After Hours

802.11: Time to clear up some antenna misconceptions

Antennas for the most part are ignored. That's too bad as they have a huge impact on whether a Wi-Fi link works well or not.

This article is due in part to a recent experience I had with a new client. Needless to say, he had some erroneous notions about how Radio Frequency (RF) propagation worked.

I'll admit, it's hard to visualize how electromagnetic radiation acts when it exhibits both wave-like and particle-like properties, and propagates in three dimensions. Still, understanding the fundamentals is simpler than most people think.

As funny as it sounds, all it takes is a balloon. Many years ago, while studying for my first amateur radio license, one of my mentors handed me a balloon, telling me to inflate it. Ignoring the quizzical look of a sixteen year old, he then proceeded to explain RF propagation.

Isotropic radiator

Understanding RF radiation begins with the concept of an isotropic radiator. The Sun and the results of the Big Bang Theory are both real-world examples of isotropic radiators.

When it comes to antennas, an isotropic radiator is theoretical. So, imagine a RF source emitting electromagnetic radiation in three dimensions with equal intensity and 100 percent efficient. At first, I didn't get it. My mentor told me the spherical shape of the balloon represents an isotropic radiator, with the balloon's surface being where RF radiation stops. OK, that works.

dBi

One more theoretical construct needs to be understood. That is Decibels (isotropic) or dBi. It is the gain in radiation intensity of an antenna when compared to the isotropic radiator. Here is the tricky part. Think of my balloon again and imagine that the air inside is RF energy. There is a set amount, so how does an antenna achieve gain? It does by concentrating the RF energy. Looking at some real-world antennas will help explain.

The ubiquitous dipole antenna

The 15 cm long vertical element you see on most Wi-Fi equipment is actually a dipole antenna. It consists of two elements and is popular because of its omnidirectional radiation pattern. It has approximately two dBi gain over the isotropic radiator. Let's see why that is.

The dipole schematic (courtesy of Wikipedia) hints at how the gain is achieved. Let's take my balloon once again and squeeze it in the center as shown in the image. The air/energy inside the balloon is forced into a different shape. This shape is representative of the dipole antenna radiation pattern.

Notice the balloon is longer, that would be considered gain in directions perpendicular to the axis of the antenna. The amount of energy is the same, it's just being redirected. The following antenna radiation diagram is of a typical dipole.

More power, Scotty

Looking at antenna specifications, you may have noticed that similar antenna styles have different dBi ratings. Ever wonder how that works?

Back to the balloon; it's still simulating a dipole antenna, but let's squeeze the balloon between two pieces of cardboard. Notice how it gets longer, that's additional gain. Did you also notice there is less vertical coverage? The next radiation diagram shows what's going on.

Client's first mistake

This is how my client started getting into trouble. His Wi-Fi router was in the center of the building and offices along the outer walls were not connecting. He bought two antennas similar to the one shown here and was pleased as the offices with weak reception were now connecting.

Guess what, computers on the second floor right above the router lost access to the network. That's because dipole antennas achieve gain by squishing the radiation pattern along the axis of the antenna. They are still omnidirectional, but only in the space perpendicular to the axis of the antenna as shown in the radiation diagrams. So, does he need more gain?

Directional antennas

Enter directional antennas, they are the power houses when it comes to gain. Once again the balloon easily depicts the radiation pattern. Satellite TV dishes are an example of directional antennas.

As you can guess, to increase gain, direction antennas further restrict the radiation pattern. In fact, the pattern is no longer omnidirectional along either axis of the antenna. Notice in the radiation diagram below, there is only one chart. That's because the pattern is the same for elevation and azimuth.

Client's second mistake

I bet you're wondering if I was going to get back to my client. He thought he still needed more power. So on the recommendation of a salesperson, he bought two panel (directional) antennas. Their logic was the building would be fully covered by pointing the antennas in opposite directions.

Can you figure out what problems this caused? Now additional people on the second floor were complaining of lost connections. We know that's because directional antennas radiate less in the vertical plane when compared to dipole antennas. But, why were people on the first floor complaining of slow connections?

MIMO and Multipath

I was remiss in not mentioning the client's network was built with 802.11n equipment. I'll bet you can see where this is going. By using directional antennas, my client lost two of the best features of 802.11n, MIMO and multipath propagation.

802.11n leverages something called multipath interference. It does this by using multiple signal streams and conditioning the disparate feeds into a stronger more reliable signal at the receiver. That's why some of my client's employees whose computers had RF Non-Line of Sight to the Wi-Fi router were able to make a connection. But, that ability goes away when directional antennas are used.

Final thoughts

Setting up Wi-Fi networks so they meet expectations can be challenging. Knowing a bit about RF propagation and how antennas work should be one of the first steps, not an after thought.

For a well-written explanation of basic antenna principals and radiation diagrams, check out Dr. Trevor Marshall's article Antennas enhance WLAN security. I also want to thank the fellow hams at Force 12, Inc for allowing me to use their balloon pictures.

About

Information is my field...Writing is my passion...Coupling the two is my mission.

70 comments
mjd420nova
mjd420nova

There are many, many more things to consider when dealing with microwaves. At the higher frequencies, coax and antennas reduce the actual signal that gets converted to the EM waves. Ideally, the antennas should be more along the lines of an emitter and a reflecting dish. This makes it highly directional. Taking this into account can solve some problems but as already explained, create some too. A quick and simple approach would be to use the power line adapters that take a WIFI signal and attach it to the AC power in a facility. Not a sure cure but would help those outlying areas to get adequate signals for good communication.

brnkumar
brnkumar

Can you please let me know if there are any problems if an internet tower is placed on top of the building? does it have any effect on humans?

bgfores
bgfores

I use 802.11g still just inserted a home-made tin foil parabolic antenna and gained 2 bars of signal strength!

dcmccunn
dcmccunn

My router is in the basement against the west wall and the computer is upstairs against the east wall. Putting a foil covered reflector behind the router seems to intensify the signal. Does flattening the "balloon" in the back push it out the front towards the upstairs? There are several web sites that show ways of making the signal more directional using foil covered reflectors. Search "wireless reflectors"

bowenw
bowenw

Mike, that was an excellent article. The ham that "Elmered" me used the example of a balloon, a donut and a tortilla to demonstrate antenna "gain" to me. I think that example is even more appropariate to demonstrate the issues of signal strength in the vertical plane. One "trick" I've used to help resolve vertical plane issues is to mechanically tilt directional antennas. There are actually some antennas (both panel & dipole) that are designed with electrical "downtilt" or "uptilt" - they are VERY common in cellular and paging deployments. I'd also add that a good look at the RF propogation section of any recent ARRL "Radio Amateurs Handbook" would be a good thing to do for anyone working on the deployment of wireless networks. P.S. I like your call - very appropeiate in the line of work you're in. 73s OM from ex-K8YGT

garyfizer
garyfizer

Ok. But what was your recommended fix. Will you do a part two of this story?

jeslurkin
jeslurkin

Hadn't thought to use a reflector to improve privacy and gain. Thanks.

mike.motes
mike.motes

Having been a RF engineer for many years, the concept of an isotropic radiator has always been difficult to visualize to new students. I used to use an orange as an example, but a balloon works so much better! Consider taking this further, such as explaining the skin-effect in coaxial cable and cable loss, as in many situations these can be more important than the radiator being used.

trcnet
trcnet

Loved the balloons, they make great visuals for the propagation. However, I am unclear as to why adding 2 more omnidirectional antennas to the first floor would affect the second floor connectivity? If they connected before with one antenna, why does adding two more verticals affect the second floor? I am assuming the original router and antenna stayed in place and antennas 2 and 3 were placed at opposite ends of the building? I am also assuming that the original antenna was vertically mounted so that the signal reaching the second floor would be the same whether there was just one antenna on the first floor or if there were three antennas on the first floor. Could you clarify?

mjd420nova
mjd420nova

Michael: Like you I hold an amatuer licence, since 1967, but haven't had any FCC notices yet. When it comes to harmonics, I always believe in filtering. Grounding is the biggest part of any filter network. Elimination can be difficult but careful design of pre and final stages will help to reduce them to a point where filtering can be minor without affecting primary signals at the final stage. I like to use an old military transmitter that had no regard for harmonics and it needs triple stage filters on the output to prevent those spurious signals from even leaving the shack. Working with reflectors helps to reduce signals from going where I don't want them.

Ron K.
Ron K.

I have one wireless user that only has 2% signal using our 802.11n router. It's enough to connect and use the Internet but just barely. I can hardly wait for them to get home so that we can try something different.

Michael Kassner
Michael Kassner

Typically it's not a problem as the power levels are quite low. You will have to tell me if it is Wi-Fi and whether it's 802.11 or 802.16.

brnkumar
brnkumar

Can you please let me know if there are any problems if an internet tower is placed on top of the building? does it have any effect on humans?

Michael Kassner
Michael Kassner

As long as you are not needing omni-directional reception/transmission that will be a big help.

Michael Kassner
Michael Kassner

You just need to remember that the amount of energy does not change. Where you are shoving it can and does depending on the antenna. A member pointed out this link. It is an excellent one and describes how to make a parabolic reflector for this frequency range. Give it a shot: http://www.freeantennas.com/projects/template/index.html But, you have to promise to report back and tell us how it worked.

Michael Kassner
Michael Kassner

Thank you for commenting. I am amazed and encouraged at how many Hams are TechRepublic members. You have pointed out one aspect of microwave technology that I am clueless about. I've read about RF down-tilt, but no one explains how they do that. If you know, please clue me in.

Michael Kassner
Michael Kassner

Hey, Gary What do you think would be needed to fix the situation?

Michael Kassner
Michael Kassner

That means a great deal coming from someone who knows. I agree about cable loss, actually I suspect coax is less considered than antennas. I will have to add that to my list.

bowenw
bowenw

Mike, Some excellent points. Anything you can do to reduce losses between the antenna and the tranceiver helps - too many people scrimp by using cheap cable and connectors and then wonder why the system does not work in real life like it did on paper.

Michael Kassner
Michael Kassner

The antennas were right by the router. If you look at the second dipole coverage diagram, you will see that the vertical coverage is considerably less. In my client's case it was enough to create connection problems.

todd
todd

None of the wifi routers I have used have provided any guidance as to how to adjust the orientation of the various antennae to get the signal where you want it to go. For example my my linksys N router had 2 long thin antennae and one thicker flat one.

Michael Kassner
Michael Kassner

Another Ham, excellent. There are more here than I thought. I got my Novice in 1966. So we are old-timers to be sure. The FCC notice, I was referring to was my first crystal-controlled 40m CW transmitter. I think it had about 10 watts. At the time I lived in Wisconsin and received the warning from the FCC in Seattle. I could believe it was getting out that well.

Michael Kassner
Michael Kassner

You absolutely have to keep us up-to-date with what you have found. I love hearing from others and their experiences. If 802.11n is having trouble, I doubt the other protocols would even associate. Are both ends of the connection using 802.11n equipment?

HAL 9000
HAL 9000

It's defiantly not going to do who ever it lands on any good. Also if the Building isn't up to Specification it could cause a Catastrophic Failure of the building and cause death or injury when the building collapses or is otherwise damaged by the falling Aerial. But in regards to LFEMR [i]Low Frequency Electro Magnetic Radiation[/i] this is as yet unproven as to it's lack of safety. Though if you are directly under the Aerial it's less likely to be as big a problem as if you where 100 feet away from the Aerial exposed to the Full Power of the Radio Transmissions. As Radio has been used for well over 70 years with no documented problems to Human Health the reasonable person would have to say that it's relatively safe. Col

mike.motes
mike.motes

That I have used in explaining the parabolic reflector is to imagine the bulb in a flashlight (the isotropic radiator...sort of) and the reflector it is mounted in (the parabolic dish). The lumens being emitted by the bulb have not increased, but you are now directing the light in a single direction increasing light intensity. You can also visualize why proper spacing between the focal point of the bulb and reflector becomes important (wavelength). 73's

Michael Kassner
Michael Kassner

But, vendors do not give many choices. Besides, not many people would like working with LMR 400.

trcnet
trcnet

Oh, ok, I had made the assumption that the antennas he added were of the same gain and didn't realize that he added hi-gain antennas which would definitely reduce the potential for the signal to reach the next floor. Makes sense now, thank you! Nice article and great diagrams and photos. Nice to see all the old hams (I'm one, also) chime in with comments, too.

Michael Kassner
Michael Kassner

I am not quite sure what Linksys has in mind. I suspect the middle antenna is a panel (directional antenna), not sure though. The vertical antennas are omni-directional and should work together for multi-path propagation. That is if the other end of the link is using 802.11n as well.

Ron K.
Ron K.

Our router is a D-Link DIR-655 and the USB adapter is a D-Link DWA-140 RangeBooster N USB Adapter. The router is here in the house, situated roughly 80-100 feet from where my user, my daughter, has her desktop computer. It's behind the corrugated steel roof of a converted barn. I picture that roof bouncing the signal upwards at an angle. Having the adapter down as low as possible gives the best signal. My memory is certainly faulty but I believe that we had a 25% signal when I'd hooked it up downstairs. I've thought of drilling a small hole thru the floor and splicing in new wiring to lengthen the USB cable to put the adapter downstairs and use Velcro to attach it to the wall. It's still an option. I didn't think that wireless was going to be our solution. I wanted CAT 5e in the ground, strung through a pole-barn over to the computer. I wasn't informed about their schedule so they had the trench dug and refilled for the electrical wiring before I knew it. I don't know what they expected me to do. I decided to take a chance on wireless and I'm thankful that it works well enough that they're satisfied. Repositioning the router and antennas, a few moments ago, had no positive effect. I was just out there to install some software. I updated it in just a few minutes. It's not too slow for me. Then too, I'm used to slower speeds and latency. Our ISP is Hughesnet satellite on the Pro Plus plan. We had to go that route because were out here in the boonies. C'mon cable companies! Bring us some bandwidth! I'm off track. To complete the picture, our 4 wired computers are running on the motherboard's adapters. That was an easy setup. Plug in the cable and go. I'd recommend the D-Link products we have. They're performing well. I'd like to try a D-Link wireless NIC out there for the extra antennas but I think, "if it ain't broke, don't fix it."

dawgit
dawgit

You are correct in the low power radiation EMF. But, If you not sure stay away... In quite a few High Power Antenna sites, it might not be so healthy. There was a reason we used sheep to keep the grass down. ;)

Michael Kassner
Michael Kassner

Thanks, Mike for mentioning it. The only problem with the light example is when you try to explain a dipole antenna and gain in the horizontal plane.

seanferd
seanferd

Although I am no expert, certainly. ;) I could be getting the wrong idea from what I've read so far. The most references I find for this seem to involve cell towers that have perhaps a bit too much elevation over the coverage area. But again, maybe I have taken a skewed sample.

seanferd
seanferd

I was a bit puzzled by the slope in coax which requires adjustment, as it was obviously, as you note, not the same as RF antenna electronic downtilt. I didn't find a good reference for this quickly, but read some stories where the slope was horribly wrong in cabling setups.

bowenw
bowenw

Seanford, the reason for tilt or slope in coax cable is that coax, electrically, is actually a series inductance and a parallel capacitance: a distributed low-pass filter. This is NOT the same as "electrical beam tilt" in an antenna, which is tilting the main lobe of an antenna's pattern while mechanically the antenna frame remains vertical.

Michael Kassner
Michael Kassner

Good information. I have used mechanical tilt before, but not electrical. It seems to me that way would create out-of-phase signals.

mike.motes
mike.motes

But I know of a client that had a LONG run that there was no way of getting around. We wound up using Andrew 1 1/4 Heliax for that one. BTW, it wasn't actually WiFi, but used the same ISM band.

bowenw
bowenw

Mike, LMR400? How about 6" Andrews? Just kidding - but yes, I have run LMR400 & 1/2" Heliax for microwave backhauls.

Neon Samurai
Neon Samurai

I'll have to check my angles. I've one virticle and one horizontal. The horizontal is about 45 degrees twist on the horizontal plane. I was basing it on visible antenna area so I'll have to look at it for polarization instead. Luckily, I have a tiny area to cover so both sticks up would work unlike my previous place that had to blast up through two floors in addition to the basement area.

Neon Samurai
Neon Samurai

On my linksys with two strait and the one flat antenna, I've spread them out as best I can. One stick strait up, the other nearly horizontal and the flat panel at twist/angle somewhere half in-between. This is based on the idea that an R/C plane works better when controller antenna is not pointing directly at the plane (ie. smallest display surface visible). I'm also not doing 11n as only one client node supports it and dd-wrt doesn't give a mixed g/n option.

Michael Kassner
Michael Kassner

Where as the sat link is 1.6 Mb/s as you mentioned. So her connection should not be the bottleneck.

Ron K.
Ron K.

Her wireless signal comes from my DIR-655 802.11n router connected to the satellite. There is no other Internet access available here in the boondocks. That's why we've been on the satellite for the past 8 years.

Michael Kassner
Michael Kassner

A Wi-Fi connection should blow away any connection from a sky bird. To be honest, it should be better than any connection to the Internet. It is fixable, but we need to know more details.

Ron K.
Ron K.

Our service plan from Hughesnet is the ProPlus plan with only 1.6Mbps Down and 250 Kbps Up. I've never checked her speed out there but it seemed slower than what we have in the house using Cat 5E. I'm definitely open to suggestions.

Michael Kassner
Michael Kassner

But are the users happy with the bandwidth? I have more tricks if they are aren't.

Ron K.
Ron K.

After reading this discussion I dug out an old DirecTV satellite dish, half an hour ago and put the wireless USB dongle near the focal point. We have 22% signal strength now. I appreciate your advice and this blog, Michael.

Neon Samurai
Neon Samurai

.. in this case, it's a specialty tool so doing an antenna hookup rather than cradling the USB dongle in the bowl would do it.

Michael Kassner
Michael Kassner

The only thing you will lose is antenna diversity for 802.11a/b/g. If you are using 802.11n, you will have to decide if the gain benefit outweighs the loss of multi-path propagation.

Neon Samurai
Neon Samurai

I've been meaning to build one of those myself after I find a good USB wifi dongle. Should be able to use the same strainer for both wifi and bluetooth dongles.

Ron K.
Ron K.

I may try a strainer. I like the looks of that. I'll post back to let you know if I do it and what the results are.