Demystifying the Dark Art of Antenna and Frequency Selection

by cwozny | July 24, 2012 | (9) Posted in Tips

DISCLAIMER: Transmitting/receiving on some of the frequencies listed below and/or transmitting above certain powers on certain frequencies may be illegal in your country. I claim no responsibility for this so check with your local government (FCC for the U.S.) for any restrictions.

I decided to make this guide because I was tired of seeing people speculate as to what the range was of their FPV setup (or any RF setup) and wanted to find some cold, hard technical facts. This will be totally theory-based as I do not have the resources to try each and every combination I will list below. These calculations are based on an unobstructed line-of-sight signal with no electronic interference. However, the real world presents many variables that result in less-than-perfect wireless performance, such as mismatched impedance, electronic noise, building obstructions, reflected signals, etc. This means that you can use these values as an absolute maximum for transmission. Any further and the signal-to-noise ratio will drop below 1.

All of the work I did was done in MATLAB and I will be including the (fairly simple) source files so you all can recreate it as well. It was nearly completely based off of Friis' transmission equation.

I'm going to preface this by saying that essentially any range for an FPV setup is possible, but range grows logarithmically with transmission power and antennas with high enough gains may not even be realizable. Alas, we can still get some pretty serious range out of some simple antenna setups. When I was thinking about designing my FPV setup, I tried to think what the limiting factors would be. I came up with:

  1. Range of my Spektrum DX6i transmitter (about 9/10 of a mile)
  2. Range of line-of-sight view in case of FPV failure (about 1/4 of a mile)
  3. As a corollary, I thought I might also want to integrate an autopilot into my FPV setup so I figured I wouldn't stop with those distances listed above and try and push the envelope.

For the figures below, I used the following gains as generalizations for different types of antennas. Also, I couldn't find specific values for receiver sensitivity so I chalked it up to manufacturer quality. I went with -81 dBm for it. I also included a 15 dB fade margin to make the values a bit more realistic. The following values were retrieved from spec sheets of commonly purchased antennas:

  1. Skew-Planar Wheel: 1.0 dBi
  2. Cloverleaf: 1.2 dBi
  3. Dipole (Rubber Ducky): 3.0 dBi
  4. Patch: 8.0 dBi

The great part about including the source means that if you want to try out other specific values you have for antennas then you can feel free to do so! If you don't have a MATLAB license, the source code should still compile in GNU Octave/PyLab/SciLab/etc. Here's a link to another calculator I used to estimate/verify transmission range which includes Fade Margin calculation and signal loss due to cable impedance: RF Link Budget Calculator

 

Finally, enough talk! Let's see some pretty plots!


Here we have an example of a Cloverleaf transmitting on the aircraft and a Skew-Planar Wheel receiving on the ground station. The range is not that great, but it is omni-directional and circularly polarized:


Next up we have an example of a Cloverleaf transmitting on the aircraft and a circularly polarized patch antenna receiving on the ground station. The range is significantly better than the previous example, but you always have to have the patch antenna pointing towards the aircraft or else you'll get 26 dB loss:


Now here we have two dipole antennas. Keep in mind these are omni directional, but the radiation pattern means they both either have to have the same polarization (vertical and vertical, horizontal and horizontal, etc) otherwise there is a 26 dB loss:


A linearly polarized dipole antenna with a linearly polarized patch antenna:


I also found some high end directional antennas that claimed 16 dBi so I wanted to include one of those in use with a Cloverleaf:

And finally, just out of morbid curiosity, a high-end circularly polarized patch antenna transmitting to another high-end circularly polarized patch antenna!



After seeing all of these findings, the ideal setup (in my opinion) would be a Cloverleaf transmitting to a Skew-Planar Wheel receiver (and a Circularly Polarized Patch Antenna if you have a diversity controller.)


I hope that everyone enjoys this article and finds it informative and also that I've fulfilled one of the three Es from the FliteTest mission statement! If you have any questions or think I made a mistake (I more than likely did) feel free to let me know in the comments and I'll do my best to remedy them!


COMMENTS

colorex on July 24, 2012
Looks like 5.8 is quite crappy for range... How does it compare in image quality? I suppose it can transmit more data somehow? Higher resolution maybe?
Log In to reply
cwozny on July 25, 2012
I've heard the same thing, I'm not 100% sure of it. I know that composite video signals are modulated on UHF which is 300 MHz to 3 GHz. 5.8 GHz is technically under the Nyquist sampling rate so there is signal aliasing, just as there is with 900 MHz, 1.3 GHz, and 2.4 GHz. Composite video signals are all at the same resolution 640x480 so 5.8 GHz wouldn't give a higher resolution. With 5.8 GHz there is a lot more atmospheric distortion than the lower frequencies which will cause picture degradation. It's really a toss up so you could lean either way as to which frequency has a better video quality. Even actual video footage isn't definitive in terms of quality.
Log In to reply
colorex on July 25, 2012
So it looks like it's 1.2 - 1.3 for me... more precision in homemade antennas, better range, better going through trees and humidity...

-cellphones prevent me from using 900MHz
-radio gear is 2.4
-5.8 looks bad at range...

only drawback is that HK doesn't sell this frequency...
Log In to reply
cwozny on July 25, 2012
You can find those frequencies at http://www.rangevideo.com. It's a company based out of Florida.
Log In to reply
colorex on July 25, 2012
Thanks! I'll see what works out for me. I live in Ecuador so I need intl shipping.
Log In to reply
ycopter on August 3, 2012
cwozny, great article that has got some lively debate going! In my experience your models for Cloverleaf, Skew Planar and patch antennas are conservative. I did some testing over 6km line of sight 1.3Ghz 400mw and got some surprising results. Take a look at http://www.youtube.com/watch?v=T3vwgz1paVo

In the real world there are too many variables to model this accurately (receiver sensitivity, antenna match, feedlines, propogation etc, etc)but your article is a great starting point.
Log In to reply
lobstermash on August 1, 2012
OK, this is interesting theory, but real observations don't reflect those curves in the slightest. I'm a scientist and pretty well versed in modelling, and if a scientist came up with a model that was that far removed from real observations, they'd tear it up and start again. I'd be very curious to get people to plot their observations of distance from transmitter, altitude, topography, terrain (eg houses, trees etc.) tx power, antenna setup and atmospheric conditions (temperature, humidity, wind direction and strength, cloud cover etc.) at the very least. A project like that, collecting data from as many people as possible, would give a much better indication of the relationship between transmitter output, frequency used and effective distance.
Log In to reply
cwozny on August 1, 2012
Let me start off by saying I know that this is only theory. Which piece of information do you disagree with? With the higher dBi antennas, I'm a bit skeptical as well, I have a feeling that there is an inversely proportional relation between receiver sensitivity and antenna gain. If you were to tell me that then I would agree wholeheartedly and fix the plots in my post. Obviously the last plot of two 16 dBi antennas is highly suspect, but I've seen white papers where people have tried LoS communication across a valley in these frequencies. With regards to the first couple plots, however, I think they are spot on as they match the transmission distances I've gotten in the past. I lumped in a generic 15 dB environmental loss and I think the data correlates well to real world usage.
Log In to reply
lobstermash on August 1, 2012
I have to say that you've inspired me to collect a heap of data on the subject and will do some statistical analysis to try and identify some trends and hopefully indicative range curves. The ones you have here far under-shoot the real life performance of both 2.4 and 5.8Ghz systems, for which there are many documented flights of 2-5km with cp antennas on 200mW systems.

It'd be great if you want to collect data for me. I've started a thread under FPV in the forum, and I'll be posting the parameters and measurements I need for some boffin-esque statistical analysis on representative (hopefully) data sets for the range of different systems being used.
Log In to reply
cwozny on August 1, 2012
Once you collect the data I'd love to use it to refine my model!
Log In to reply
lobstermash on August 2, 2012
Sure thing! I'm keen to get a bit of a guide happening for both experienced and newbie (I'm the latter) FPVers on the best setup types for different conditions, topography and terrain. A lot of the people I've seen advice from in forums etc. have 'expertise' in one stream/frequency/setup, and their assumptions about stuff outside their niche don't match up with other people's reported experience.
Log In to reply
lobstermash on August 2, 2012
Here's the link to a Google Spreadsheet so you can add your data: https://docs.google.com/spreadsheet/ccc?key=0ApqXMzvBbsbddHdkTGpWWFV6TFBwYmxVWkJpNVB3d1E#gid=0
Log In to reply

You need to log-in to comment on articles.


Demystifying the Dark Art of Antenna and Frequency...