Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Amplitude modulation

  1. Jan 5, 2012 #1
    Hello, I am trying to understand how the most basic form of AM works.
    Can someone please confirm if my brief description below is essentially correct ?

    I know that upper and lower sidebands are created when a signal is modulated onto a carrier frequency. These sidebands result in a signal that consumes a larger bandwidth that the original single frequency carrier signal. There is now a signal that spans a range of frequencies rather than the single carrier frequency.
    The amplitude modulation itself is created from the sum and difference application of the input signal to the carrier signal.
    Therefore the resulting amplitude modulated waveform, when drawn, would consist of a sine wave of varying frequency and amplitude ?

    Thanks for any help especially on confirmation of the line immediately above this one.
     
    Last edited: Jan 5, 2012
  2. jcsd
  3. Jan 5, 2012 #2

    vk6kro

    User Avatar
    Science Advisor

    If you just looked at an amplitude modulated signal on an oscilloscope, it would look like a constant frequency sinewave which changes amplitude.

    If you look at it with a spectrum analyser, you would see a constant amplitude carrier with sidebands equally spaced on either side of it.

    This is not a contradiction. The sideband signals interact with the carrier to produce the varying amplitude sinewave you see on an oscilloscope.

    You can bring 3 different signals from 3 different signal generators and create an amplitude modulated signal or you can amplitude modulate a sinewave and produce the extra sidebands in the process.
    The final result is the same.
     
  4. Jan 6, 2012 #3

    NascentOxygen

    User Avatar

    Staff: Mentor

    It is called amplitude modulation because the amplitude is modulated. Where the frequency is modulated, then you have frequency modulation.
     
    Last edited: Jan 6, 2012
  5. Jan 6, 2012 #4
    With amplitude modulation the frequency is still modulated indirectly, right? In that sense, similar to phase modulation, it's just a slightly different form of frequency modulation.
     
  6. Jan 6, 2012 #5

    vk6kro

    User Avatar
    Science Advisor

    No.

    If a 1 MHz carrier is amplitude modulated by a 1 KHz sinewave then the output is a sinewave at 999000 Hz, (the lower sideband), a sinewave at 1000000 Hz (the carrier) and a sinewave at 1001000 Hz, (the upper sideband).

    There is no frequency modulation at all.
     
  7. Jan 6, 2012 #6

    Averagesupernova

    User Avatar
    Science Advisor
    Gold Member

    Sometimes I think AM is the most misunderstood thing in electronics.
     
  8. Jan 6, 2012 #7
    It's important to remember that the carrier frequency is usually much higher than the modulating signal frequency. If you want to draw an AM signal, first draw the envelope of the unmodulated carrier (which will be two horizontal straight lines at +Vc and -Vc). Then draw the modulating signal on top of those lines. In other word when the modulating signal is +1V the upper envelope is Vc+1 and the lower is -(Vc+1), when the modulating signal is -1V the upper envelope is Vc-1 and the lower one -(Vc-1). The resulting picture will be symmetrical about zero. This assumes a sensitivity of one volt per volt.
    It is not really useful to try to draw in the carrier. Firstly it is too high a frequency to easily draw and secondly, if you were looking at this on an oscilloscope triggered by the modulating signal you won't see the carrier sine wave anyway, because there is no requirement for it to be in sync with the modulating signal so it will just look "filled in" unless you freeze the picture on a digital scope.
    An amplitude spectrum is much more useful that a time graph.
    AM is a situation where a bit of school level trigonometry explains everything more than any description!
     
  9. Jan 6, 2012 #8
    So what would the bandwidth of your transmitted signal be? If your signal were only amplitude modulated with no frequency modulation then shouldn't the signal be monochromatic or at least have only the three frequencies you mention? If amplitude modulation truly doesn't modulate frequency then where are all those extra frequencies coming from? Why isn't it possible to have a monochromatic AM signal with zero or close to zero bandwidth?
     
  10. Jan 6, 2012 #9

    vk6kro

    User Avatar
    Science Advisor

    The bandwidth of this signal would be 2 KHz and it would be composed of only 3 components: the carrier and two sidebands as above.

    What extra frequencies?

    You could have a signal with close to zero bandwidth if you switched it on and off VERY slowly.
    For normal modulation with speech the bandwidth is twice the highest modulating frequency.

    I have a receiver that is quite capable of listening to just the carrier of an AM signal and I can tell you that it is just a very stable sinewave with no frequency shifts at all.
     
  11. Jan 7, 2012 #10
    Maybe it's my understanding of the concept of bandwidth itself that is faulty. I have been thinking of it as a measure of the number of different frequencies being transmitted simultaneously. Wikipedia defines at as "the difference between the upper and lower frequencies in a contiguous set of frequencies." I notice how it says in a contiguous set of frequencies. In AM are the frequencies contiguous or discrete? That is are there any frequencies between the sideband frequencies and the carrier frequency being transmitted? Are the sideband frequencies fixed and monochromatic or do they change frequency or just continuously occupy a whole range of frequencies?

    What if you switch it on and off very quickly but don't modulate it per se as in deliberately sending information over a carrier? I was thinking about an example where you have a very short pulse once per day.

    This is interesting. I wonder what a very short pulse would look like to that receiver. Presumably the short pulse length would not alter the carrier wave at all and only the sidebands would be very far apart due to the shortness of the pulse.
     
  12. Jan 7, 2012 #11

    vk6kro

    User Avatar
    Science Advisor

    If you apply a pure sinewave modulation of one frequency, then the output will be as before. Just two sideband signals and a steady carrier.
    There are no other signals generated.
    The bandwidth is the difference between the lower and upper frequencies.

    Speech is much more complex, but the result is similar. For each component of the speech, there are two corresponding sideband frequencies generated. (Each group of frequencies is called a sideband and there are two of them. One lower and one upper.)

    What if you switch it on and off very quickly but don't modulate it per se as in deliberately sending information over a carrier? I was thinking about an example where you have a very short pulse once per day.

    If you switch it off all day and then just give one short pulse, I don't think you can argue that the carrier is there all the time. It plainly isn't, but I have heard people try to argue that it was.

    Slow Morse code can be passed through a narrow filter but as the Morse code gets faster, the pulses just blur together because the rise and fall time of the pulses are so long due to the narrow bandwidth..
    Narrow filters give better noise performance, but the filter must be able to cope with the bandwidth of the signal.
     
  13. Jan 7, 2012 #12
    To back up what Vk has said here is my diagram of the frequency spectrum of a 100kHz carrier wave Amplitude modulated with the full range of audio frequencies in the range 0.1kHz to 10kHz.
    The 2 side 'bands' show up I hope
    The bandwidth is 20kHz
     

    Attached Files:

    • AM.jpg
      AM.jpg
      File size:
      72.2 KB
      Views:
      146
  14. Jan 7, 2012 #13
    Based on that diagram it appears that each sideband has width in the frequency domain presumably proportional to the frequency of the baseband signal. The sidebands are NOT monochromatic. That's a big part of what I was not understanding. All of the spectral 'space' between the central carrier wave and the two sidebands is filled up with RF energy. With frequency modulation you also get sidebands in the frequency domain, but those bands actually represent the carrier wave changing its frequency. With FM, if you could freeze time, at any given point you would just have a carrier wave at a given frequency. If you could freeze time with amplitude modulation would each of those sidebands of RF energy in the frequency domain actually represent a single wave that is continuously changing its frequency as with a carrier wave in FM or many waves transmitting at different frequencies simultaneously? Somehow a large number of different frequencies is being transmitted. So either you have 2 discrete sideband waves which change frequency or a great many sideband waves at different frequencies.
     
  15. Jan 7, 2012 #14

    vk6kro

    User Avatar
    Science Advisor

    A carrier modulated by speech or music will have two sidebands like that.
    Every frequency in the modulating waveform mixes with the carrier to produce a sum and a difference product in the output.
    That is why they are called "bands".

    This may be quite simple if you have just a violin playing or very complex if you have the whole orchestra playing at once.

    BUT there is no FM.
     
  16. Jan 7, 2012 #15
    Okay. So with AM you are essentially broadcasting at many different frequencies at the same time. The RF energy just kind of spreads out. With FM the RF energy is concentrated at one frequency, but that frequency is continuously changing.

    So getting back to a concrete example. If you have a 4 kHz carrier wave and a 100 Hz modulating signal. You would have a contiguous stream of carrier waves being transmitted in parallel from 3900 Hz to 4100 Hz. At 3968 Hz you would have a wave. At 4033 Hz you would have a wave. At every frequency you looked in that range you would detect an identical carrier wave. Almost like you had one really 'thick' carrier wave that took up all of that spectral room.
     
  17. Jan 7, 2012 #16

    vk6kro

    User Avatar
    Science Advisor

    No.

    You would have a signal at 3900 Hz, then a great big empty space until 4000 Hz then another great big empty space until you got to 4100 Hz.

    Three signals with nothing in between.
     
  18. Jan 7, 2012 #17
    Now I'm totally lost. It's kind of funny because it seems like the OP was satisfied with the answers. Do you know of any books that explain these concepts well? I'm looking at various books on signal theory. I suppose this will all get a lot easier to understand once I learn how to do Fourier transforms.

    Technician's frequency domain graph and your statement:
    both led me to believe that the side "bands" are not monochromatic. Now it sounds like you are saying that they are.

    When you say that you would have a signal at 3900 Hz and a signal at 4100 Hz would those signals themselves have bandwidth? Would the 3900 Hz carrier for instance only be present at exactly 3900 Hz?

    I just checked with wikipedia on amplitude modulation and it says
    . That statement would lead me to believe that in this example you would have one band of frequencies from 3900-4000 Hz and another from 4000-4100 Hz. I can't remember the last time that I had so many seemingly contradictory answers to the same question.

    And then this graph seems to indicate that the sidebands are centered halfway between the the carrier frequency and the sideband frequency with continuously decreasing power going towards or away from the carrier. If that graph is correct then the sidebands are actually centered at 1/2 the baseband frequency from the carrier.
     
    Last edited: Jan 7, 2012
  19. Jan 7, 2012 #18

    Averagesupernova

    User Avatar
    Science Advisor
    Gold Member

    Meti, you need to reread this thread a bit more carefully. I can understand your confusion based on you quite likely skimming a bit too fast and missing out on some important details.
    -
    On the sidebands and the amount of spectrum that they occupy: If we take a carrier of 1 Mhz and modulate it with a pure sine wave of 100 KHz the final spectrum will contain the carrier of 1 Mhz, an upper sideband of 1.1 Mhz, and a lower sideband of 9.9 Mhz. All other spectrum in between the sidebands and carrier will be empty. This case is with a pure sine wave modulating the carrier and this is an important distinction. If we take the same carrier and modulate it with speech which is a comlex set of sine waves we will have the same result except the sidebands are more filled. The upper sideband will look EXACTLY like the baseband audio spectrum except that it is obviously shifted above the carrier. The lower sideband will look exactly like the upper sideband except mirrored. The important distinction here is the type of signal modulating the carrier. A complex modulating signal will get complex sidebands.
    -
    Amplitude modulating a carrier adds power to the final signal. So multiple signals exist at the same time.
    -
    In FM however, it is a bit different. Full modulation or zero modulation the final power output will be the same which is unlike AM. And to confuse you even more, FM does NOT sweep the carrier quite the way we think it does. If we take a carrier of 1 Mhz and frequency modulate it with 1000 Hz at 100 Khz deviation we will get sidebands all the way from the carrier to 100 Khz out either side of the carrier at 1000 Hz intervals. Even though we think of the carrier sweeping, the final result is individual sidebands spaced 1000 hertz apart with NOTHING in between.
     
  20. Jan 7, 2012 #19

    vk6kro

    User Avatar
    Science Advisor

    . That statement would lead me to believe that in this example you would have one band of frequencies from 3900-4000 Hz and another from 4000-4100 Hz. I can't remember the last time that I had so many seemingly contradictory answers to the same question.

    And then this graph seems to indicate that the sidebands are centered halfway between the the carrier frequency and the sideband frequency with continuously decreasing power going towards or away from the carrier. If that graph is correct then the sidebands are actually centered at 1/2 the baseband frequency from the carrier.


    A pure sinewave has only one frequency at a time and if you mix two of them together, you get the original signals as well as the sum and difference of the two.
    So, you get four discrete signals out, but one of them is the original audio signal which is not radiated.

    If you have a carrier and modulate it with 10 discrete sinewaves, you will get 21 signals out. The carrier and 10 signals with a lower frequency than the carrier and 10 with a higher frequency than the carrier.

    The "bands" shown in the diagram are only potential areas of spectrum which may get filled if the appropriate sound is produced to modulate the carrier with. The exact shape of these depends on the frequency response of the modulator.
     
  21. Jan 7, 2012 #20

    NascentOxygen

    User Avatar

    Staff: Mentor

    When a 4000 Hz carrier is amplitude modulated by a 100 Hz sinusoid, the output will be three sinewaves, one at 3900Hz, one at 4000Hz, and a third at 4100Hz.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook