A closer look at the strength and shape of 2.4 GHz radio signals

[redbird300]

A closer look at the strength and shape of 2.4 GHz radio signals

A) Introduction:

Radio signals are perhaps the most important element in our hobby, but also the least visible, and by consequence the most ignored and misunderstood element. This article is specifically about 2.4 GHz signals, and tries to reveal a few mysteries about one of the dark sides of this subject, being the strength and shape of the signals, as well as the way of functioning of the actual available systems in terms of signal transmission.

Allow us to state VERY clearly that this post has no intention what so ever to decide which system would be better or worse. We know very well that there are a whole lot of Spektrum and Futaba fanboys out there for example, but we only used these brands as an example because they are very popular, and because they use different ways of operating. Which makes them interesting to take a closer look at, from a technical point of view. We also don’t intend to write down a complete technical review about all transmitters (which would be a huge job), but will only tell a few things about the transmission principle and the strength of the signals. By consequence, some aspects won’t be covered, like for example the different ways in which these systems try to provide redundancy, differences in speed or resolution, etcetera. And absolutely no comparison in terms of price or the level of service. What we will cover is the strength of the signals, the way they are transmitted and the according data, to get an idea about the transmission power levels of R/C transmitters.

Why this approach ? Well, while you’ll find information easily about a lot of these subjects, it appears to be harder to find numerical data about the signals themselves. One reason might be that you need a suitable 2.4 GHz spectrum analyzer. The following results in chapter “E. Measurements” were obtained with a rather simple, portable analyzer with these specs:
. Amplitude resolution: 0.5dBm
· Dynamic range: -110dBm to 0dBm
· Absolute Max input power: +5dBm
· Average noise level (typical): -105dBm
· Amplitude accuracy (typical): +-3dBm


B) First a small piece of essential theory:

We’re talking about the frequency band from 2.4 up to 2.483 GHz. The two most important methods used for frequency modulation in this band are “frequency-hopping spread spectrum” (FHSS) and “direct-sequence spread spectrum” (DSSS). A few real-world, everyday examples: Bluetooth uses FHSS modulation, while wireless USB and 802.11b/g/a (better known as Wi-Fi to many of us) use DSSS.


DSSS:
“Direct-sequence spread spectrum” is a modulation technique in which the transmitted signal uses more bandwidth than the modulated information signal. The information stream is divided into small chunks. Each of these chunks is divided over one or a few fixed frequency channels. A piece of data signal ready to be transmitted, is combined with a mutual code that distributes the data. This redundant code helps to resist signal interference, and to reconstruct the original data if some data would be damaged during the transmission. An example, with data distributed over 2 fixed channels:

Read more here: http://en.wikipedia.org/wiki/Direct-sequence_spread_spectrum

FHSS:
“Frequency-hopping spread spectrum” is a method to transmit radio signals by rapidly shifting a carrier signal among many different frequency channels, using a pseudo random order known by transmitter and receiver. This offers a few benefits compared to a fixed frequency:
- FHSS signals don’t suffer from narrowband interference. The process of reconstructing a spread signal after the transmission causes interfering signals to be pushed to the background.
- FHSS signals are difficult to intercept. A FHSS signal shows itself as a simple increase of background noise to a narrowband receiver. A listening station would only be able to intercept the transmission if it would know the pseudo random order.
- FHSS transmissions can share frequency bands with many types of normal transmissions with minimal interference. The spread spectrum signals add only a small amount of noise to the narrowband transmissions, and the other way around. The result is a more optimal use of the bandwidth.

An example:

Read more here: http://en.wikipedia.org/wiki/Frequency-hopping_spread_spectrum


The fundamental difference between DSSS and FHSS in a simplified diagram:

DSSS/FHSS hybride:
There are also hybrid systems, which combine the advantages of both technologies. An example is the Jeti system. It uses a DSSS signal, but which shifts rapidly over 16 different channels, using a fixed scheme. It looks like this:

Pictures by courtesy of www.rcmodelreviews.com and www.hometoys.com

C) The units:

Signal strength can be expressed in dbm or milliwatt, here’s a conversion table:

dBm --- milliwatt

20 ----- 100
15 ----- 32
10 ----- 10
5 ------- 3,2
4 ------- 2,5
3 ------- 2
2 ------- 1,6
1 ------- 1,3
0 ------- 1
-1 ------ 0,79
-5 ------ 0,32
-10 ----- 0,1
-20 ----- 0,01
-30 ----- 0,001
-40 ----- 0,0001
-50 ----- 0,00001
-60 ----- 0,000001
-70 ----- 0,0000001
-80 ----- 0,00000001


The formulae to calculate x [dbm] as a function of P [Watt] is:
x= 10log10(1000P) or x = 10log10P + 30

And the formulae to calculate P [Watt] as a function of x [dbm] is:
P = 10(x/10)/1000 or P = 10(x-30)/10

D) What about 2.4GHz R/C transmitters:

The modulation techniques which are used nowadays by some major manufacturers:

· Spektrum and JR use (amongst others) DSM2, that’s a DSSS system. The available frequency band between 2.4 and 2.48 GHz is divided into 80 channels, each 1 MHz wide, and of which exactly two free channels are fixed when booting the system. Only these two channels will be used during that session: the transmitter searches for two free channels when powering up, and as soon it has found these, the “listening” receiver (which only reacts on the right GIUD code that has to be sent by the transmitter) will lock on the same channels. But one channel can be used by multiple transmitters at the same time, as the channel isn’t occupied the whole time, so don’t think that 40 transmitters would be the maximum allowable. Note: these brands use receiver antennas equal to a quarter wavelength which makes for decent sensitivity, that’s around 30 millimeter. We know that wavelength = speed of light in air divided by the central frequency, which gives us roughly 299.000.000 m/s divided by 2.440.000.000 /s = 0,122 m. A quarter of this is slightly over 30 mm. Now you know why these little Spektrum receiver antennas are about 30 mm long.

· The newer DSMX protocol from Spektrum uses 23 randomly selected channels.

· DMSS is a new protocol from JR, and stands for Dual Modulation Spectrum System. This combines the best of both worlds, read “FHSS and DSSS”. FHSS is very resistant to interference because of the frequency swapping method, and DSSS makes for fast response times. In a nutshell: also a hybrid system, just like Jeti.

· Futaba uses FHSS: the signal changes its frequency several hundreds of times each second. The order is determined by the GUID code of the system (Globally Unique Identifier). Some people get confused when looking at the Futaba gear: Futaba sells “FHSS” and “FASST” receivers, but both use FHSS technology. You need to know that “FASST” is a brand name used for their more sophisticated equipment, which implement the FHSS technology in a different way. As a consequence, both systems are not compatible.

· Multiplex M-LINK, Hitec Aurora and Graupner HOTT also use some kind of FHSS implementation.


E) The measurements:

E.0) We’ll always measure using the “MAX” and “AVERAGE” mode, allowing us to see what (fast moving) peaks have occurred. To be sure that there are no other disturbing signals during the measurements, we’ll start with a test measurement in a suspected “clean” environment, meaning that there are no disturbing 2.4 GHz sources like Wi-Fi, Bluetooth, etcetera. We notice an average noise level of -95 dbm, and a peak level equal to -90 dbm, which we’ll consider as normal from now on, given the used equipment:




E.1) To get a little feeling for the following values, we’ll first measure the signal output of a common Wi-Fi router, as many of you own and operate at home. These devices have a real life power output of around 30 to 100 mW, being +15 to +20 dbm, so (very !) roughly the same magnitude as the normal 2.4 GHz R/C transmitters, which are rated 100 mW theoretically. Here you see a D-Link DIR-825 in action: we measure peaks of -15 dbm of 0,03 mW, as close to the antenna as possible:



Take care though, the previous mentioned power values of 30 to 100 mW are TRANSMITTING power values, and our measurements are captured signal strength, which is a totally different thing, but they relate in a linear way. Note: 100mW is the legal maximum in many countries, spread over a maximum of 14 different channels who’s center is 5 MHz apart.

Another interesting test shows that the active Bluetooth module (read FHSS) of for example a HTC TOUCH HD GSM equipped with Bluetooth v2.0 class 2 (theoretically 4 dbm or 2,5 mW transmitting power and a nominal range of 10 meters) emits a maximal signal strength of about -25 dbm or 0,003 mW. This low value sounds logical, as this technology is meant to cover only small distances.

And now a number of measurements using R/C transmitters. We’ll look for the maximum value that can be measured, by holding the sending antenna of the transmitter and the receiving antenna of the spectrum analyzer as close to each other as possible, and in an optimal orientation, to obtain comparable values:

E.2) Spektrum DX8 in range test mode, meaning a limited power output: -20 dbm or 0,01 mW. In this mode, the transmitter is supposed to maintain contact up to 30 meters in open air, according to the Spektrum manual. This makes sense when comparing that value with the weaker Bluetooth module and the stronger Wi-Fi router. We clearly notice the two frequency peaks that are so typical for the Spektrum DSSS system:




E.3) Spektrum DX8 at normal transmission power: we measure two peaks equal to -10 dbm or 0,1 mW. The signal is now clearly stronger than the signal of the Wi-Fi router, no surprise here, as we know that the DX8 should have a superior range compared to the router:




Note: at 1 meter: -25dbm / at 2 meter: -35dbm / at 3 meter: -45dbm / at 20 meter: -50dbm.

Knowing these values, it now becomes easy to check if a given R/C transmitter produces the transmitting power it should. We want to cover at least several hundreds of meters, and preferably even more.

E.4) Even more interesting is a measurement using two Spektrum DX8 transmitters at the same time, and measure close by both antennas, with the transmitters at full power:






We clearly recognize four instead of two frequency peaks now, as each transmitter occupies two channels. The two middle channels are still those of the transmitter above in E.3), switched on first. The two outer peaks appeared as soon as the second transmitter got switched on.

E.5) A cheap Spektrum compatibel transmitter, the E-Flite MLP4DSM: this one only outputs -15 dbm, instead of -10 dbm as his big brother does, the DX series. That’s about the same power as our Wi-Fi router, and this makes sense again. This is the kind of cheap transmitter that you get when buying a RTF Blade mSR:







E.6) Futaba T10CP with a TM-10 FASST module in range test mode, meaning a limited power output: -25 dbm or 0,003 mW.





E.7) Futaba T10CP with a TM-10 FASST module at normal power: we notice (again) an increased signal strength over the entire range as above. Remember that this is FHSS, shifting rapidly between many channels. The value is higher now of course: -10 dbm or 0,1 mW. Exactly as much as the Spektrum DX8 at normal power.




Note: at 1 meter: -30dbm / at 2 meter: -35dbm / at 3 meter: -45dbm / at 20 meter: -50dbm.

[redbird300]

E.8) Now we switch on a Spektrum DX8 and a Futaba T10CP at the same time. We deliberately put the Spektrum close by and the Futaba at 1 meter away from the analyzer antenna this time, resulting in a weaker signal from the Futaba, so that we’re able to see the different signals more clearly. Otherwise, the signals would be of the same magnitude, making it impossible to distinguish the separate radio waves. We now notice the two large frequency peaks of the Spektrum transmitter, and a number of smaller peaks over the total frequency band from the Futaba FHSS system:




You can see this even better in the “MAX-HOLD” mode of the analyzer:




F) Summary:

The (slighly rounded) measured signal strengths:

Wifi router (DSSS): -15 dbm 0,03 mW
Bluetooth class 2 (FHSS): -25 dbm 0,003 mW
Spektrum DX8 range test (DSSS): - 20 dbm 0,01 mW
Spektrum DX8 (DSSS): -10 dbm 0,1 mW
Futaba T10CP range test (FHSS): - 25 dbm 0,003 mW
Futaba T10CP (FHSS): - 10 dbm 0,1 mW

Hopefully this gives a little more insight in these invisible radio signals that are so important for our hobby. The development doesn’t stop, and other frequencies are also used, like video transmitters at 5,8 GHz for FPV. But that’s another story.

Just a last tip: don’t ignore the range tests, they really are essential, even if the possibility of interference has become very unlikely using 2.4 GHz systems. A whole lot of other factors are also into play. Even brand new transmitters can fail, as I had to discover to my own shame. “Quality Control” is as only a little sticker, not an absolute certainty.

P.S.: Sorry for any spelling mistakes or grammatical bad sentences, I’m not a native English speaker. Feel free to add or correct whatever you wish of course, this is only my 2 cents. Enjoy.

[Commodore8888]

This whole page.....is made of win.

For jamming resistance I remember reading a test where an EE dissected the signals of the FAAST and other FHSS systems and found that they were constantly agile DSSS systems, the hybrids you mentioned which JR is switching to. DSSS systems can quirk at times if they happen to select two close channels or if there are a ton of operating radios. Hopefully DMSX fixed that, but in the contest of jamming resistance, the hybrid systems have proven to be more resilient. I wonder if all high end RC systems will go that way eventually. Maybe a firmware flash at some point?

Really good read though, I enjoyed it.

[BaconRaygun]

Apparently redbird300 has tigers blood! Total warlock status...

[Redlight11]

Not enough info. Care to elaborate? jk


Thanks for the info. I will do my best to wrap my mind around it.

[dyuen888]

Okay... armed with this information... What is the best antenna position for a 2.4 TX ?

Pointed straight out as in the picture?
Folded with the point off to one side of the TX ?
Folded with the point straight up?

This asumming the TX is level to the ground while holding it.

[redbird300]

Originally Posted by dyuen888

Okay... armed with this information... What is the best antenna position for a 2.4 TX ?

Pointed straight out as in the picture?
Folded with the point off to one side of the TX ?
Folded with the point straight up?

This asumming the TX is level to the ground while holding it.

All you wanna know about that is here for example:
http://henrycountyrc.com/blog/2009/1..._dynamics.html

[Mike2112]

Thanks redbird300 and the United States military!

[thrak76]

Thank you for taking the time to post this valuable information. I really appreciate your efforts.

[nocontrol1]

Wow! Nice post. Thanks!

Rob D.

[desertstalker]

Any chance of a test using DSMX? or would that just look like the futaba one?

[jperkosk]

+1, Raf. We won't see the channel hopping sequence on the spectrum analyzer, of course, but maybe we could see how channel width compares between F and S when both are in FHSS mode?

Excellent job, as always

[redbird300]

Thank you fellow freakers.

I'll post a graph of a DSMX system soon, in fact I already measured it, but didn't capture a graph until now. With the analyzer in hold mode, you'll see the 23 peaks, as the system is shifting between them all the time. So yes, it looks a little like the Futaba graph.

And it eliminates the problem where a Spektrum DSM2 system that has chosen two close by frequencies at boot up time, could theoretically be locked out by a stronger, continuous source, wide enough to overlap the two channels. Very unlikely, but doable with the right equipment.

[stevec]

Originally Posted by dyuen888

Okay... armed with this information... What is the best antenna position for a 2.4 TX ?

Pointed straight out as in the picture?
Folded with the point off to one side of the TX ?
Folded with the point straight up?

This asumming the TX is level to the ground while holding it.

The picture attached by the OP is a fair representation of the way the signal radiates.
it is a wave which comes out of the sides of the aerial.

by bending the tx antenna to 90 degrees you are maximising the amount of wave going towards your rx.

imagine a snooker table and take your cue as usual and push it the length of the table. the tip may or may not hit a ball as you run the cue down the length.
if you turn the cue 90 degrees and run it down the length of the table you will hit everything on the table.

with radio obviously there are other factors conspiring with gravity to end your day badly, like carbon side frames and other signal blockers on the airframe but you stand a better chance if the whole of the wave is pointed at the rx rather than a beam of RF from the end of the antenna.

hope this helps.

great post by the way.

cheers
Steve

[nexgen]

Originally Posted by Commodore8888

This whole page.....is made of win.

[rdlohr]

Awesome post. I found it to be very clearly written and understandable. Thanks for taking time to compose and publish this for us!


I seldom take the time to rate a thread but you get a 5 star!

Rick

[Redlight11]

So much good info...I vote to make this post a sticky.

[WillJames]

Thank you VERY much for the thread. This is some very cool info to actually see presented so easily for people to understand. Thank you.

[cudaboy_71]

WOW!

can you provide a TLDR summary for the ADD in the crowd? ;-)

[Cailid]

Originally Posted by dyuen888

Okay... armed with this information... What is the best antenna position for a 2.4 TX ?

Pointed straight out as in the picture?
Folded with the point off to one side of the TX ?
Folded with the point straight up?

This asumming the TX is level to the ground while holding it.

http://www.archeli.com.au/forums/sho...d.php?t=136223