Understanding AM/FM Radio Stations

[enigma]

Can we please take the hostility down a notch, guys?

 

[Averagesupernova]

Averagesupernova:

As always you seem to get things mixed up. The nature of the device that produces the RF is of no consequences. You only need to be concerned with the process of modulation, the transfer function. As you have pointed out the RF voltage varies in a linear manner with applied voltage. You also agree that the amplitude of the side band also varies in a linear manner with applied audio. If you really don’t understand modulation at this point you must be the only one on this forum that does not.

Look at a 1 MHz signal modulated with a tone on a spectrum analyzer. The carrier does not change in amplitude and neither does either sideband as long as the strength of the tone is constant. What you see on the scope changing in amplitude is the complex waveform of the carrier in combination of the 2 sidebands. All of them combined produce the waveform. Remember when I mentioned taking 3 independently generated signals and summing them? The waveform you would see will look like the signal you are referring to.

As for the source producing the RF being irrelevant? Well, it IS relevant. If you run the PA transistor/tube linear such as class A/B and leave plenty of headroom then all that superimposing audio onto the power supply will do is SUM the RF and the audio. It will do this because you can lower the supply voltage and the sine wave won't change because it still has headroom left before it clips. (To a point of course) I will admit it is easy to forget that the PA in a class C output is running non-linear because the collector circuit has built in filtering and the output appears as a sine wave. It just seems like a good source for a sine wave but it is actually part of the modulator.

Of course you know I don’t agree that frequency mixing and amplitude modulation are the same thing. You also know that it is not the same thing.

Well, technically they are. I can recall seeing it in an ARRL publication. I'd have to dig it out but it was there. I can do it if you want, hopefully I have the book. I wouldn't have said it if I weren't sure.

At this point I can see that you are starting to see this in a whole new light. I also know from what you have posted that you will never state that you are ever wrong.

Why should I state I am wrong? You see I have this tendency to post about things I am certain of. I am not saying I have never been wrong but I don't open my mouth unless I am darn certain. You should try it. I will admit that there have been certain times in this thread that I have posted and assumed that the reader knows what I am talking about. The example I can think of is in a post when I said mixing and modulation are the same thing. What I should have said is that frequency mixing and AM modulation are the same thing. Technically you could read all modulation including FM was the same thing as frequency mixing. I am guilty of it in many threads. It is part of my personality.

I have no problem teaching electronics but I don’t care to do it in a confrontational manner. If you whish to know how mixers work or any thing else please change your approach. The approach I recommend is that you state you understanding listen to the other understanding, if different. then pose questions and give replies to the point. If you are confused on any point, resolve the confusion first without confrontation.

I don't need to know how mixers work. I already do. When did you assume you were the one doing the teaching? You seem to contradict things out of both books I have quoted now. How many books do you want to go for because as long as this keeps up I will quote out of books until I run out of them. And I will make a point of quoting things that prove your view wrong. Funny thing is though, the texts mention that in order to produce new frequencies the device has to be driven into it's non-linear region. You seem to ignore it. If new frequencies are generated with an AM modulator without non-linear operation then why can't the same be done with mixers? The book says it is not possible and from what I can recall working on video equipment modulators and mixers it is not.

 

[Averagesupernova]

Enigma I am very sorry if this is getting a little too hostile. I guess at this point I would like some more input from other users. I know that there are several other posters on here who should have some input. I know my tone can be a bit confrontational at times. But I don't see how quoting textbooks and having my opponent completely ignore facts from those books adds anything to this thread. So as an apology to everyone, I want to ask all the posters on this thread what they would like me to do? If posters have a specific question about something I have posted by all means ask.

 

[4newton]

Enigma you are right.
I am sorry that my frustration got out of hand.

having my opponent completely ignore facts”

I apologize Averagesupernova I should know better, but I did not realize, until now, that I was considered an opponent.

 

[enigma]

Thank you both.

 

[4newton]

Averagesupernova;

But I don't see how quoting textbooks and having my opponent completely ignore facts from those books adds anything to this thread.

I am sorry if you have taken offense at my not responding to your quotes but I did not feel that it was necessary. I do not disagree with your quotes. I thought I stated that once before. They are just not relevant to the disagreement about modulation.

I have tried to hold the discussion to the basics of modulation. I did not feel that it was of any use going in any other direction without that basic agreement between us.

I pointed out and you agreed that the voltage out of the RF amp was linear with applied voltage. I also pointing out to you and you agreed that the voltage in the sideband from the modulation was also linear with applied audio, I thought that should have resolved the issue. I still don’t know why it has not. It is axiomatic that a result that is linear is a linear function.

Maybe I should introduce you to the idea of a transfer function. The RF source maybe thought of as a black box. It has only three terminals. Common, supply voltage, and output. The only thing you can change is the supply voltage. The only output you have is the RF. I have shown. And you have agreed, that with any change of supply voltage or supply voltage change at an audio rate the output is always linear. If you disagree all you need to do is show that the output will change in a nonlinear manner with a change of supply voltage.

The black box could be a constant RF source that has a variable resistor the output of the box is linear with the position of the resistor. As you change the resistor, by definition, you modulate the output. This box will produce amplitude modulation with the same waveform of any other AM modulator. There is no nonlinear component. This is why the nature of the RF source is of no importance. Only the transfer function is relevant.

 

[Averagesupernova]

You didn’t realize you were an opponent? Well, I didn’t consider you so at first, but somewhere along the line….. And don’t we have opposing views? Anyway, on to the disscussion.

My goals have not changed. They are:

1) To convince you that mixing and amplitude modulation are considered the same thing.

2) To convince you that what you call the voltage out of the RF amp varying linearly with applied audio is nothing more than the carrier and 2 sidebands forming a complex waveform.

3) To convince you that the source of the RF is in fact relevant and that the output device is part of the modulator.

4) To convince you that a device running linear mode on the output would NOT be capable of modulation by superimposing audio onto the power supply.

I have once again turned to the good old ARRL to back my side of this whole deal. This time the book I am quoting is entitled:
THE RADIO AMATEUR’S LICENSE MANUAL. It is the 80th edition around 1984.

On page 2-19 there is a good columns worth of text that talks about amplitude modulation. Here we go.



Amplitude Modulation

The modulation system that gave rise to the carrier idea is called amplitude modulation (a-m). It’s a system for doing just what we have described above.



What they mean when they refer to what they described above is a description of sidebands and carrier. They descibe the mathematical relationship between the carrier, sidebands and frequency of the modulating audio.


But when we come to the actual mechanics by which the signal is modulated, it’s easy to thinik of it from a different viewpoint. That viewpoint is this: That the amplitude of the carrier (that is, the value of the carrier’s current or voltage) is made to follow faithfully the instantaneous changes in the audio-frequency voice voltage. The general idea is shown in Fig. 32. In the amateur bands, the carrier frequency is thousands of time greater than the audio frequency, so the carrier will go though a great many cycles during one cycle of even the highest audio frequency we want to transmit. This is shown by the shading in the figure; we couldn’t begin to draw the actual radio frequency (rf) cycles because there would be far too many to be printed.




Figure 32 is the typical AM 100% modulated signal as viewed on a scope.


You’re probably thinking that something is actually being done to the carrier. But this is only because it isn’t possible to draw a picture of more than one aspect of modulation at a time. The picture you see in Fig. 32 is really a composite one showing the result of the action of three separate frequencies in a circuit that will pass all of them. The three are the carrier frequency, the upper side frequency and the lower side frequency. They all add together in such a way as to give the appearance of a single frequency (the carrier) whose amplitude is changing just the same way that the signal doing the modulating is changing.


Isn't this what I've been saying?


But appearances can sometimes fool us. It would be more accurate to say that the actual modulation process is one of mixing, which we mentioned in our discussion of receiver circuits.



I do believe I have been saying this too.


Nevertheless, it’s easier to grasp some things about amplitude modulation with the help of a picture such as Fig. 32, and we’ll take advantage of it. Never forget, though, that the carrier does not actually vary in amplitude, and that the modulation is all in the two sidebands.

Averagesupernova;
I am sorry if you have taken offense at my not responding to your quotes but I did not feel that it was necessary. I do not disagree with your quotes. I thought I stated that once before. They are just not relevant to the disagreement about modulation.

The quote out of the above book should prove otherwise. The one about modulation actually being a process of mixing.

I have tried to hold the discussion to the basics of modulation. I did not feel that it was of any use going in any other direction without that basic agreement between us.

Once again, mixing, modulation, can't have one without the other.

I pointed out and you agreed that the voltage out of the RF amp was linear with applied voltage. I also pointing out to you and you agreed that the voltage in the sideband from the modulation was also linear with applied audio, I thought that should have resolved the issue. I still don’t know why it has not. It is axiomatic that a result that is linear is a linear function.

Ummm, not quite. I believe I said that the envelope follows the audio in a linear fashion. You talk about 'RF' coming out of the amp. You seem to refer to it as a quantity of 'stuff'. The 'envelope' is an illusion of sorts, a complex waveform made up of 3 signals as the text says.

I've done a little math. I will show you how modulation is not linear. Take a 20 volt peak to peak sine wave that is NOT modulated. It is driving a 50 ohm load. The power dissipated in the load is 1 watt. I think we can agree on that.

Now let's modulate it at a given frequency with a NEAR square wave. I hope you are ok with this method. We will modulate it 100%. So what we have is the 'RF output' going on and off. The voltage NOW swings up to twice the original peak to peak voltage but only half the time. The other half of the time the output is ZERO. Average power dissipated is 2 watts. Now let's modulate at 50%. The envelope peaks up to 15 volts now for a 30 volt peak to peak signal. Incidently, this condition is defined in the book I quoted as 50% modulation. So now we have power dissipated in the load at the peak of the envelope of 2.25 watts and the trough of the envelope of .25 watts for an average of 1.25 watts. Hmmmmmm. Is this linear? I don't think so.

Incidentally, notice that twice the power is dissipated at 100% modulation as with no modulation. Ever notice at 100% modulation using a tone the sidebands are each exactly 3 db down from the carrier when viewed on a spectrum analyzer? Each is half the power of the carrier which adds up to what we figured in the above math. I used square wave modulation for simplicity.

Maybe I should introduce you to the idea of a transfer function. The RF source maybe thought of as a black box. It has only three terminals. Common, supply voltage, and output. The only thing you can change is the supply voltage. The only output you have is the RF. I have shown. And you have agreed, that with any change of supply voltage or supply voltage change at an audio rate the output is always linear. If you disagree all you need to do is show that the output will change in a nonlinear manner with a change of supply voltage.

I just did that.

The black box could be a constant RF source that has a variable resistor the output of the box is linear with the position of the resistor. As you change the resistor, by definition, you modulate the output. This box will produce amplitude modulation with the same waveform of any other AM modulator. There is no nonlinear component. This is why the nature of the RF source is of no importance. Only the transfer function is relevant.

You can do that, but the math I did still proves it is not linear.


Now I want to ask you a question: I am going to take a class A or class A/B amplifier configured common emmiter and take the output off of the collector. I have set it up so that the voltage swing of + and - 5 volts and the average DC voltage on the collector at 10 VDC. I am using a 20 volt power supply. So I have some headroom left so to speak. Now I superimpose a tone on the powersupply but it never gets down to 15 volts. Just low 'modulation' so to speak. What would the output look like?

 

[4newton]

Averagesupernova

1) To convince you that mixing and amplitude modulation are considered the same thing.

Do you think summing two audio frequencies together with resistors is modulation or mixing?

2) To convince you that what you call the voltage out of the RF amp
varying linearly with applied audio is nothing more than the carrier and
2 sidebands forming a complex waveform.

Are you saying that the wave form you see on the scope is or is not the change of voltage with time and what you see is not taking place?

3) To convince you that the source of the RF is in fact relevant and
that the output device is part of the modulator.

4) To convince you that a device running linear mode on the output
would NOT be capable of modulation by superimposing audio onto the power
supply.

Did you follow the concept of transfer function? What is the transfer function of a class C amp with applied voltage? What is the transfer function of a class A amp with voltage?

That the amplitude of the carrier (that is, the value of the carrier’s current or voltage) is made to follow faithfully the instantaneous changes in the audio-frequency voice volttage.

What do you think that says? This is a description of a linear function.

They all add together in such a way as to give the
appearance of a single frequency (the carrier) whose amplitude is changing
just the same way that the signal doing the modulating is changing.

You note the word ADD.

Now let's modulate it at a given frequency with a NEAR square wave. I hope you are ok with this method. We will modulate it 100%. So what we have is the 'RF output' going on and off. The voltage NOW swings up to twice the original peak to peak voltage but only half the time. The other half of the time the output is ZERO. Average power dissipated is 2 watts. Now let's modulate at 50%. The envelope peaks up to 15 volts now for a 30 volt peak to peak signal. Incidently, this condition is
defined in the book I quoted as 50% modulation. So now we have power
dissipated in the load at the peak of the envelope of 2.25 watts and the
trough of the envelope of .25 watts for an average of 1.25 watts.
Hmmmmmm. Is this linear? I don't think so.

Are you aware that you went from voltage to power. How does power vary with a change of voltage in a resistor? Is changing voltage across a resistor a linear function? Look again at you example.

 

[Averagesupernova]

Averagesupernova
Do you think summing two audio frequencies together with resistors is modulation or mixing?

No. I have NEVER said that. In fact, I was the first one in this thread to dispute it. When I speak of mixing I mean frequency mixing. Maybe for someone who remembers ‘old AM’ I should call a mixer a converter. When I mean summing I will say it.

Are you saying that the wave form you see on the scope is or is not the change of voltage with time and what you see is not taking place?

Well the last time I looked up the definition of an oscilloscope I believe it was defined as a device which is primarily used for displaying a graph of voltage over time. Could the questions become any simpler?

YES, the waveform viewed is in fact a voltage changing over time and what is seen is the carrier and 2 sidebands summed together. Those 3 are in fact summed linearly. The non-linear part came before that and the 2 sidebands are the result of it. It is what I call a complex waveform.

Did you follow the concept of transfer function? What is the transfer function of a class C amp with applied voltage? What is the transfer function of a class A amp with voltage?

Yes I follow the concept of transfer function. However, we cannot agree on the behavior of a class C versus a class A amplifier when superimposing audio onto the power supply. At least I cannot assume we agree because you haven’t answered my question concerning the class A or class A/B amplifier from my last post. So what is the point of talking about transfer function? In order for us to agree on it we have to first agree on the behavior of the considered amplifiers.

What do you think that says? This is a description of a linear function.

How convenient for you to leave out the surrounding text. You KNOW that the paragraph is talking about how it is viewed on a scope and nothing else.

You note the word ADD.

Of course I do. What is your point? See my comment earlier in this post concerning the addition of 2 sidebands and the carrier.

Are you aware that you went from voltage to power. How does power vary with a change of voltage in a resistor? Is changing voltage across a resistor a linear function? Look again at you example.

Do you really think that I accidentally went from voltage to power?

I did some more math. The conditions are the same as my last post with the near square wave. Here is what I came up with:

Take a sine wave at 20 volts peak to peak driving a 50 ohm load. The power dissipated in it is 1 watt. Double the voltage and the power dissipated in the resistor is quadrupled. This is nothing new to me either, but I am glad you brought it up.

Excuse the crude chart, but it should get the point across.

% mod_____pk volts_____min volts_____pk pwr_____min pwr_____Avg pwr
0_________10__________10__________1__________1__________1
10________11__________9___________1.209______.81_________1.0095
20________12__________8___________1.44_______.64_________1.04
30________13__________7___________1.69_______.49_________1.09
40________14__________6___________1.959______.36_________1.1595
50________15__________5___________2.25_______.25_________1.25
60________16__________4___________2.559______.16_________1.3595
70________17__________3___________2.889______.09_________1.4895
80________18__________2___________3.239______.04_________1.6395
90________19__________1___________3.609______.01_________1.8095
100_______20__________0___________4__________0__________2

I think the chart is fairly self-explanatory. But I will go over it anyway. The first column should be obvious. The second column is the PEAK voltage on the envelope. The third column is the PEAK voltage in the ‘trough’ of the envelope. The fourth column is the AVERAGE power in watts during the crest of the envelope. The fifth column is the AVERAGE power in watts during the ‘trough’ of the envelope. The 6th column is the average total power of the whole envelope.

The load used is a 50 ohm load. Notice how NONE of the power columns wattage is linear with the % of modulation the same way it would be if you just were doubling the voltage on a resistor. Now, take and subtract one watt out on all entries in the average column (6th). This represents removing the carrier. You have 2 sidebands left. The power STILL does not increase the way a plain old sine wave would when doubling the voltage across a resistor. Now divide each entry in half in the 6th column. This represents removing on sideband. NOW the power is linear. The same way it would be by doubling the voltage of a plain old sine wave into a resistor. Which is exactly what it is. One single frequency. If you were to filter each sideband out you would find that each one increases in power linearly. We’ve agreed on that. In reality, there would be many sidebands because the modulating signal is not sinusoidal in my example. But the power in each group of sidebands would add up to the same power as if it were one single upper and lower sideband.

One last thing. I want an answer to my question about modulating the supply voltage on a class A or class A/B amp. What will the output look like?

 

[4newton]

Averagesupernova:

Yes I follow the concept of transfer function. However, we cannot agree on the behavior of a class C versus a class A amplifier when superimposing audio onto the power supply. At least I cannot assume we agree because you haven’t answered my question concerning the class A or class A/B amplifier from my last post. So what is the point of talking about transfer function? In order for us to agree on it we have to first agree on the behavior of the considered amplifiers.

The transfer function is the behavior of the considered amplifiers. A class A/B amp has a constant output and does not change with any change of voltage. It therefore is irrelevant to the question at hand. You can not modulate it you can not use it for summing. It is a constant output device. The transfer function of a class C amp changes linearly with applied voltage. Just as you show in you chart. The voltage changes linearly with the modulation. Your chart will show the same results if you have no RF at all and you put a step function that has the same differential voltage as you show in col. 2 and 3 into you resistor.

If we just stay with the oscilloscope I think we can make some progress. You keep drifting off to waveforms and devices that are irrelevant to the question. As you agree the oscilloscope is a true picture of what is taking place. Do you agree that the whole topic can be resolved in this manner.

 

[Averagesupernova]

So now you are saying that a class A amplifier has a constant output that doesn’t change with supply voltage. That is not what you said several posts ago. You said that the source of the RF is irrelevant. So you now see my point? Class A is in fact different than class C? If so, then the output transistor/tube has to be considered part of the modulator.

Yes, you are right. The instantaneous voltage on the envelope changes linearly with applied audio. But the AVERAGE voltage out does not. If it did, then the power would also, and it doesn’t. But that is sort of irrelevant because as I have said countless times now the envelope having a shape that follows audio is just the summing action of the 2 sidebands and the carrier. Take 3 signals. A 1 MHz carrier, and an upper and lower sideband each at half power of the carrier. Make the sidebands each oh, say 5 KHz either side from the carrier. Now combine these 3 signals with a summing circuit. What you will see is the same thing that is seen if you modulated a 1 MHz carrier with a 5 KHz audio tone at 100%. But, there is no 5 KHz tone. Just the 3 signals combined. So you cannot say that responsibility for generating the envelope in question is solely placed on the audio frequency. Because in this case there is no audio frequency.

Incidentally, why should we just stick to the oscilloscope? Right is right, and ALL instruments involved need to lead us to the same conclusion. You just can’t pick your test equipment based on whether or not you like the results that they give you. And discussing modulation while prohibiting discussing the sidebands and other involved frequencies is like discussing the methods for making chocolate chip cookies while prohibiting the mentioning of chocolate chips. I have never said that the scope is the complete picture of what is taking place. And I do not believe that the whole topic can be resolved just by looking at what is on the scope.

 

[4newton]

Averagesupernova:

Yes, you are right. The instantaneous voltage on the envelope changes linearly with applied audio.

Thank you. I don’t know what the other points are that you are talking about. This was the only difference I know of that we had

You said that the source of the RF is irrelevant.

That is right. It is only important that the transfer function is that same as it is for high level modulation. You can’t take a rock and say that modulation is not linear because the rock can not be modulated. You must have a device that can be amplitude modulated.

why should we just stick to the oscilloscope

The main reason is to decide on an instrument to prove the question. The oscilloscope will do the complete analysis, only of course if you know how to read an oscilloscope. But at this point the issue is settled. If you wish I will interpret the oscilloscope trace for you and show you how it includes the sidebands.

Have a Good day.

 

[Averagesupernova]

Nice cop out...

 

[4newton]

What is you problem? What point do we not agree on?

Ummmm, not quite. A mixer or modulator requires a NON-LINEAR device. Using resistors will form an adder which is NOT a mixer.

Nope, a linear circuit is incapable of true modulation. You will not get sum and difference frequencies out of a circuit that you describe. It may be capable of adding 2 signals together to send down the same wire or something but this is not a mixer. I've worked on a lot of RF equipment and I have never seen a mixer/modulator that didn't use some sort of non-linear device.

But the FUNCTION of modulation is not linear. A true linear amplifier will sum two signals and not create new frequencies. A non-linear amplifier will sum them AND create new frequencies. Creating new frequencies is what modulation is all about.

Your last statement.

Yes, you are right. The instantaneous voltage on the envelope changes linearly with applied audio.

I pointed out to you this function.

The black box could be a constant RF source that has a variable resistor the output of the box is linear with the position of the resistor. As you change the resistor, by definition, you modulate the output. This box will produce amplitude modulation with the same waveform of any other AM modulator. There is no nonlinear component. This is why the nature of the RF source is of no importance. Only the transfer function is relevant.

This shows that you do not need a nonlinear device and that a linear change produces AM modulation. You did not comment on this statement.

All your other posting were off topic to the disagreement. You can not resolve a disagreement by expansion to non related topics. All you need do is reduce to basics as I have done. Do you wish me to comment where you are wrong on other point that you brought up? If we start into that we will be here forever. Like I said before. You have a lot of information but you just don’t put it together right.

 

[Ouabache]

Edwin Armstrong

Very perceptive discussion! Fascinating to learn about AM & FM transmissions.
BTW, some light reading describing the subtleties of FM, may be found in an original paper written by Edwin Armstrong,
inventor of frequency modulation, published May 1936, Proceedings of the Institute of Radio Engineers

 

[GENIERE]

Everyone here is right to some extent. No one it seems is old enough to remember vacuum tubes...

I’m not sure of the date but I think it was in 1949 that my friend’s dad drove us to NYC to look at the transmitting equipment, WABC I think it was, and his dad may have been a broadcast engineer.

I recall little except the huge size of a vacuum tube, in my memory it was 50 feet high, but it was probably more like 8-10 feet. Years later I realized it was a very high wattage water-cooled transmitting tube that I had seen.


...

 

[Antiphon]

I'm a professional EE and I'm not taking sides, but I must simply state the facts.

Modulation is mathematically linear in that multiplication & addition are linear,
that is (a working definition of linearity is) a*(X+Y) = aX + aY.

In fact, you do not need a nonlinear device to perform modulation. All you need
is a device with two inputs which is capable of forming a (linear) product. A
transistor operated in its linear range (with small signals) would effect modulation without
non-linearity if one input was on the base and another was on the collector.
They make special multiplying op-amps for this purpose as well.
These are still linear devices if operated properly.

A nonlinear transfer function will result in a different type of spectral change than
a linear mixer such as in a radio receiver.

An audio mixer is an adder. A modulator is a multiplier. Both are typically linear.

Hope this helps...

Edit: BTW, an audio mixer become a modulator if you slide the volume
pots up and down really fast. But you are multiplying the volume pot signal
with the audio signal.