www.hoperf.de/upload/rf/ANTENNAS_MODULE.pdf. It's basically a loop trace antenna an the border of the main PCB.

However I was always concerned about the polarization mismatch - horizontal small magnetic loops radiate with horizontal polarization. Good news on that front: I can't really measure much degradation with a simple TH mote using a wire antenna at different angles to the gw dipole indoors. And http://wireless.stanford.edu/papers/Kyritsivtc01_fall.pdf comes to the same conclusion - they measure -3db cross coupling across polarization in an indoors lab.

It seems that reflections, refraction etc mess with the polarization enough that it doesn't matter much at the receiver. One step closer to the ultra-thin disk life-time th mote  :)

It's great to hear that you now have such a large link budget.   :)  If Felix still has any left, maybe you'd want to try out Felix's trace antenna Mega as a mote in this configuration.  I seem to recall they were nicely compact, but unfortunately I don't recall anyone posting anything about them after they were first put up for sale.  Would anyone who has tried them care to comment about their  performance as compared to the usual wire antenna?
« Last Edit: May 17, 2016, 03:56:25 PM by WhiteHare »

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: Antenna Polarization indoors: doesn't matter much
« Reply #2 on: May 17, 2016, 08:24:16 PM »
I just took a closer look at the Moteino Mega trace antenna, and it looks quite different than other 915Mhz trace antennas that I've seen on commercial products.  Is it perhaps some kind of loop antenna or something?  There seems to be one or two passive components that close the circle.



Not sure how that's the same or different (or better) than other types of trace antennas, but it sure does look interesting!

[Edit: Answering my own question:
]

« Last Edit: May 17, 2016, 10:34:18 PM by WhiteHare »

joelucid

  • Hero Member
  • *****
  • Posts: 867
Re: Antenna Polarization indoors: doesn't matter much
« Reply #3 on: May 18, 2016, 07:28:16 AM »
Quote
It's great to hear that you now have such a large link budget.   :)  If Felix still has any left, maybe you'd want to try out Felix's trace antenna Mega as a mote in this configuration.  I seem to recall they were nicely compact, but unfortunately I don't recall anyone posting anything about them after they were first put up for sale.  Would anyone who has tried them care to comment about their  performance as compared to the usual wire antenna?

Yeah Felix antenna is nice and that it worked so well in part motivated me to research the topic. I think for the th mote the semi loop might be even better. These loop antennas are very much dependent on loop area size and I can't put a 2x1cm loop next to my mote - it would destroy the form factor. But increasing diameter to 2.5-3cm and running a loop around the pcb - now that's a viable proposition.

We're talking about this after all: https://lowpowerlab.com/forum/index.php/topic,1617.msg11856.html#msg11856

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: Antenna Polarization indoors: doesn't matter much
« Reply #4 on: May 18, 2016, 07:48:30 AM »
For itty bitty things, chip antennas have lately seemed popular.

joelucid

  • Hero Member
  • *****
  • Posts: 867
Re: Antenna Polarization indoors: doesn't matter much
« Reply #5 on: May 18, 2016, 08:04:46 AM »
Quote
For itty bitty things, chip antennas have lately seemed popular.
ReplyQuoteNotify

Certainly another option for 868/915 MHz. For 433 they're just so big...

I have a couple lying around i also need to test. However I do like the isotopic radiation in the loop plane. You just don't get that with chips. Plus I love that no ground plane is necessary. And it's cheaper. Let's hope it works.

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: Antenna Polarization indoors: doesn't matter much
« Reply #6 on: May 18, 2016, 11:10:15 AM »

joelucid

  • Hero Member
  • *****
  • Posts: 867
Re: Antenna Polarization indoors: doesn't matter much
« Reply #7 on: May 19, 2016, 12:50:22 AM »
http://www.digikey.com/product-detail/en/taoglas-limited/ILA.01/931-1080-1-ND/2332744

This one does look interesting: a ceramic chip loop antenna which would give you all the advantages of a loop with higher efficiency. However it comes at a prohibitive price point.

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: Antenna Polarization indoors: doesn't matter much
« Reply #8 on: May 19, 2016, 06:55:50 AM »
Interesting.  Your research is providing some improved perspective.  Let's see:
1.  Lower frequencies have better propagation characteristics, but
2.  those lower frequencies sacrifice more db's to get to a compact size, unless
3.  cost goes up
Is that right?

If that is true, then is it all just a wash, or does any one particular ISM band standout as best suited for both range and compact size and low cost?  Is it 433Mhz or 850-950Mhz (or possibly an even higher frequency?).

joelucid

  • Hero Member
  • *****
  • Posts: 867
Re: Antenna Polarization indoors: doesn't matter much
« Reply #9 on: May 20, 2016, 10:22:15 AM »
Check this out: http://www.mikrocontroller.net/attachment/19206/Ringantenne70cm_CQDL4_83.png

It's a 433mhz loop antenna with 96% efficiency and less than 5cm diameter. Now 5cm is a little large for a Mote. But how about 3.5cm with 50% efficiency. That would be amazing. And the copper ring could work in a thin disk form factor, too. It would use the real estate more efficiently by going into the z direction.

If this actually works it would be a great solution for the gateway as well.

Quote
If that is true, then is it all just a wash, or does any one particular ISM band standout as best suited for both range and compact size and low cost?  Is it 433Mhz or 850-950Mhz (or possibly an even higher frequency?).

You absolutely can't deduce from the existance of counteracting forces that it's all a wash.

Quote
If this actually works it would be a great solution for the gateway as well.

Which would even give back the 3db polarization loss.

TomWS

  • Hero Member
  • *****
  • Posts: 1926
Re: Antenna Polarization indoors: doesn't matter much
« Reply #10 on: May 20, 2016, 03:07:56 PM »
Check this out: http://www.mikrocontroller.net/attachment/19206/Ringantenne70cm_CQDL4_83.png

Uh, one question, why is the German paragraph about 10 times longer than the French or English paragraphs?

I'm glad you can read this so that you can do the first implementation!

Tom