I was getting quite long with that last post, and it is a pain when the forum tells you you used too many words, so I'll finish up here.

The AAC book has several good sections on crystal radios, as it should as they are fundamental to RF in general.

Using one of the illustrations I'm going through how one of the other concepts taught beginners is a bit off. Again, it is so useful it will never go away.

http://www.allaboutcircuits.com/vol_3/chpt_9/6.html

Here we are shown that we rectify the RF, and the varying amplitude spits out the audio. Jives nicely with varying amplitude RF carrier wave theory, doesn't it?

But try looking at it another way instead. The carrier wave is being heterodyned by the diode with the side bands. Given the difference frequencies are the audio, audio out is what you get.

Back before diodes could be bought off the shelf, you had to make your own. Once scheme involved using razor blades and a cats hair whisker. It wasn't a real cat hair, just a fine wire. They worked extremely well. Another method was to uses a quartz or germanium crystal and find a sweet spot. The commercial equivalents are now a Schottky Diode (recommended) or a germanium diode (also superior). Both work much better than a silicon diode, due to the the vastly reduced voltage drop, plus a Schottky diode is fast, much faster than most conventional diodes, a good thing when dealing with RF.

Why are models important? Because if you really understand what is happening you can do new and unique things.

Let us say that we choose to suppress the carrier. That carrier is costing a lot of money, if it is a million watts this power has to be paid for. If you suppress it, you no longer have to pay for it. OK, so lets get rid of one side of the sidebands. Another 50% power reduction, more money saved.

Plus, something interesting happens. Your power is now a fraction of what it was, but your transmitter may still be capable of a million watts. So you boost the power of the remaining sideband. It's range is vastly extended, and if there is no modulation there is no power transmitted. You are broadcasting, but not using electricity unless there is information to be shared, and when you are broadcasting information it travels much, much further with no extra cost. You are also using half the RF spectrum that you were, more free swag.

The down side is that ultra simple receiver shown above no longer works. It needs the carrier, to recover the modulation.

With the advent of precision electronics we can now derive the carrier to an exact degree though, so transmitting sidebands by themselves is practical. This wasn't the case in the beginning of radio, to make it work someone had to tweak a cranky oscillator that drifted frequently to make it work, it was not worth the hassle.

The old NTSC television signal and FM stereo use both AM and FM modulations schemes merged together. The stereo on the TV audio is a 38Khz SSB Upper sideband riding on the FM modulated signal that carriers the audio. I once built a TV stereo sound receiver by modifying a cheap stereo FM radio to use a cable box whose output was channel 3 to pick up on the sound channel (I bumped the FM sound channel of the TV signal to the FM band with a balanced mixer, in other words, frequency conversion), the schemes used between NTSC stereo sound and FM stereo are very similar, adjusting the resistor that controlled the PPL on the FM receiver to the new carrier carrying the stereo channel was a piece of cake.

There are a lot of AM schemes that are part of something else out there, it is very common.

 

So the basic explanation is without considering detailed signal analysis, and without prerequisite to understand Fourier maths (for instance).

It's a bit off topic but I consider it quite interesting.
Having a small oscillator, based on two coils, and one transistor, I modified it for PNP (original circuit uses PNP). There are two capacitors as well.

One day I built it again in NPN technology, but I aligned the coils close to each other. I figured out, the capacitors can be omitted. And as I moved the coil in 3 dimensions, I would get variation in LED intensity (the circuit is some kind of a joule thief).

However, I was not able to do this with the PNP circuit, even if it is exactly symmetrical, and is using the same components.

The PNP circuit however also is weird, no matter which way I insert the transistor, it works. The NPN does not. 2n3904/2n3906 by the way.

In the PNP circuit alignment of the coils seems to have little significance.

 

http://www.allaboutcircuits.com/vol_3/chpt_9/6.html

good work, if you for instance consider various webpages, where bad copies from original ATX-type SMPS are cobbled together, without explanation, sometimes the quality is so bad it's even impossible to view it properly.

One single well explained schematic is far more useful.

And it's not always the case the person who wants to use it would be unable to follow the technical concepts- just specialization might be totally different.

So it's not unusual to encounter web pages with perfectly complete explanation- from the view of the author. From the view of the user, these are often hopelessly incomplete, the prerequisites are simply not there. I mean the users are not always hobbyists or total beginners who can't count things together. Sometimes they might be of minor age and not fully able of abstract thinking, but some people out there are surprisingly clever.

Wikipedia-like formula's and calculations are not very useful as such- some people never really use these. They rather belong inside universities, or printed literature.

 

We once had a gentlemen who was a major troll (banned now), who could not get the concept that sometimes a wrong concept is OK, if it taught the basics. He just could not wrap his head around that sometimes a beginner needed it kept simple, and if the interest was there they could build on it later.

He kept insisting on correcting the basics, even though people new to the field had to understand the textbook first, the reality later.

In many ways conventional vs. electron flow is about this. When you need better models they will still be there for you, waiting.

His assertion that BJT transistors were actually voltage controlled would have been humorous, but he manage to make a lot people pull their hair out. Funny thing, the model is valid, if viewed in a certain light, but he was unable to accept that his models were not the only ones.

When Tony, the writer of the AAC book, came back stating he was thinking of adding that model to the book the reception he got was major funny, we thought the other guy had come back as was pretending to be Tony.

There is a section of the AAC book I keep loosing and have trouble finding. If some sees it please send me the link. It is a hand drawing of a crystal radio. It is not referenced very well, next time I see it I am going to pin that sucker down.

The AAC book is very deep, and yet has blank spots. It continues to surprise me.

 

http://www.allaboutcircuits.com/vol_3/chpt_2/7.html

I added a couple of chapters to the posts.

Why am I going to all this trouble? For my blog, of course. I will clean them up and repost them, this is not the first time I've referred to the AM paradox, where a flawed model is taught to beginners with the understanding true understanding of RF can come later.

 

Tony, the writer of the AAC book, came back stating he was thinking of adding that model to the book the reception he got was major funny, we thought the other guy had come back as was pretending to be Tony.

That's hilarious. Is that thread still around? Although I guess if the troll was banned his messages would be gone.

In the book "Art of Electronics" they start out with the current-controlled model with beta, but very quickly point out the limitations of it and introduce the voltage-controlled model and simplified Ebers-Moll and then keep both models close at hand for the rest of the book. I only mention it because I've been looking at that book a lot lately.

 

I think one could readily recover the modulating signal using a simple envelope detector. I don't think you would need to re-instate the 'missing' sideband in a typical bog standard AM receiver.

My understanding is that the reason that SSB isn't used very widely (is was (is?) used extensively for specific applications, such as telephony) is precisely because of the needed complexity of the receiving equipment. I haven't delved into it enough yet to explain it any better, but I would suggest that if it was that easy to recover the signal that they would have used it for broadcast AM in order to increase the number of channels available. But, again my understanding, is that they chose AM-DSB-C specifically because it kept the receivers simple and cheap.

 

Amplitude modulation is the modulation of an AC signal by an AC signal with different frequency.

This seems pretty... useless. The fact that the word "modulation" is used both as part of the definition and as part of the term being defined makes it circular. Unless the definition is specifically saying what the "amplitude" descriptor means. But, in either case, this implies that FM, PM, and any and all other modulations schemes are NOT the modulation of an AC signal by an AC signal with a different frequency.

 

One of the reasons I dispute the model where you vary the carrier amplitude is the carrier does not actually vary amplitude, it is the sidebands that do all the varying.

But this is pure semantics. If I pass a signal through an amplifier that have variable gain and I reach over and turn up the gain, have I not varied the amplitude of the signal? If I sit there an turn the the gain up and down in response to the beat of the music, am I not varying the amplitude of signal? If I hook up a circuit that varies the gain in response to the signal coming from a microphone as I speak into it, am I not varying the amplitude of the signal? Whether you speak of the variations imposed on the amplitude of the signal as creating the sidebands or you talk about the sidebands (spectrum of the signal controlling the gain) varying the amplitude is six-to-one half-a-dozen to the other (each description has merit and value).

A statement was made that AM is not linear, in an ideal system I would dispute that too. The sidebands are spectrally identical to the audio.

But you can only move the audio to the sideband of the carrier through a nonlinear process. In your first post in this thread you talk about a 1MHz signal and a 1.001MHz signal and how that produced the same spectra as mixing a 1kHz signal onto a 1MHz carrier. Leaving aside the other sideband, the question still remains how you got the 1.001MHz signal from the 1kHz signal to begin with. It had to be through a nonlinear process because linear systems can only change the amplitude and phase of frequency components in the signal, they can't produce signals at new frequencies.

 

I know that, as so many other matters, explaining something like AM implies a certain degree of complexity.

In my activity I am used to go straight with actual definitions of whatever is involved, where what is told could vary in the degree / amount of details given but never a model that later, strictly speaking, could prove wrong / inexact (and maybe misleading but certainly confusing if not giving way to doubts.)

Sorry Bill, I know I cannot be player in this game, not an EE myself, far from that, but I feel that teaching things straight from the beginning is the way to go.

I do not like the idea of being driven into confusion later. What for? Let the effort of understanding be spent in the right track from square one.

Isn't that being paternalist with a newbie thinking he will not understand?

I hate the idea of doing that myself.

 

My understanding is that the reason that SSB isn't used very widely (is was (is?) used extensively for specific applications, such as telephony) is precisely because of the needed complexity of the receiving equipment. I haven't delved into it enough yet to explain it any better, but I would suggest that if it was that easy to recover the signal that they would have used it for broadcast AM in order to increase the number of channels available. But, again my understanding, is that they chose AM-DSB-C specifically because it kept the receivers simple and cheap.

I suppose it depends what SSB means in the context of this discussion. To me SSB would mean only one sideband transmission without carrier - SSBSC [Single Sideband Suppressed Carrier]. The abbreviation seems to often leave of the suppressed carrier bit. If one transmits one sideband plus the carrier will the basic AM receiver reproduce the applied modulation without any issues? I can't see why it couldn't happen. Certainly receiving and correctly demodulating SSBSC on a basic AM receiver isn't a given - rather the contrary would be expected.

 

I suppose it depends what SSB means in the context of this discussion. To me SSB would mean only one sideband transmission without carrier - SSBSC [Single Sideband Suppressed Carrier]. The abbreviation seems to often leave of the suppressed carrier bit. If one transmits one sideband plus the carrier will the basic AM receiver reproduce the applied modulation without any issues? I can't see why it couldn't happen. Certainly receiving and correctly demodulating SSBSC on a basic AM receiver isn't a given - rather the contrary would be expected.

My understanding at this point is that there are lots of variants, but they come down to a combination of: {DSB, SSB, VSB} and {FC, PC, SC}. Pick one from each group.

DSB: Double-side band
SSB: Single-side band (choice of upper or lower, usually upper)
VSB: Vestigial side band (SSB on one side and part of the other sideband)

FC: Full carrier
PC: Pilot carrier (or RC, reduced carrier) (carrier partially filtered, but some remains)
SC: Suppreseed carrier (carrier removed as completely as possible)

My understanding is that the problem with not having a carrier signal to sync to, the receiver is extremely difficult to tune since a mismatch of a few tens of hertz can make the signal annoying or even unintelligible by the time you start getting toward 100Hz.

Now, what I don't have a good feel for is what "syncing to the carrier" means in the context of a simple, ordinary receiver. I get that with a really symbol receiver, we tune it manually until we get a sound we enjoy. Does this really mean that we have tuned it to within a handful of hertz of the received carrier? If so, then what roll did the presence of the carrier play in any of this, since it would seem that we are adjusting it based on the quality of the reception of the sideband and getting it frequency translated to its correct place in the baseband? Wouldn't we have tuned it just as well if the carrier hadn't been there? If so, then why wouldn't we have tuned it acceptably if there were only one sideband present? I have to assume that we wouldn't have, otherwise there would have been a huge press to use SSB-SC for AM Broadcast back in the days when almost all receivers were manually tuned and before the advent of commercial FM as a big player.

I don't know the answers to these questions, but I plan to.

 

But this is pure semantics. If I pass a signal through an amplifier that have variable gain and I reach over and turn up the gain, have I not varied the amplitude of the signal? If I sit there an turn the the gain up and down in response to the beat of the music, am I not varying the amplitude of signal? If I hook up a circuit that varies the gain in response to the signal coming from a microphone as I speak into it, am I not varying the amplitude of the signal? Whether you speak of the variations imposed on the amplitude of the signal as creating the sidebands or you talk about the sidebands (spectrum of the signal controlling the gain) varying the amplitude is six-to-one half-a-dozen to the other (each description has merit and value).

I guess the overall point is that the transmitted AM signal carrier power is invariant with the modulating signal amplitude & frequency.

 

Now, what I don't have a good feel for is what "syncing to the carrier" means in the context of a simple, ordinary receiver. I get that with a really symbol receiver, we tune it manually until we get a sound we enjoy. Does this really mean that we have tuned it to within a handful of hertz of the received carrier? If so, then what roll did the presence of the carrier play in any of this, since it would seem that we are adjusting it based on the quality of the reception of the sideband and getting it frequency translated to its correct place in the baseband? Wouldn't we have tuned it just as well if the carrier hadn't been there? If so, then why wouldn't we have tuned it acceptably if there were only one sideband present? I have to assume that we wouldn't have, otherwise there would have been a huge press to use SSB-SC for AM Broadcast back in the days when almost all receivers were manually tuned and before the advent of commercial FM as a big player.

I don't know the answers to these questions, but I plan to.

The simplest of receivers - such as a crystal receiver - doesn't need to exactly tune to the carrier frequency because the carrier of itself has no modulation information. As anyone knows who has built a crystal set, the issue with such a simple receiver is the poor selectivity.

 

Hello,

AM modulation will need a non-linear element to perform the modulation.
Just adding the signals will not work.
See this page for more info on modulation techniques:
http://www.williamson-labs.com/480_mod.htm

Bertus

 

I guess the overall point is that the transmitted AM signal carrier power is invariant with the modulating signal amplitude & frequency.

I don't see how that could really be the case. If a transmitter is putting out a constant total amount of energy into the spectrum, that energy has to be divided up between the carrier and the sidebands. If there are no sidebands (the transmitter is transmitting, but there is no signal to modulate the carrier), then all of the power is in the carrier. If there is a signal, then power in the carrier has to be reduced.

I suppose that a transmitter could be designed so that it adjusted its output power so that the carrier always had the same power, but I don't see that as inherent in the modulation scheme.

 

I don't see how that could really be the case. If a transmitter is putting out a constant total amount of energy into the spectrum, that energy has to be divided up between the carrier and the sidebands. If there are no sidebands (the transmitter is transmitting, but there is no signal to modulate the carrier), then all of the power is in the carrier. If there is a signal, then power in the carrier has to be reduced.

I suppose that a transmitter could be designed so that it adjusted its output power so that the carrier always had the same power, but I don't see that as inherent in the modulation scheme.

I would base the statement on the general AM function of the form

\(f(t)=\(1+msin(\omega_m t)\)A_csin(\omega_ct)\)

where m is the modulation index.

The effective RMS value of f(t) for various values of m varies but the carrier RMS value [at frequency ωc] remains unchanged. Only the total sideband RMS value changes in response to changing modulation index.

 

But this is pure semantics. If I pass a signal through an amplifier that have variable gain and I reach over and turn up the gain, have I not varied the amplitude of the signal? If I sit there an turn the the gain up and down in response to the beat of the music, am I not varying the amplitude of signal? If I hook up a circuit that varies the gain in response to the signal coming from a microphone as I speak into it, am I not varying the amplitude of the signal? Whether you speak of the variations imposed on the amplitude of the signal as creating the sidebands or you talk about the sidebands (spectrum of the signal controlling the gain) varying the amplitude is six-to-one half-a-dozen to the other (each description has merit and value).



But you can only move the audio to the sideband of the carrier through a nonlinear process. In your first post in this thread you talk about a 1MHz signal and a 1.001MHz signal and how that produced the same spectra as mixing a 1kHz signal onto a 1MHz carrier. Leaving aside the other sideband, the question still remains how you got the 1.001MHz signal from the 1kHz signal to begin with. It had to be through a nonlinear process because linear systems can only change the amplitude and phase of frequency components in the signal, they can't produce signals at new frequencies.

Fraid not. The fact is, if you measure the carrier it does not vary. If you filter the sidebands out you will have a steady state signal. This is more than semantics, this is a reality.

The act of modulation, which is what you are describing, creates the sidebands. This is the reality. The act of modulating a RF carrier is pure heterodyne. You are creating new frequencies, this does not mean you are modifying the old ones in any significant manner.

The reason you can suppress the carrier, or reject a carrier, is because they are separate and distinct signals, so you can pick and choose among them.

A spectrum analyzer is a piece of test equipment that measures exactly that. So is a selective voltmeter. Interestingly, they are modified AM radios. On the shop floor the techs would take selective voltmeters and configure them to be AM radios, before I figured out what they were doing it would puzzle me where the audio was coming from.

When we build things like transmitters we are trying to get as close to theory as we can, anything else is distortion. Mathematically though, a signal that is slowly cycling has a sideband, and what you are actually seeing is an interference pattern. I am not well versed in Fourier Analysis, but as I understand it this is pretty much the basis for it. Any repeating waveform, be it AM modulation or a triangle wave, can be broken down into many steady state (constant amplitude) frequencies.

If you have a amplifier whose gain can be varied electrically, up to 20Khz, you have an AM modulator. Interestingly, you also have a precision heterodyne circuit. You can not separate the two.

When I worked for Collins Radio I saw a lot of interesting devices, some of them I never really understood but they were neat to play with. Mechanical filters are a bunch of metal disks connected by wires spot welded on the sides of the disk (disks being about 1" X 3/8"). Somehow these made precision filters with really great rejection outside their bandpass. It made Art Collins rich, and was the basis of a lot of telephone circuits. Every voice channel had two of these, RX and TX. Needless to say, back then there were huge numbers of racks that were nothing but these modems (not to be confused with a computer modem). The hard part for the techs in the field was which cards to pull and replace among the thousands.

When I went to college in 77 I figured I would wind up with a job at a radio station. Fact is, college was a bare start, I have never stopped learning, and Rockwell Int., Collins Division was fun. I really miss that job.

Anyhow, aren't models fun?

 

If it is as you explain, Bill, calling it AM (amplitude modulation) is deadly wrong. No amplitude is changing here...

And how is then, they get this in the scope?

 

The act of modulation, which is what you are describing, creates the sidebands.

Which is what I've been saying. You're the one who has been claiming that AM modulation is the linear addition of the sidebands.

This is the reality. The act of modulating a RF carrier is pure heterodyne.

Which is a nonlinear process.

If you have a amplifier whose gain can be varied electrically, up to 20Khz, you have an AM modulator. Interestingly, you also have a precision heterodyne circuit. You can not separate the two.

Which why it is semantics to say that one description is wrong and the other is right.