LowPowerLab Forum

Hardware support => RF - Range - Antennas - RFM69 library => Topic started by: john k2ox on October 25, 2013, 01:36:23 PM

Title: Antenna Tutorial or Antennas in a Mote
Post by: john k2ox on October 25, 2013, 01:36:23 PM
        Practical Information Regarding Antennas for use with Moteino’s.      John k2ox 10/25/13


1.   Antenna performance is usually referenced to that of a dipole antenna.

2.   If your antenna design radiates the same power in all directions as a dipole it is said to have zero gain referenced to a dipole antenna.  Its gain is 0dbdipole.


3.   Antennas are reciprocal.  Their transmit gain is equal to their receive gain.  If the antenna transmits best in one direction it also hears best from that direction.


4.   You cannot get more power out of an antenna than you put into it!

5.   So what determines the size antenna?  Well now things get complicated.  I’ve spent years devising techniques to explain antenna theory without referring to Maxwell’s equations.  Hopefully, this will get you started.
 

6.   The smallest resonant antenna (1/2 wavelength) is 6 inches at 1000 MHz!  Now you have a method of calculating antenna lengths without formulas, calculators, etc.  At 500 MHz, 12 inches (2x).  For the FM radio band (100MHz), 60(10x) inches. All referenced to a Moteino antenna! The lower the frequency the longer the antenna.  Real antennas are a little shorter because radio waves are slower in air than in a vacuum.


7.   Why is Moteino’s wire antenna only 3 inches long?  It was discovered in the early days of radio that ¼ wave antennas (hundreds of feet tall for the low freq’s used at that time) mounted vertically on the ground acted like full size ½ wave antennas.  Since the earth is so much larger than the missing quarter wavelength it acts like an image of the real quarter wave vertical.  If a conducting object on the ground side of the antenna is a ¼ wavelength or longer in some dimension the antenna will be an effective radiator.  The asymmetry will result in a nonsymmetrical radiation pattern though.  For the 3 inch ¼ wave Moteino antenna to be most efficient it needs a rf conductive mass at least 3 inches long.  The ideal ground plane is a disk with a radius of 3 inches with the 3 inch wire sticking out of its center.  The disk being connected to the Moteino ground terminal and the monopole the ‘ant’ terminal.


8.     What happens if I have a little coin cell mounted in as small of a package as possible?  Well, it will certainly be smaller dimensionally than the 3 inches needed for the ground plane.  There are two concerns, it will be an inefficient radiator and something we haven’t discussed yet, it will have an impedance mismatch.



9.   Antenna users always wish their antennas were smaller. AM radio stations use ¼ wavelength verticals that are two hundred feet tall.  These are very expensive antennas.  The installation costs, real estate, maintenance, power for aircraft warning lights and on and on.   Moteino users also don’t like dealing with antennas dangling from their projects.  Small antennas do have their down sides.  If they didn’t AM Broadcast stations would be extremely happy to eliminate the expense of those gigantic ¼ wave verticals.



10.   The smaller the antenna the more you have to compromise.  That doesn’t have to be a bad thing though.



11.   Antennas radiate Electro Magnetic waves.  The orientation of the electric wave describes its polarization.  A dipole antenna positioned with its elements parallel to the ground has horizontal polarization. One pointing up and down is vertically polarized.  The largest signal is received when the received signal has the same polarization as the antenna.  If the polarization of the received signal is opposite of the antenna polarization then no signal will be received!  In this case there may be a lot of signal but it has the wrong orientation to fit through the ‘mail slot’.



12.   Small printed circuit antennas.  Electrically steerable RADAR systems use very small antennas arranged in arrays on a flat surface. These small antennas are built on ceramic substrates and other low loss materials.  Early cellphones had dipole antennas poking from the tops of their cases. Antenna engineers have adapted the RADAR designs for use in modern cellphones.  These are often referred to as patch antennas.  A 915 MHz patch antenna should easily fit on a 2 by 2 inch piece of printed circuit board. 

Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Felix on October 26, 2013, 12:26:34 AM
Excellent stuff!
Really looking forward for the follow up :)
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Vince on October 26, 2013, 09:52:04 PM
John,

     Excellent information!

Vince - WB2FYZ
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: jbeale on October 27, 2013, 12:20:52 AM
Nice writeup. I'm curious how you design patch antennas, do you work from first principals, scale a generic existing layout, or is there a a good software package for that? I remember using MININEC back in the day, but it just analyzed what you already had, did not suggest new ideas.

cheers, another john (n8juf)
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: john k2ox on October 28, 2013, 10:36:23 AM
I was just thinking I'd start with something that's follows published design rules, make measurements, and then scale it.  I need to see how dependent it is on the PCB substrate.

My tool of choice is EZNEC.  I've used it quite a bit to model long thin antennas.  I should take a look to see how it might be used for the patch.  They all use the NEC engines and have custom UIs.  I'm not sure it is setup for planar stuff.

Good thought.  I'll check it out.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: billchurch on October 28, 2013, 06:43:48 PM
This was a great read.  Right now my range requirements are pretty minimal, but when I start going outside with this stuff I'm probably going to need to be a little more exact about things.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: jbeale on October 29, 2013, 12:23:19 AM
I was just thinking I'd start with something that's follows published design rules, make measurements, and then scale it.  I need to see how dependent it is on the PCB substrate.
I was tangentially involved in a FR4 PCB slot-antenna design for 433 MHz, long ago. The PCB was space-constrained (aren't they always?) so the design was fairly narrow bandwidth (high-Q). In that case we had to measure each batch of boards that came in and select a tuning capacitor based on the measured properties. Within the batch, the FR4 dielectric constant was fairly constant, but across batches it varied a lot.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: sketchy on December 29, 2013, 04:27:10 AM
Thank you for the excellent info. 

What is the best wire to use for monopole antennas in an in-home network?  Is there a certain kind you use?
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Cabe on December 29, 2013, 11:14:23 AM
Anything copper that will hold its shape really, solid core ethernet network cable is pretty good stuff (everyone of the hopeRF style systems I have come across use it).

Personally I have a stash of helical bead antennas I am keen to try and some "rubber ducky" style as well.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: sketchy on December 29, 2013, 01:44:04 PM
(thank you for the reply).

So the 22 AWG gauge copper wire I use on my bread board will work as well as the wire that comes with the Mote?
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: john k2ox on December 29, 2013, 02:02:30 PM
Thanks sketchy.

Your solid core copper wire will work just fine.

Enjoy your motes.

john
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: tve on December 30, 2013, 03:35:19 AM
Interesting read, I looks forward to the patch antennas!

As far as I 've read, thicker 1/4 wave antennas (and dipoles as well) have wider bandwidth, which should provide a bit more tolerance on the length. I've used some hollow brass rods for a dipole (using a toothpick inserted into the rods to mechanically join the halves) and it looks good plus holds it shape well. Dunno whether the extra diameter really makes a difference, though...

Can someone explain why the coiled wires are so popular? As far as I can gather, the gain is lower over a 1/4 wave monopole, but there's less polarization. What am I missing? The cool factor?
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Cabe on December 30, 2013, 08:54:55 AM
They allow a physically shorter antenna, but still resonate.

But of course you sacrifice reception gain for that shorter length.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: photogphred on January 04, 2014, 02:49:43 PM
I'm curious, if a 1/4 wave antenna needs a 1/4 wave sized ground plane, what does a 1/2 wave antenna need? I'm assuming nothing? If so, then if you wanted to coil your antenna would it be better to coil a 1/2 wave instead of a 1/4 wave antenna given the layout of the Moteino? Is there an optimal coil size too?

Great discussion, by the way.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: LazyGlen on January 04, 2014, 03:24:58 PM
In another post, I suggested including an antenna on a sheild:
Has anyone tried using a PCB trace antenna? I laid it out in AutoCAD and found that about 1.5 perimeters would provide the right length antenna for 413MHz. "Cut Here" marks could allow the other 2 frequencies to use the same trace. Analog RF design is WAY outside my knowledge base, so I don't know if this is even possible.
I won't include the picture again to help with backup size's.

I understand the ideal is to have antennas perpendicular to the ground plane. However, as you point out in your post, antenna design is an exercise in compromise. Furthermore, if it transmits the distance you need it to, it's good enough, and most of us coil up the wire to make it more convenient anyway.

What are your thoughts on trace antennae? Being on a shield would require it to be connected to the current connection point anyway, so if it didn't achieve the desired range, a different antenna could be simply swapped in.

I'm not at the point that I'm ready to design a shield, but thought I would put the idea out. Both for input on feasibility and in case someone else could use it.

LG
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: john k2ox on January 04, 2014, 09:39:13 PM
Trace antennas are the way to go. Once you have a good design it will be reproducible and stable. Cellphones have the phone, wifi and bluetooth antennas printed as traces.

The biggest issue is get it to match 50 ohms at the desired frequency.  I use pcb material for my half wave dipoles. It's low cost and mechanically stable.

BTW. The thinner an antenna is, the narrower the freq range it will operate over, compared with a fatter antenna.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Felix on January 04, 2014, 09:53:27 PM
John, several people are asking me for trace antenna.
While the wire is currently the recommended cheapest and most flexible option, do you have any tweaked working trace antenna designs for 433/915mhz, anything you could share?
I imagine this means the antennas will take another half a Moteino, but still would be great to know what the alternatives to the wire are.

UPDATE: there are now Moteinos with Trace antennas available in the shop (https://lowpowerlab.com/shop/moteino).
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Cabe on January 05, 2014, 09:30:33 AM
http://df2ck.de/tech/helix/ and http://www.micrel.com/_PDF/App-Notes/an-23.pdf

Are what I have used as my primers on Helical antenna, I suspect most people are after a "Normal Mode" Helix, so the diameter and length of the helix should be smaller than the wavelength.

692.36mm is 433MHz wavelenght, so as you can imagine, most coils will be normal mode.

Most pre-made helicals I have found use a 6mm diameter and equal spacing. This technique has worked for me (if I find the website, I will post it too) Take a 1/4 wave lenth of wire and a something like a pencil and wrap it tight, then to get an even spacing just pull it apart slightly from each end. You can use this "springyness" to tune a helical to a degree to get the best out of them, but they will be about half the gain or less than if it were just a straight whip.





Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: jbeale on January 06, 2014, 08:17:51 PM
I'm no expert either, but I sat next to one at a previous job :-). You may find that putting your antenna on the same PCB as your CPU can cause some noise issues which reduces the effective receive sensitivity, unless you are very clever with the routing. I am still amazed by how cellphones manage to integrate all the antennas that they do (3G, wifi, bluetooth, GPS etc) but of course it becomes easier at GHz frequencies. The 315 MHz wireless home security sensors I've seen use a U-shaped wire loop that comes out perpendicular to the PCB.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: CapnBry on January 07, 2014, 03:03:15 PM
I have previously made a trace antenna for 915MHz RFM12B modules using the Texas Instruments reference design (http://www.ti.com/lit/pdf/swra227) for a meandering monopole antenna.
(http://i.imgur.com/Q4QA1zp.png) (http://imgur.com/Q4QA1zp)

I didn't have any method for matching the impedance so I just hooked it directly to the module. The performance was noticeably lower than just a piece of wire and much more directional (laying the board down so the antenna radiated up/down caused dropouts). I also know nothing about antenna design other than the basics of how antennas in general work, so there might also be something massively wrong with the design.

I've attached the EAGLE brd/sch if anyone wants to play with it.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: LazyGlen on January 09, 2014, 11:32:47 PM
ANNNNNnnnnnd THAT is why I'm not an RF engineer. Just looking at and attempting to read the app note makes my head hurt.

LG
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: kobuki on February 23, 2014, 06:23:39 AM
I wonder if one uses a monopole or a dipole that is not a half the wavelength but a multiple, like 2..4 times the 1/4 of the wavelength for a monopole or 2..4 times the 1/2 of the wavelength for a dipole, would it increase the antenna gain? How about the impedance? I've read a lot about dipoles and monopoles with a ground plane but it's always just shortly mentioned that an antenna should be N times the half wavelength and such but they never write about possible benefits, drawbacks, gain etc. For example, would a full wavelength dipole have 2 times the gain of a half-wave one? This way it would be very easy to increase gain (at the expense of longer, more inconvenient whips).
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: john k2ox on February 24, 2014, 09:50:14 AM
You are thinking correctly.  The dipole antenna needs to be an odd number of half wavelength so that it's impedance is near 50 ohms.  If it's a monopole cut the dipole length in half.

Just remember, as I've already stated, that to get gain in signal strength in a particular direction  gain will be removed somewhere else.

Looking at the 3D gain distribution of a 1/4 monopole vertical antenna it resembles a doughnut laying flat.  If you increase the length of the antenna to 3/4, 5,4,.. of a wavelength the gain distribution will look like a flattened doughnut.  The gain will increase perpendicular to the antenna, but less above and below.

This can be a problem with FM radio stations.  High gain antennas increase the far away distance, at the expense of close in listeners.  Since the antenna is hundreds of feet in the air the disk like high gain radiation pattern shoots right over the head off nearby receivers.

If your goal is omnidirectional gain you are on the right track.

john
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: kobuki on February 24, 2014, 11:05:27 AM
I see, thanks for the clear answer. Yes, my goal is to increase gain horizontally, so it appears to be the right approach. Naturally, if the radiated energy is the same, the spatial distribuition would need to be arranged differently. And now that I think about it, there are various high-gain wifi antennae on the market advertising a gain well above 10 dB, which are all reported to increase reception quality perpendicular to the whip.

PS1: If I wanted to learn a bit about antenna theory, what literature would you recommend for a beginner? This subject seems interesting.

(PS2: how is your trace antenna desing coming along?)
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Cabe on March 03, 2014, 06:16:51 AM
Just catching up on my electronics news, and this note on PCB antennas popped up on Hack-A-Day

http://hackaday.com/2014/02/28/pcb-antenna-reference-designs/

Thought it might be interesting for some.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Nuudeli on March 17, 2014, 01:43:33 PM
This doesn't relate directly to Moteino but I'll share this anyway... I use Hope RF modules (RFM22/23/69) for my own designs and have made a couple of boards with 'ready-made' antennas. I've used two different Linx antennas for 868MHz ISM-band and those are "ANT-868-USP" and "ANT-868-SP". Smaller (uSP) model is for remote controller based on Serpac CA-4 case and larger (SP) is for server/base station based on Hammond Manufacturing 1591XXSSBK case.

Didn't use any simulations before PCB orders but simulated matching cirtuicts before applying those (easy/accurate to simulate for low RF frequencies) and final performance was measured with network analyzer. PCB's were ordered from IteadStudio (actually quite high quality for DIY projects).

First layout for remote was my fist PCB ever and it had few optimization points which I corrected to second layout version. Antenna related observations were that PCB should have as solid ground as possible from antenna ground pads to core ground. I made some changes and used layer 2/2 only if absolutely needed ( + prefered vertical traces instead of horizontal) and added three component matching circuit between RFM and antenna. Also added more GND vias between layers to reduce parasitic inductance. I didn't have to use any matching (excluding 68pF series component which doesn't have impact on the impedance at that frequency) for the second layout and the performance difference between two layout versions were quite high. See attached figure.

For server board I used much larger PCB and different antenna. Learned from remote desing that should prefer solid ground, vertical traces and a lot of GND vias. Added also three matching component to make sure that I can tune the antenna if it's needed without new PCB order. It wasn't so bad than I thought it would be (not so solid ground after all). Still decided to tune the impedance with 18nH shunt coil (between RFM and series cap). Tuning improved performance a little but it would be better with smaller coil (didn't have time/interest to fine tune).

Now the performance for both should be quite ok, but of course near field effects (hand etc.) will cause some performance degradation. Don't have any range measurements yet but definetly will make some after I get software somehow functional. Attached also the board layouts if someone is interested in PCB size [mm] (doesn't include first layer which is quite full in both boards).
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: john k2ox on March 19, 2014, 09:17:43 AM
Thanks for posting Nuudeli.  What kind of distance are you reliably able to get from these?  What power level are you using?

tks,
John
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Tommy T on July 22, 2014, 11:40:25 AM
Hi Guys,

I'm a newbie at this so I was hoping to get some clarification on the following that I quoted from John's very first post:

Quote
My tuned dipole is trimmed to 5.8 inches for use at 915 MHz.  The antenna is fed at the center, it is cut into two with one side connected to ground and the other to the radio’s ant pin.

Does that mean that he has 5.8 inches of wire soldered on the ANT pin and another 5.8 inches of wire soldered on the GND pin?  Or is it a total of 5.8 inches with one end of the wire soldered onto the ANT pin and the other end soldered onto the GND pin?  Or is it something completely different?  A picture would help immensely.

What would the length be if it was monopole?

A lot of this antenna theory is a bit over my head so I apologize in advance.

Thank you,

Tommy
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Felix on July 22, 2014, 11:57:22 AM
I think he meant something like this:
(http://www.agi.com/resources/help/online/stk/10.1.1/source/extfile/comm/images/CommRadar03-5.gif)

If you search google for dipole antenna (https://www.google.com/search?q=dipole+antenna&espv=2&source=lnms&tbm=isch&sa=X&ei=fojOU-vaBZOLyATovoAo&sqi=2&ved=0CAcQ_AUoAg&biw=1920&bih=967#imgdii=_) you will find lots of images that explain the principle.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Tommy T on July 22, 2014, 12:36:44 PM

Hmm...I've looked at a bunch of sites on Google but none of them seem to mention how GND is figured in.  Do I simply ignore GND, bend the wire in half, and attach to ANT?  If so, why did John mention GND? 

Sorry, if it's obvious to everyone else.

Thanks again,

Tommy
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Tommy T on July 22, 2014, 03:49:18 PM

It looks like Mike over at www.airspayce.com also mentions attaching to GND.  Here's the link:

http://www.airspayce.com/mikem/arduino/RadioHead/classRH__RF69.html

The following is quoted from there:

Quote
10cm RG58C/U soldered direct to RFM69 module ANT and GND

10cm also sounds quite a ways off from the 82mm that's listed for 1/4 wavelength here: http://farm4.staticflickr.com/3818/10585364014_d028c66523_o.png



Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Lensdigital on August 11, 2014, 01:55:34 PM
Great info here!
So I have a question. Would I see any benefit if my transmitter's antenna is longer than receiver (T=1/2 with R=1/4)?  And vice versa, if my receiver's antenna is longer than transmitter (T=1/4 with R=1/2).  Which option is better?  Obviously I realize that if both have 1/2 antenna it would be the best, but might not be possible.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Felix on August 11, 2014, 02:11:29 PM
I think one of the more expert RF guys posted in one of the threads that odd number of quarterwaves lengths is better than even. So 1/4, 3/4 etc..
BTW the radios are both receiver and transmitter, unless you're using them only has RX or TX, but it still makes sense to listen for ACKs even when only transmitting.
I would just keep them all the same.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: KanyonKris on August 11, 2014, 07:40:01 PM
The RF principal of reciprocity says it won't matter. However you may have some environmental effects (RF being absorbed by or bouncing off objects in or near the link path) may favor one  antenna setup over the other so you may want to try each and look at the RSSIs to decide which to use. But my gut feel is you'll only see a little difference, if any.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: john k2ox on August 18, 2014, 08:51:29 PM
The length should be the same for both rx and tx.  The environment effects it equally in tx and rx.  Different lengths apply when rx and tx are on different freqs.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Charly86 on August 28, 2014, 06:45:57 AM
Hi,

I'm using CAT5 network cable for my antenna (so single inside) but I don't know what is the best, single wire or multi-wire, any tough on this ?

By the way, I'm searching to thickness the wire and I don't know which one to buy, anyone as a good reference (ebay is okay) of thin wire to use a antenna ?

Thanks for your help.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Felix on August 30, 2014, 09:21:43 PM
Single core 22-24GA works well. Don't use stranded wire.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Charly86 on August 31, 2014, 07:37:07 AM
Felix,

thank's for the tip about stranded wire, so much time I'm asking this question, it's good to know now  ;)

So except the fact that Cat 5 wires are not the thickened one , they're good ?
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: Felix on August 31, 2014, 10:18:51 AM
Yes
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: john k2ox on November 09, 2014, 10:39:16 PM
Electrically or RF wise there is little difference between single and multistrand wire. Practically, single strand is preferred because it retains its shape. Stranded wire is floppy.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: joelucid on September 05, 2015, 06:02:59 AM
Interesting antenna experience: With my Tino (https://lowpowerlab.com/forum/index.php/topic,1269.0.html) I've been experiencing sub par RF performance. RSSI was about 10-12 lower than with a regular Moteino with 9V battery.

I did some research and some experiments and found that:

1.) The Tino had great RF performance if powered from a cylindrical Lithium battery.
2.) Different environments (e.g. sizes of ground plane) sometimes lead to very significant resonance shifts.
3.) Resonance shifts can be compensated for by adjusting antenna length.

So I hooked a 25cm antenna to my 433mhz Tino and started clipping. Turns out the optimal antenna length with CR2450 mounted is 22-23cm instead of the around 17cm that work for a Moteino. At that length RF performance is similar to a Moteino with 9V battery. I would have never expected such a large divergence.

So if you've built elaborate enclosures etc and get RF problems, just attach a LONG antenna and take out your scissors.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: WhiteHare on September 05, 2015, 09:30:43 PM
Would that have to do with item #8 in the OP?  I've read that even the PCB epoxy's dielectric constant can play a role in determining the impedance, so getting all the factors perfectly specified and built seems like it could be quite a challenge.
Title: Re: Antenna Tutorial or Antennas in a Mote
Post by: TomWS on September 05, 2015, 09:37:02 PM
Would that have to do with item #8 in the OP?
Ha!  Good one!  Maybe probably a lot!

Tom