• Ever wondered why we use 2.4 Ghz and 5 Ghz frequencies for 802.11b/g and 802.11a Wi-Fi systems ?

    Why not 300 Ghz or 22 khz ?

    It???¡é?¡é?????¡é???¡és all to do with what???¡é?¡é?????¡é???¡és called the ???¡é?¡é?????¡­?¡°Microwave Window???¡é?¡é????????. Many things in nature exihibit noise properties ???¡é?¡é?????¡é?€?? lightning, sun-spots etc. Movement of electrons in a conductor due to heat [ even with no battery or AC source applied ] generates random noise.

    The planets and stars generate noise and some of it is directed towards earth, where it can be picked up by certain types of antennas. This noise will interfere with signals which happen to be at those frequencies, which is not good.

    Ever listened to the 3- 30 Mhz Short Wave band [ HF ]. ???¡é?¡é?????¡­?¡°This is The BBC Calling From London???¡é?¡é???????? etc???¡é?¡é???????|. ? Many of the crackles and clicks you are generated from noise originating zillions of miles away in space !!

    Between about 1 Mhz to 1000 Mhz, this Galactic Noise is present. It starts off high at the lower frequencies and drops rapidly to a practical limit around 1 Ghz. We call this the lower limit of the ???¡é?¡é?????¡­?¡°Microwave Window???¡é?¡é????????.

    If you look at the picture you can see that the noise temperature [ just think noise interference ] drops from 1 Mhz to 1 Ghz [ the x-axis scale is logarithmic ???¡é?¡é?????¡é?€?? yes, those dam**d dB???¡é?¡é?????¡é???¡és and things again !! ]. So that???¡é?¡é?????¡é???¡és the lower limit. Attenuation also follows the same curve.

    Imagine a child on a swing, and you want to push them. If you give a sudden shove, the swing will shake violently and the kid will probably fall out. However if you give that swing a nice gentle push, then wait till it comes back to you and give it another gentle push at just the right time, the swing will go up higher. This is called ???¡é?¡é?????¡­?¡°resonance???¡é?¡é????????. Injecting energy at just the right amount of time with just the right amount of force. Everything in nature has a natural frequency of resonace.

    The Tacomo Narrows bridge was hit by winds at just the right frequency and amplitude and the bridge shook itself to bits.

    During Napoleon???¡é?¡é?????¡é???¡és time, soldiers crossing a small bridge [ ???¡é?¡é?????¡­?¡°Left, right thump thump, left right thump thump???¡é?¡é????????] actually caused the bridge to collapse by hitting it???¡é?¡é?????¡é???¡és natural frequency of resonance. This is why soldier???¡é?¡é?????¡é???¡és are told to ???¡é?¡é?????¡­?¡°Break Step ???¡é?¡é?????¡­?¡°when crossing a bridge. There are some nasty weapons out there for crowd control which use sound waves to hit the natural frequency of oscillation of the internal organs [ causing some terrible effects ].

    When you see the steam coming off the pad when the Space Shuttle goes up, that does not come from the engines. They burn hydrogen and oxygen. That burns clean. On the first STS mission, sound waves got set up between the orbiter???¡é?¡é?????¡é???¡és main body and the launch tower, standing waves [ just like our mismatched impedances between 50 and 75 ohm cable ] were set up and the shuttle came within a foot of hitting the gantry. The sound energy was so intense that the wildlife within a mile radius was all killed due to internal damage. That steam actually comes from huge amounts of water used for the sound suppression system.

    The atmosphere contains gases such as water vapor and nitrogen. A water molecule contains a hydrogen atom and an oxygen atom. If these are hit with energy at or near 22.5 Ghz, they will resonate [ like the kid on the swing ], absorb energy and give out noise. From about 10Ghz upwards, atmospheric gas absorption and resonance becomes important.

    If you look on the chart you can see that the noise [ there is another chart that looks the same basically for attenuation ] increases very rapidly from 10 Ghz upwards.

    This is the upper limit. So, we now have a window between 1 and 10 Ghz where noise and attenuation are at a minimum. That window is called the microwave window, and that is the main reason why we use the frequencies that we do in Wi-Fi systems outside.

    Inside is a slightly different story, but so much equipment was already [ ???¡é?¡é?????¡­?¡°on-shelf ???¡é?¡é?????¡­?¡°] at those frequencies that it made sense to use them. All things being equal, the higher the frequency , the higher the bandwidth available for transmission.

    As the 2.4 Ghz and 5 Ghz bands get busier and busier, WMAN designers are forced to go higher in frequency. However, although higher bandwidths are available, attenuation and noise go up dramatically. It is not just a linear jump, it is a logarithmic jump. The losses of a 20 Ghz radio link are much, much higher than those of a 2.4 Ghz system. Special design techniques have to used.

  • dave1234 Escribi?3:

    All things being equal, the higher the frequency , the higher the bandwidth available for transmission.


    Again, excellent information. You had one sentence in particular that caught my attention.

    A colleague and I have gotten into a friendly discussion concerning your statement.

    I state that the higher the frequency, the more actual data throughput that can be achieved. He contends that it isn't the frequency that matters, it is the channel width and modulation. So, given a different frequency, but the same channel width and modulation, the data rate and throughput would be equal.

    I've researched the topic and it seems that in digital wireless communication, he is correct. However, your statement seems to differ. It seems as if the difference is in digital vs. analog???

    We do appreciate having an RF (and maybe Wi-Fi as well!) expert here on the boards. Some people such as myself that grew up on the wired networking side and have a general weakness in understanding RF.



  • Hi Grant

    Many Thanks.

    " He contends that it isn't the frequency that matters, it is the channel width and modulation. So, given a different frequency, but the same channel width and modulation, the data rate and throughput would be equal."

    He is right. I didn't want to elaborate on that issue too much as it opens a bunch of other issues and I needed to get to the telly to watch George in Seinfeld have a bed built under his desk at the Yankees Office !

    Anyways, am back from the telly after laughing so much my sides hurt. [ I watch all the re-runs of Seinfeld and Frazier every night before sleep. Have stopped watching the news before bed, and sleep much better ].

    This is a bit tricky and may need a few reads. [ To get the quick answer, jump down to the section called ITU Bands ]

    Lets take the case of a radio [ doesn't have to be 802.11 ] which has a number of modulation schemes available and is capable of taking in a number of data rates, say 1Mbit/s and 10 Mbit/s of customer data. Let's also assume that it produces an exact square output [ impossible due to a number of issues I'll put in a another post, but makes it easier to understand ]. This radio is to be able to change the center frequency of the carrier signal. It can change it to 20 Mhz or 20 Ghz.

    Let's also assume that we have two nice spectrum analyzers which can look at the square output [ again would have sidebands etc in reality ] of the radio.

    The last thing is that when we have a 1 Mbit/s data input, we get a 10 Mhz output width of the spectrum and when we have a 10 Mbit/s data input, we get a 100 Mhz output [ doesn't quite work like that, but it'll help make things easier ].

    We take one radio and set it to 50 Mhz center frequency.

    We take the other radio and set it to 50 Ghz center frequency.

    We now feed in a 1 mbit/s data stream to both.
    On the screen of the spectrum analyzers [ one ???¡é?¡é?????¡­?¡°tuned???¡é?¡é???????? to 50 Mhz and the other to 50Ghz ] we would see the same 10 Mhz wide output.

    The first analyzer would have a center frequency of 50 Mhz, with ???¡é?¡é?????¡­?¡°half???¡é?¡é?????¡é???¡é the signal below that value and half above it. The spectrum would run from 45 Mhz to 55 Mhz [ 10 Mhz of bandwidth ].

    The second analyzer would have a center frequency of 50 Ghz with ???¡é?¡é?????¡­?¡°half???¡é?¡é?????¡é???¡é the signal below that value and half above it. The spectrum would run from 49.995 Ghz to 50.005 Ghz [ 10 Mhz of bandwidth ]

    The same data rate has given us the same output bandwidth on the same type of radio outputs, the only difference is that the center frequency has changed.

    If we now change the input data rate to 100 Mbit/s, the output bandwidth will change in both cases to 100 Mhz wide [ Total width ???¡é?¡é?????¡é?€?? half below the center frequency and half above ].

    The first analyzer would have a center frequency of 50 Mhz, with ???¡é?¡é?????¡­?¡°half???¡é?¡é?????¡é???¡é the signal below that value and half above it. The spectrum would run from 0 Mhz to 100 Mhz [ 100 Mhz of bandwidth ].

    So this means that the same radios with the same modulation type and same data rate will give the same output bandwidth even though the center frequencies are different.

    ITU Bands

    So why did I mention about higher bandwidths being available at higher frequencies. The bandwidths I was talking about here, are not individual channel bandwidths, but ???¡é?¡é?????¡­?¡°total band bandwidths???¡é?¡é????????

    This needs a little explaining.

    Radio frequencies run from very low frequencies in the khz range to very large frequencies in the Ghz range.

    These frequencies are divided into bands as per the following:

    Let???¡é?¡é?????¡é???¡és say someone is allowed to operate in the VHF band [ Local FM stations, police etc ]. From the chart we can operate from 3 Mhz to 30 Mhz. This is 27 Mhz of bandwidth. Can we just send out a signal with 27 Mhz of bandwidth ? No, because so many people want to use it, it is regulated and you would be allocated a very small CHANNEL BANDWIDTH. The total ???¡é?¡é?????¡­?¡°BAND???¡é?¡é???????? BANDWIDTH is 27 Mhz, and the channel bandwidth is small, in many cases in khz. This band is jam packed with users and each user has a fairly small bandwidth.

    Now lets zoom up to the SHF [ Super High Frequency ] band. This is where 802.11 lives. Can we just start sending out signals as we wish ? No, some of the band from 3 to 30 Ghz is reserved [ licensed ] for satellite, weather radar etc. However we are given certain bands like the 2.4 to 2.4835 Ghz and UNII bands for our use. Our bandwidths are not bad [ a few tens of MHz ]. Better than VHF. Because we have more band [ in this case SHF band ] bandwidth, we can give users more channel bandwidth as compared to VHF. We could have everyone operating at very low bandwidths and have zillions of channels, but that would defeat us in hoping to get higher data rates, as these are proportional to bandwidth.

    Now let???¡é?¡é?????¡é???¡és look at EHF [ Extremely High Frequency ]. This band???¡é?¡é?????¡é???¡és bandwidth [ sounds awful but it???¡é?¡é?????¡é???¡és the only way to say it, unless we say this ???¡é?¡é?????¡­?¡°frequency range???¡é?¡é?????¡é???¡és bandwidth???¡é?¡é????????] is from 30 Ghz to 300 Ghz [ 270 Ghz of bandwidth ].

    So, as we go up in frequency, the bandwidth of the bands [ VHF, UHF, SHF, EHF etc ] increases. That is what I meant.

    This is probably as clear as mud, but basically in the higher frequency ranges, there is more bandwidth per range than at the lower ones. However as we go up to higher frequencies we hit absorption problems [ water vapor absorption at 22.5 Ghz and O2 oxygen absorption at 60 Ghz ]. However designers of WMAN???¡é?¡é?????¡é???¡és already looking at these frequencies as the others are becoming so congested.

    Whew???¡é?¡é???????|???¡é?¡é???????| ???¡é?¡é?????¡­?¡°Malcolm in the Middle???¡é?¡é????????is calling me???¡é?¡é???????|..

    To: Kevinsandlin: ???¡é?¡é?????¡­?¡°Who am I ? ???¡é?¡é?????¡­?¡° I started off as a man with six fingers on my right hand, which have now been worn down to five after this post


  • dave1234 Escribi?3:

    Hi Grant

    To: Kevinsandlin: ???¡é?¡é?????¡­?¡°Who am I ? ???¡é?¡é?????¡­?¡° I started off as a man with six fingers on my right hand, which have now been worn down to five after this post


    X-D X-D X-D

    LMAO. I can actually believe it!!!

  • Dave,

    Thanks again; I actually do understand what you wrote.

    Here is where I get messed up.

    When I think of modulating a signal, I imagine the radio wave changing in some way. Phase, amplitude or frequency, or a combination of them.

    Let's say that one wave (cycle) at 900 MHz is being modulated with phase modulation. Take the same phase modulation and apply it to 5 GHz (random higher frequency). Given that the frequency is higher, can't more data travel on the "wave" because there are more opportunities for phase (or amplitude or frequency) change?

    Thanks again!


    P.S. My first name is actually Gene, not Grant. I go by GT (Gene Thomas) because Gene is quite a geeky name and I'm trying to reduce my geekyness, as it runneth over. :)

  • Hi GT

    Firstly, in this Arena, "Thou shalt wear thy badge of geekiness with honour as you are among fellow geeks".

    From smudged spectacles to ink-stained shirt pockets, all are medals of honor !!

    When I took a book on link budget analysis on my honeymoon that did not go down well with the missus !!

    Bye the way anyone else out there "Princess Bride" fans ?

    You have asked an excellent question re modulation. Unfortunately it's also one of the hardest to explain without a lot of calculus.

    I'll have a think and get back in a while.


  • Firstly some terminology. In TRUE technical sense, the carrier signal is the UNMODULATED single frequency that is modulated by the input data. Once it has been modulated , it is technically speaking no longer the carrier signal. In fact in some AM systems, you get rid of the carrier almost completely !! [ I won???¡é?¡é?????¡é???¡ét get into that one at this stage ].

    Once the carrier signal has been modulated, the entire output should just be referrred to as the modulated output signal. The carrier is a component of the output, but there are other frequencies as well. However, historically we just refer to the entire output as the ???¡é?¡é?????¡­?¡°carrier???¡é?¡é????????. Wrong, but we have to live with it.

    Let me start off with an analogy.

    Imagine that there is a straight, horizontal railroad line from Smalltown California on the West coast to Bigtown New York on the East Coast.

    Smalltown represents 900 Mhz and Bigtown represents 5 Ghz

    There is a motorized cart which sits on the railroad track.

    You stand on the cart vertically upwards. You are the carrier signal.

    You have a cardboard sign, nice and square which you hold up at chest level. The sign is two feet wide.

    The cardboard sign is the modulated output spectrum. The bandwidth is 2 Mhz.

    On the sign it says ???¡é?¡é?????¡­?¡°Modulation is ???¡é?¡é????????orrible !! Bloody ???¡é?¡é?????¡­?¡°orrible !!???¡é?¡é????????

    This is the data that is being sent.

    You eat a big lunch, and are ready to go.

    Someone takes your photo at Smalltown. Your sign is there, the words are there and you are there.
    The cart starts to move. You stand perfectly still. Even though you are moving down the line and distance [ frequency ] is increasing, your sign and the data it contains has not changed.

    Finally you end up in Bigtown New York.

    Another photo has been taken and your sign is still the same width [ output bandwidth ] and the same words [ data ] are there.

    The carrier signal merely carries the data, it does not participate in determining the modulated output bandwidth.

    So what does determine that output? Lots of things:

    The input data rate

    The modulation Type [ BPSK, QPSK, 64 QAM etc ]

    Error Correction being used

    Output filtering

    Overhead [ for some long range microwave systems ]

    A long time ago, in one book, someone said ???¡é?¡é?????¡­?¡°because a 10 Ghz frequency signal is a higher frequency than a 10 Mhz frequency signal, there are more ???¡é?¡é?????¡­?¡°opportunities???¡é?¡é???????? for it to be modulated in phase, frequency etc.???¡é?¡é????????

    I wish I could find out who said that, because it has spread like wildfire all over the place and is completely wrong.

    So where did they get that idea from?

    I think it was from this:

    Imagine we have a 10 Mhz output signal [ output bandwidth ].

    Now imagine we have been given a 20 Mhz center frequency. The bandwidth limits would go from 20 ???¡é?¡é?????¡é?€?? 10 = 10 Mhz to 20 + 10 = 30 Mhz.

    No problem.

    Now imagine that we want a 100 Mhz output spectrum.

    The upper limit will be at 20 + 50 Mhz = 70 Mhz. No problem.

    The lower limit will be at 20-50 = ummmmm???¡é?¡é???????| -30 Mhz ?

    We cannot have that. The problem is that we have not go enough elbow room at the lower range. However if we go up to 1 Ghz as the carrier frequency, it???¡é?¡é?????¡é???¡és no problem.

    1000 Mhz ???¡é?¡é?????¡é?€?? 50 Mhz = 950 Mhz Lower limit

    10000 Mhz + 50 Mhz = 1050 Mhz Upper limit

    I think that this issue of ???¡é?¡é?????¡­?¡°elbow room???¡é?¡é???????? has somehow started all of this business over.

    In summary, the carrier signal merely carries the data as a modulated signal.

    Historically we have referred to the unmodulated carrier as the carrier and the modulated carrier as the carrier. Not good, but we can???¡é?¡é?????¡é???¡ét change it now.

    I???¡é?¡é?????¡é???¡éll put up a post some time on how modulation really works, but???¡é?¡é???????|.

    Must sleep???¡é?¡é???????|???¡é?¡é???????|need food???¡é?¡é???????|???¡é?¡é???????|???¡é?¡é???????|???¡é?¡é???????|brain hurts....


  • My eyes hurt...

    They built a BED under his DESK ?????


  • If you are a Seinfeld fan, you will know that George Costanza is the laziest person on the planet, but this takes it to a whole new level:

    If the YouTube clips don???¡é?¡é?????¡é???¡ét come up, just put ???¡é?¡é?????¡­?¡°The Nap???¡é?¡é???????? in the search box.

    Enjoy or delete as appropriate.


  • Dave, PM sent, some contact details.

Page 1 of 2