Need advice

[epilot]

Hello there,


Today when I was thinking about AM modulation I thought why AM modulation requires bandwidth?
The carrier has certain freq and the modulated signal (audio) mixes with that carrier and changes the carrier Amplitude. So here we have changes in the carrier domain not carrier freq (like FM modulation), so?
I think a very low bandwidth is ok, but why my guess is not true!

 

[dknguyen]

Different amplitudes = different rates of change that are not the same as a single sinusoid = different frequencies to put it very very simply. A single frequency component means that there is a constant amplitude, constant frequency UNCHANGING sinusoid. Any change that causes the sinusoid to deviate from itself will cause more than one frequency to be present- even if it's just momentary transition to a new maximum amplitude, and then it goes back to a different, single frequency once that transition is done. Well...that's just what comes to mind since I forget why. That's an interesting question. I'll have to pull out my communiations textbook and go over it again and get back to you.

The same thing happens in FM if you do not think about it completely (ie zero bandwidth required).

 

[Sceadwian]

RF transmission's most ultimate method of imagining it's output as a function of frequency vs time is the rate of change in voltage. AM modulation causes variations in the maximum and minimum voltages created hence the ultimate RF output changes because the impedance of an antenna changes with the rate of change of the voltage across it, even if the frequency of that change is the same. Remember that single point, rate of change. Also remember that the carrier itself contains no information and is not actually required to transmit data. The 'data' is contained in the side bands which are a result of the rate of change in the instantanious amplitude of the sine wave.

 

[dch222]

The modulating signal generates sum and difference frequencies either side of the carrier.

eg a 6Khz sine wave modulated onto a 1MHz carrier will produce frequencies of 1 + 0.006 and 1 - 0.006 that is 1.006MHz and 0.994MHz.

A more complicated modulation (like music) will have multiple modulation frequencies and will produce multiple sum and difference frequencies either side of the carrier.
These are called sidebands and need to be transmitted as well as the carrier, although it is possible to get away with only transmitting one since both carry the same information.

 

[Sceadwian]

Suppressed carrier SSB is used by amateur radio users sometimes on lower frequencies as it allows more efficient use of very limited bandwidth. Both the carrier and one side band can be filtered out at the transmitter, decreasing the bandwidth and power levels required for a signal. All that's required for regeneration is a local oscillator on the receiver tuned to the same frequency as the suppressed carrier.

 

[RadioRon]

Hello there,


Today when I was thinking about AM modulation I thought why AM modulation requires bandwidth?
The carrier has certain freq and the modulated signal (audio) mixes with that carrier and changes the carrier Amplitude. So here we have changes in the carrier domain not carrier freq (like FM modulation), so?
I think a very low bandwidth is ok, but why my guess is not true!

You know, this concept of the frequency domain characteristics of a modulated carrier is one of the most difficult to understand for a beginner. It took me quite a while to get a feel for it. In university the concept is taught using mathematics. To go this route, you need to grab a textbook. For example, I grab my old technical school text called "Electronic Communication Systems", by George Kennedy (McGraw Hill, 1970) and find a good explanation on page 72. This explanation makes the point that any modulation, including AM, distorts the sinusoid that is the carrier. We learn in the mathematics of electronics that when you distort a sinusoid, you create other spectral components. Different kinds of distortion create different quantities and frequencies of spectral energy. This is a fundamental point. Another fundamental point is that it is, at least for me when I was a beginner, impossible to understand the frequency spectrum of a modulated carrier by looking at the time domain waveform. You really have to look at the math to prove what is exactly happening.

 

[Sceadwian]

Personally rate of change describes it all to me without any math. The rate of change of a perfect sine wave is always exactly the same so the frequency is flat. When the rate of change (amplitude) increases or decreases the instantanious frequency changes as a result. The frequency can also change at the same given amplitude by altering the rate of change. Mathmaticly AM and FM modulation are really the same, looking at the visual representation of AM and FM modulation makes people think it's fundamentally different, and the various methods of creating and receiving AM and FM signals have different charactoristics and ease of implementation with AM or FM, but the underlaying principle of rate of change is the only cause of a frequency.

 

[Sceadwian]

If I'm not mistaken, the reason it doesn't look like a triangle wave is because the actual energy goes up exponentially with increasing voltage, which gives it the curve even though the rate of change is linear.

 

[epilot]

The modulating signal generates sum and difference frequencies either side of the carrier.

eg a 6Khz sine wave modulated onto a 1MHz carrier will produce frequencies of 1 + 0.006 and 1 - 0.006 that is 1.006MHz and 0.994MHz.

A more complicated modulation (like music) will have multiple modulation frequencies and will produce multiple sum and difference frequencies either side of the carrier.
These are called sidebands and need to be transmitted as well as the carrier, although it is possible to get away with only transmitting one since both carry the same information.

Yeah that's right,

But I am not able to guess why the sidebands have sum and difference freq while the amplitude of the carrier is changed by modulating signal (audio)?!

the attachment is a pic of AM modulation by "Ron-H". can you show me the sidebands?
if you call the upper side of the wave "upper sideband) and the lower side of the wave "lower sideband" like you said ""...produce multiple sum and difference frequencies ;;either side of the carrier;;"",so please let me know why I am seeing the upper the lower sidebands the same in freq? it seems the upper and the lower sidbands have the same freq?!

Please, I don't like the math here

 

[Hero999]

AS was mentioned perviously you need a frequency domain plot to understand it - looking at the time domain only confuses things.

 

[RadioRon]

No, you can't look at it like that. You cannot see the sidebands in a time domain (oscilloscope) view of the signal.


Like I said before, this is one very tough part of electrical engineering. Unfortunately math is an essential part of the language of electrical engineering. If you can't handle the math, then you must fall back on faith that an AM signal as seen on your scope has sidebands as others explain to you. You may not ever truly understand more than that without the math.

 

[JimB]

OK lets try a few pictures.

The attachments are photos taken from an oscilloscope and a spectrum analyser.
They are looking at the output from a signal generator which was running at 3 Mhz.
The first two pictures were taken with the sig gen amplitude modulated by a 5kHz tone.
The second two pictures were take with the sig gen amplitude modulated by audio from a radio (speech and music).

On the spectrum analyser, the large centre peak is the carrier, either side of which can be seen the sidebands, upper sideband to the right (higher frequency), lower sideband to the left (lower frequency).

The scaling on the spectrum analyser is:
vertical 10dB per division
horizontal 5kHz per division (calibration is a bit off!).

With the tone on the spectrum analyser, the small blips at each side of the main components are due to distortion in the audio signal and the modulator in the signal generator.

JimB

 

[Sceadwian]

What program are you using to simulate that waveform? Most of them allow you to do FFT and view the frequency output, you can plainly see the sidebands there. A program like LTspice (free) would allow you to do it if yours doesn't.

 

[epilot]

OK lets try a few pictures.

The attachments are photos taken from an oscilloscope and a spectrum analyser.
They are looking at the output from a signal generator which was running at 3 Mhz.
The first two pictures were taken with the sig gen amplitude modulated by a 5kHz tone.
The second two pictures were take with the sig gen amplitude modulated by audio from a radio (speech and music).

On the spectrum analyser, the large centre peak is the carrier, either side of which can be seen the sidebands, upper sideband to the right (higher frequency), lower sideband to the left (lower frequency).

The scaling on the spectrum analyser is:
vertical 10dB per division
horizontal 5kHz per division (calibration is a bit off!).

With the tone on the spectrum analyser, the small blips at each side of the main components are due to distortion in the audio signal and the modulator in the signal generator.

JimB

Thanks to all,

So can someone explain me the below pic?

 

[Roff]

Thanks to all,

So can someone explain me the below pic?

That waveform can be broken down to AM (carrier plus sidebands), linearly added to the baseband modulating signal. Assuming the modulating signal and the (unmodulated) carrier are both sinusoids with no harmonics, on a spectrum analyzer you would see blips at Fmod, Fcarrier-Fmod, Fcarrier, and Fcarrier+Fmod.

 

[epilot]

That waveform can be broken down to AM (carrier plus sidebands), linearly added to the baseband modulating signal. Assuming the modulating signal and the (unmodulated) carrier are both sinusoids with no harmonics, on a spectrum analyzer you would see blips at Fmod, Fcarrier-Fmod, Fcarrier, and Fcarrier+Fmod.

The carrier Is a 40KHz square wave, but the modulating signal is a 1KHz
(sine wave).

I am not able to seed the lower sideband so how are you saying that it is an AM (carrier plus sidebandS)? why not saying SSB plus carrier?!

 

[JimB]

The carrier Is a 40KHz square wave, but the modulating signal is a 1KHz
(sine wave).

That picture looks like it was taken from a digital scope, and I think the sampling rate is too low, you seem to have aliases in there.
I am not sure I completely agree with RonH about what is in that trace, my first impression is YUK!.

I am not able to seed (see ? JimB ) the lower sideband so how are you saying that it is an AM (carrier plus sidebandS)? why not saying SSB plus carrier?!

You can't see the upper sideband either. You can't see the sidebands on a scope, you need a spectrum analyser.
The upper and lower sidebands are not the upper and lower part of the scope trace.

JimB

 

[epilot]

That picture looks like it was taken from a digital scope, and I think the sampling rate is too low, you seem to have aliases in there.
I am not sure I completely agree with RonH about what is in that trace, my first impression is YUK!.




You can't see the upper sideband either. You can't see the sidebands on a scope, you need a spectrum analyser.
The upper and lower sidebands are not the upper and lower part of the scope trace.

JimB

Yes the pic is taken from my digital scope it has a spectrum analyser too and I have seen to FFT of the wave too

the below are the time domin and FFT of the wave at 500uS

Are you sure that the upper level of the wave is not the upper sideband as I have assumed in the papers:?**broken link removed**

yes I meant 'see', sorry for mistype. what YUK means?!

 

[Nigel Goodwin]

Are you sure that the upper level of the wave is not the upper sideband as I have assumed in the papers:?

Yes, it's NOT the upper sideband, it's just a VERY highly distorted attempt at AM modulation - in fact, it looks like it's been rectified?.

 

[Roff]

I am not sure I completely agree with RonH about what is in that trace, my first impression is YUK!.
JimB

See the schematic and waveforms below.
I can plot the FFT if anyone wants to see it.