AM + FM = ?

[John Cuthber]

My point is that any FM signal is AM and any AM signal is FM.

Since the distinction is artificial, the combination is too.

[studiot]

My point is that any FM signal is AM and any AM signal is FM.

Since the distinction is artificial, the combination is too.

 

That is not strictly accurate, as shown by my simple mathematical description in post#17

[John Cuthber]

OK, what does the spectrum of, for example, a 1 MHz carrier, amplitude modulated by a 1 KHz sine wave look like?

Not all the power is at 1MHz is it?

So, some of it is at other frequencies (1,001,000 Hz and 999,000 Hz to be specific).

So you have modulated the frequencies you transmit.

 

The distinctions are artificial.

[Acme]

OK, what does the spectrum of, for example, a 1 MHz carrier, amplitude modulated by a 1 KHz sine wave look like?

Not all the power is at 1MHz is it?

So, some of it is at other frequencies (1,001,000 Hz and 999,000 Hz to be specific).

So you have modulated the frequencies you transmit.

 

The distinctions are artificial.

I'm following all the arguments trying to verstehen. Also have been reading up on tubes and their operation in regards to radio broadcasting. (I noticed that Mordred commented, "Circuitry architect will allow it." )

 

So John, as I understand it the joint modulation you are talking about has its roots in 'natural' errors if you will; i.e. a normal variance inherent in the signal as well as normal variance introduced by the tube that does the modulating. However there is no information in the 'erroneous' signal and I am trying to understand if it is possible to put information there.

I have to wonder why, if there is no distinction, engineers bother to talk about AM & FM at all.

[John Cuthber]

They talk about AM and FM, partly, because those are what the receivers expect and partly because it's a simplification- a model.

The signal that arrives at the antenna doesn't know which "sort" of signal it is.

And, as I said, you can listen to an FM signal with a (slightly detuned) AM receiver.

 

There are other types of modulation too, but the idea is a bit meaningless until you try to "decode" it.

 

And, sorry, but it's not much to do with errors, variances or normal distributions.

Those make the picture more complex

[studiot]

Like I said, calling tha AM sidebands FM is not strictly accurate.

 

The amount of modulation is a recoverable function (preferably one of simple proportionality) to the modulating signal.

 

This is not the case with AM sidebands.

 

Look at the maths.

[John Cuthber]

It seems we are using different definitions.

I'm taking a rather simple view.

If the frequency spectrum (it's never just one frequency once you have a carrier and a signal) changes "in time with the music" then there's frequency modulation happening.

 

You seem to think two mutually exclusive things

 

(a) It's only FM if it's a shift in the carrier frequency that is (ideally) proportional to the signal amplitude and

(b) It's also FM if it's a shift in the carrier phase that is (ideally) proportional to the signal amplitude.

Well, if b is true then a isn't; and once you accept that not all FM is the sort that's used for commercial stereo broadcasts, then ruling out AM is an arbitrary decision.

[studiot]

I didn't say anything of the sort.

 

The broadcast signal has an instantaneously unique frequency and amplitude.

It is one signal.

 

It is not several signals.

 

The best way I can explain this is to consider a cyclist.

At some particular instant he is cycling North East at 14 mph.

 

Now mathematically we can resolve his velocity into 10 mph North and 10 mph East.

 

But there is still only one cyclist, not two travelling on two different directions.

 

At another instant he may be travelling at a different velocity with differents components in the north and east directions.

 

 

In the same way the broadcast signal is one signal that may be mathematically resolved into components, not in space but in time or frequency.

 

There are more and different and resolutions possible for these generalised coordinates.

[John Cuthber]

You seem to be making my point for me.

There is only one voltage on the antenna at any time, but you can resolve it into two signals, one AM and one FM.

And, as I have said before, you can recover the FM signal using an AM receiver.

Clearly you can resolve the AM signal with an AM receiver too.

So what signal does an AM receiver get in this case?

 

The answer is a distorted mixture of both

[studiot]

You seem to be making my point for me.

There is only one voltage on the antenna at any time, but you can resolve it into two signals, one AM and one FM.

And, as I have said before, you can recover the FM signal using an AM receiver.

Clearly you can resolve the AM signal with an AM receiver too.

So what signal does an AM receiver get in this case?

 

The answer is a distorted mixture of both

 

No not really, but your are claiming they are the same so why do we have two terms?

 

While you are at it perhaps you would like to resolve the following.

 

In an AM signal the carrier frequency does not alter. It is constant.

 

In an FM signal the carrier frequency is constantly varying. The deviation from the carrier centre frequency is proportional to the modulating signal amplitude.

 

 

In an AM signal there is no modulation information in the carrier. In fact it is possible to remove the carrier entirely and some forms of AM do exactly this.

 

In an FM signal all the modulation information is contained in the carier. Removal of the carrier will result in total loss of the information.

 

A clear difference there.

Edited May 6, 2015 by studiot

[Enthalpy]

Also have been reading up on tubes and their operation in regards to radio broadcasting. [...]

 

Tubes disappear from radiocomms now. They have disappeared even from radars. Time to switch to transistors maybe ? This would be an opportunity to meet more recent modulations too.

[EdEarl]

I've not thought about modulation before quite so critically. I think that separating AM and FM signals from the carrier may depend on the AM and FM signals being different frequencies. Audio bandwidth is 30K Hz (more or less). Video bandwidth is about 1M Hz. Carrier frequency is 54M Hz and higher. Harmonics from video in the audio range would be 5th, which are mostly low amplitude.

 

As I've been reading, I realize the broadcast standard has changed over the years, from BW+mono to HD+stereo. The documents don't always tell which standard they refer to, which can be confusing.

[John Cuthber]

"In an FM signal the carrier frequency is constantly varying. "

No, the transmitted signal is constantly varying. The carrier isn't generally there (except, instantaneously, as the transmitted signal goes through zero).

 

"In an FM signal all the modulation information is contained in the carier. Removal of the carrier will result in total loss of the information."

No.

If you were to get a very narrow notch filter and remove the nominal carrier you would still get a signal.

 

"In an AM signal there is no modulation information in the carrier. In fact it is possible to remove the carrier entirely and some forms of AM do exactly this."

Since, as you say, you can remove the original carrier frequency and yet keep the information, it's clear that the information is held by frequencies other than the carrier.

The signal ended up on those frequencies because transmitted power was moved away from the carrier frequency.

That's a change in frequency brought about by the modulation process.

How is that modulation of the frequency not frequency modulation?

[studiot]

 

How is that modulation of the frequency not frequency modulation?

 

 

Are you asking because you actually want to find out the correct answer.

 

 

"In an FM signal the carrier frequency is constantly varying. "

No, the transmitted signal is constantly varying. The carrier isn't generally there (except, instantaneously, as the transmitted signal goes through zero).

 

"In an FM signal all the modulation information is contained in the carier. Removal of the carrier will result in total loss of the information."

No.

If you were to get a very narrow notch filter and remove the nominal carrier you would still get a signal.

 

 

Since you clearly misunderstand frequency modulation I will try once more.

 

In frequency modulation the carrier frequency is varied in accordance with the amplitude of the modulating wave.

The maximum is known as the deviation frequency f_d

[John Cuthber]

Trust me, I know what FM is.

Lets take a concrete example, a 1 MHz sine wave as a carrier and a 1 KHz square wave as the signal we want to transmit. (it looks like I'm back in the world of 1980s pop).

I build a VCO with an output frequency of exactly 1MHz per volt.

And I feed it with a signal that's a square wave centred on 1 volt and with a 1 mV amplitude.

 

 

OK so, when the signal into the VCO is 0.999 volts the VCO produces 999 KHz and when it's 1.001 volts the output is 1,001 KHz.

Since there's bound to be some switching delay on the square wave there will be times when the voltage is half way between the extremes and is 1 volt exactly so the VCO will also produce 1,000 MHz for some small part of the time.

 

Now imagine that I look at the output of that through a spectrum analyser. What would I see?

Well, the answer is simple, three spikes, one at each of those three frequencies. The middle spike's height will vary with the time take to switch.

 

OK so far?

 

Now let's look at another simple example.

I amplitude modulate a 1 MHz sine wave with a 1KHz sine waver (it doesn't matter how- but let's say I use a 4 quadrant multiplier that's obligingly linear).

Now imagine that I look at the output of that through a spectrum analyser. What would I see? I'd get the carrier and the sum and difference tones

Well, the answer is simple, three spikes, one at each of those three frequencies. The middle spike's height will vary with the time take to switch modulation index.

 

 

So, I can generate an FM signal that is exactly the same as an AM signal.*

How do you propose to stop those interfering with each other?

 

* yes, I know, I ignored the phase data in the spectral analysis. It doesn't trouble me much, I can pass the square wave through an all pass filter tailored to tweak the phase response

 

Edited May 6, 2015 by John Cuthber

[studiot]

 

Trust me, I know what FM is.

 

Please stop all this holier than thou stuff and look at the pics, courtesy g3npf.

 

 

The top one is AM the bottom one is FM

 

If you look carefully you can see visually what I am saying.

 

The frequency of the AM carrier is fixed and does not vary.

 

The frequency of the FM carrier varies continuously from (fc-fd) to (fc+fd).

 

Every cycle of the FM carrier has a different period.

 

End of.

[John Cuthber]

First, stop being patronising and wrong.

Second, it's not "end of" unless you answer the points I made

Third, really, answer the questions I asked about spectra.

[studiot]

For the benefit of the OP and others, instead of squabbling further with john cuthber, here is a physics 102 exam paper extract, with answers.

 

This shows, amongst other things, a wave that is both amplitude and frequency modulated.

 

http://www.physics.umd.edu/courses/Phys102/brill/EX2-2.HTM

[John Cuthber]

Rather than citing exam papers or other logical fallacies*, ( It's also irrelevant since my point is not that you can't modulate a signal any how you like, I'm saying that the two forms of modulation are not independent. You may recall me explaining that any FM signal that's fed to a real antenna will be amplitude modulated whether you like it or not.)

you could answer my questions.

How come I can generate two signals , both by modulating a 1MHz sine wave with a 1 KHz signal, that are identical even though one is AM and the other is FM?

 

How would a receiver know which sort of signal it was?

 

 

On a related note, perhaps you would like to tell us what sort of FM system transmission actually uses a signal like the one portrayed in post 41?

In particular, since the frequency varies by a factor of about 3 or 4 to one, what frequency would you state on the application form if you applied for a licence for it?

Citing something unrealistic like that then telling me to look at it doesn't help your case any.

 

 

* it's an appeal to authority, which is a fallacy.

Edited May 7, 2015 by John Cuthber

[EdEarl]

John, consider the case of a person modulating amplitude while sending Morse code at 550MHz. At 10 words per minute dits and dahs are about 0.25 sec long, which would emit about 138M cycles of sine wave. The only harmonics generated from keying would occur at the start and end of the 138M cycles. While the key was held on, the carrier would be a sine wave as pure as possible with modern electronics. You are right that FM modulation occurs upon key down and key up, but it does not appreciably contribute to the entire signal, and does not make reading the signal difficult. With ideal equipment, there would be no harmonics, but we don't have ideal equipment.

Edited May 7, 2015 by EdEarl

[John Cuthber]

Thanks for that.

Let's just consider the Morse code for a start.

To make it easy we can look at a set of consecutive dots, separated by gaps equal to the lengths of the dots (I'm not sure if that meets the standards for good keying, but it's not an unreasonable model)

In effect we have a square wave with a period of half a second, so the frequency of that square wave is about 2 Hz.

 

So, lets see what the spectrum of that looks like (before we use it to modulate anything)

The spectrum of a square wave is a well documented exercise in maths

the results are here

http://en.wikipedia.org/wiki/Square_wave

and the spectrum is an infinite series of increasing frequencies (with the amplitudes slowly dropping away as the frequency rises).

The frequencies are all the odd multiples of the square wave frequencies so that's 2 Hz, 6 Hz 10Hz and so on- without any theoretical limit

 

So, before you use it to modulate the 550MHz signal, there are (in principle) components with higher frequencies than that already present.

There's nothing explicit in the modulation that removes those frequencies.

So the spectrum of the transmitted Morse signal includes frequencies that are higher than the carrier.

 

Which rather kills this assertion " With ideal equipment, there would be no harmonics, but we don't have ideal equipment."

Ideal equipment ( a perfect Morse key) would give you a signal with infinite bandwidth.

 

 

Perhaps you would like to comment on my observation that it's possible to have an FM signal and an AM signal that are exactly the same; in particular, how would you stop two signals like those interfering with each other?.

[EdEarl]

I consider perfect equipment would turn the sine wave generator on starting at zero amplitude and it would turn out a perfect sine wave for 138M cycles and turn it off when it gets to zero amplitude. The square wave you are talking about would only cause harmonics on the carrier if the carrier is turned on at some amplitude other than zero, whereupon it would jump instantly to some amplitude, causing harmonics as you say.

[Mordred]

A good example of what John Cuthbert is describing is CTCSS. Which is a form of selective calling also referred to as a privacy or PL tone.

 

Here everyone shares the same carrier frequencies. The modulation scheme causes minute variations in the Carrier frequency. DTMF is another formulation. It is amplitude modulation most typically used in telephones but is also involved in two way radios at any Carrier frequency.

 

Kenwood also has an NXDN scheme which albeit digital involves several modulation levels on its digital signals this includes frequency and amplitude modulations.

 

Motorola's digital signal is also has data compression schemes that are similar.

 

The three forms of modulations are amplitude, frequency and phase shift.

 

These are oft referred to as subtones. They can and are used in combination or as standalone today in both analog and digital compression schemes.

 

The main problems with these schemes is antennae sensitivity and distortions. However that can be overcome via various filters, and amplifying the subtones for processing. Lol I lost count on the number of VCOs I've had to replace.

 

I've also lost count on radios that I've repaired that could pick up the Carrier frequency but couldn't pick up the privacy tone, or in rare cases the DTMF tone,

Circuitry wise you have your main carrier circuits, then the individual tones need to be processed through seperate filters and amplifiers to rectify. Some rectifier circuits have to use error correction schemes to batch missing data. The AM FM scheme has changed considerably in two way radio usage today. The old definitions on modulation has considerably changed.

 

Every radio company has its own unique data compression techniques, Some reveal how it's done others only hint. ICOM has some good papers on NXDM. Certainly doesn't look anything like digital on a spectrum analyzer. That's because the digital voice modulation merely rides the carrier

( in point of detail you need spdcialized radio spectrum analyzers to read specific digital schemes such as Motorola and Kenwood. Each have to have the specific scheme decoders hard wired into them.

 

One name brand is the R8000 spectrum analyzer. You can get Motorola digital specific, or Kenwood as well as the combination of the two

 

 

Here is a brief coverage of FSK as one example, (NXDN radios utilyze this scheme)

 

 

http://en.m.wikipedia.org/wiki/Frequency-shift_keying

 

it has a small list of various techniques.

Here is a multi purpose FSK/ASK VCO.

 

Frequency shift key/amplitude shift key voice control oscillator.

 

http://www.analog.com/media/en/technical-documentation/data-sheets/ADF7020-1.pdf

Lol some of those schemes I never heard of before. Not surprising I dealt primarily with Motorola and Kenwood radios. In one of my past jobs

(Side note you don't want to run too many simultaneous Shift keys at the same time causes too many distortions. Time management is a preferred prevention measure)

[John Cuthber]

I consider perfect equipment would turn the sine wave generator on starting at zero amplitude and it would turn out a perfect sine wave for 138M cycles and turn it off when it gets to zero amplitude. The square wave you are talking about would only cause harmonics on the carrier if the carrier is turned on at some amplitude other than zero, whereupon it would jump instantly to some amplitude, causing harmonics as you say.

In order to understand why that's wrong you need to look at Fourier analysis and, in particular, at apodization.

[EdEarl]

In order to understand why that's wrong you need to look at Fourier analysis and, in particular, at apodization.

I'm aware of Fourier analysis, ringing, dampening, etc. I'm trying to understand the fundamental issue being discussed. Perhaps it is people would like to communicate more than the available bandwidth permits; thus, engineers compress signals thereby reducing quality.