Did You Hear of Such an AM Receiver? For DSB-Suppressed Carrier.

[KerimF]

Note1: The academic purpose of this thread could be read on reply #72
Note2: More details on replies #9 (LTspice), #14 (AM-FM equation), #40 (DSBSC vs SSBSC), #44 (to gurus in electronics)

The receiver of interest demodulates a special double sideband suppressed carrier (DSB-SC) signal.

The modulating signal is of speech (narrow baseband) or music (wide baseband).
The frequency of the suppressed carrier is 455 kHz (IF, Intermediate Frequency in AM receivers).
The carrier is also frequency modulated by 300 Hz sinewave with a frequency deviation +/- 30 KHz. (It may be called, therefore, FM-DSB-SC.)

To my knowledge, a graduate engineer from any faculty of RF communications wasn’t taught how it could be done, though a hobbyist can build it.
It doesn’t need I-Q signals (Costas Loop) or LC selective filter (Squaring method).
Its reliable simple topology could be integrated as a low-cost IC. This IC doesn’t exist because even an IEEE professional in AM communications has no idea how to do it.

But I suspect a few members here may have heard of it, right?

Happy NEW YEAR 2025

Kerim Fahme

Hint: It is based on conventional PLL.

[pardo-bsso]

We had the same conversation some months ago but I can't find the link right now

[KerimF]

Sorry for the delay, the battery of my old lap (15 years) decided to die now (right after I started this thread).
I couldn't replace it (or buy a new lap), once every few years, because of the sad events in my region since year 2011.
To avoid a crash now, I have to keep the lap in state of charging all the time. The problem is that we have at best 4/24 mains AC only, so we depend on external inverters/batteries during the rest of the day.

https://www.eevblog.com/forum/rf-microwave/a-simple-reliable-double-sideband-suppressed-carrier-(dsb-sc)-demodulator/

As you see, the previous thread was in the RF forum (around July 2023). I didn't continue it there after I realized that discussing of this DSB-SC demodulator (still unknown universally, though I had the chance to discover its basic topology since fall 1979 as an MS thesis) is related more to electronics than today's RF.

The irony is that the local people (even so-called professors in communications) still tell me that it is impossible for a local engineer (in an underdeveloped country) to know a technical practical idea not known in the developed ones. This doesn't surprise me because, on the internet, even the engineers in the developed countries find really hard to believe a non-familiar topology (or be interested to verify it for their own knowledge in the least) as long it is not approved first by certain top representatives of today's Science.

Naturally, if no one here likes (or is capable) to discuss it (this non-familiar topology of AM demodulation) in depth, electronically (formulas, parameters... for various applications), I agree with you that there is no need to keep this thread alive in this forum too.

[radiolistener]

Let's use matlab/octave.

Usual AM can be written as:

Code: [Select]

Fs = 384000;     % sample rate
N = Fs*3;       % sample length 3 seconds

A = 1;          % Carrier amplitude
M = 0.8;        % Message amplitude
m = M / A;      % Modulation index m = M / A

carrier = A * cos(7000 * 2*pi*(0:N-1)/Fs);
message =     cos(1    * 2*pi*(0:N-1)/Fs);      % Normalized

signal_am = (1 + m*message) .* carrier;

Could you write formula for your modulation, so we can test it?

[SteveThackery]

@KerimF: could you explain more clearly what is the purpose of this thread? I think then you would get more engagement.

The thread title is "Did you hear of...", to which the only answer is "Yes" or "No" - hardly the basis for an interesting discussion. Or is it supposed to be a test for forum members, to see if they can design a circuit that will receive such a signal? Or are you hoping to discuss the benefits of double sideband suppressed carrier? Please explain clearly what you want to discuss.

I learned about various modulation systems several decades ago, and I seem to recall that DSSC was a standard AM signal with the carrier not transmitted. Even more economical for transmitter power is single sideband suppressed carrier.  My understanding is that both carriers contain the same information, so there is no benefit in transmitting both.

So there you go: the sum total of my "knowledge". 

My question is this: what is the purpose of frequency modulating the carrier at 300Hz?  And don't forget to tell us what you want from this thread.

[KerimF]

Let's use matlab/octave.

Usual AM can be written as:

Code: [Select]

Fs = 384000;     % sample rate
N = Fs*3;       % sample length 3 seconds

A = 1;          % Carrier amplitude
M = 0.8;        % Message amplitude
m = M / A;      % Modulation index m = M / A

carrier = A * cos(7000 * 2*pi*(0:N-1)/Fs);
message =     cos(1    * 2*pi*(0:N-1)/Fs);      % Normalized

signal_am = (1 + m*message) .* carrier;

Could you write formula for your modulation, so we can test it?

Hi,

As you said, the above formulas are for the usual AM signal when the carrier is not suppressed, that is 0<m<1. In this case, the polarity of the carrier doesn't reverse.

If m>1, the carrier polarity reverses at zero crossing of the modulating signal.

I don't have Matlab (or the like). I guess you know that, in case the carrier is suppressed, A=0 and m=infinity. And let us assume you will write a code for this case (A=0), will Matlab tell us how to recover the missing carrier (frequency and phase) to recover the modulating signal?

For instance, do you have LTspice?

[CaptDon]

An I.C. to demodulate such a signal is not a product of high value since there is no practical purpose or commercial usage of such a signal. There is a signal which uses two forms of modulation at the same time. It is the modulation scheme used for the VHF Omni Range or the VOR device common in 99.9% of aircraft. As far as demodulating your 'example' is the 300Hz with 30KHz of deviation steady or does it vary? Are you wanting to recover both the 300Hz signal and the DSB modulation or just the DSB riding on a carrier of useless complexity? Is your 30KHz deviation actually + and - 30KHz for a total of 60KHz bandwidth? That is over 10% of the 455KHz bandwidth and would probably lead to some residual AM component due to bandwidth rolloff of the receiver I.F. strip. Can it be designed, yes, by an amateur radio technical hobbyist, probably yes. Can a college student design it and build it as a graduate thesis, I would say yes. Is there any point to building it, any use for it, probably NO.

[KerimF]

@KerimF: could you explain more clearly what is the purpose of this thread? I think then you would get more engagement.

The thread title is "Did you hear of...", to which the only answer is "Yes" or "No" - hardly the basis for an interesting discussion. Or is it supposed to be a test for forum members, to see if they can design a circuit that will receive such a signal? Or are you hoping to discuss the benefits of double sideband suppressed carrier? Please explain clearly what you want to discuss.

I learned about various modulation systems several decades ago, and I seem to recall that DSSC was a standard AM signal with the carrier not transmitted. Even more economical for transmitter power is single sideband suppressed carrier.  My understanding is that both carriers contain the same information, so there is no benefit in transmitting both.

So there you go: the sum total of my "knowledge". 

My question is this: what is the purpose of frequency modulating the carrier at 300Hz?  And don't forget to tell us what you want from this thread.

Hi Steve,

The two purposes of this thread are:
[1] To hear a "Yes" or "No".
[2] In case someone is curious about it, and he is good in math and electronics, I can start with him a technical discussion after providing all what I did, since no one is perfect, to see how it could be updated and be applied for certain applications.

You are right, the advantages of SSB-SC are narrower bandwidth and lower power (half).

On the other hand, the advantage of DSB-SC, in case the simple reliable DSB-SC demodulator is used (not of Costas Loop or Squaring method), is that the proper reception is not sensitive to the carrier frequency variation, at the transmitter and receiver sides.
I gave an exaggerated example by modulating the carrier frequency at 300Hz with a relatively wide bandwidth +/- 30 KHz (covering 6 channels on MW band).
For instance, I used this added frequency modulation, though at 6Hz only, as a scrambling way in my private RF voice links in the 80's. Listeners on conventional AM receivers (on MW band) used to hear sort of noisy interference every time a voice signal was transmitted. (I also took advantage of this demodulator on FM band since it doesn't need a pilot. But this has another story).

By the way, in 1979 at the university, the idea of my MS thesis was to prove that receiving more information (two symmetrical sidebands) should let the recovery of the modulating signal simpler and more reliable/practical than receiving less information (one single band). This was the main idea behind this novel DSB-SC demodulator.
After 3 months of consecutive failures at the university lab, I built, during the last 3 days and as the last design I thought of, the first prototype of this demodulator. To my big surprise, it worked very well though it shouldn't work at all, in theory. Then, I had to return home for financial reason where I rebuilt it and was able to find out the 'minor' imperfection in the prototype that let the PLL be locked to the suppressed carrier.
Of course, after 45 years, I had enough time to analyze almost everything about its topology.

In brief, the big image of the DSB-SC system is completed, its transmitter and receiver are simple to build (relative to other systems) besides the frequency lock function (one example, for controlled toys, in case the low-cost DSB-SC demodulator IC is made).

Kerim

[KerimF]

An I.C. to demodulate such a signal is not a product of high value since there is no practical purpose or commercial usage of such a signal.

You are totally right since FCC doesn't give licenses for DSB in general.

As far as demodulating your 'example' is the 300Hz with 30KHz of deviation steady or does it vary? Are you wanting to recover both the 300Hz signal and the DSB modulation or just the DSB riding on a carrier of useless complexity? Is your 30KHz deviation actually + and - 30KHz for a total of 60KHz bandwidth? That is over 10% of the 455KHz bandwidth and would probably lead to some residual AM component due to bandwidth rolloff of the receiver I.F. strip.

I deliberately exaggerated the FM modulation to show its very low sensitivity to the variation of the carrier frequency.
And yes, by adding a LPF at VCO_in, the demodulator recovers two modulating signals simultaneously (from AM and FM).

Can it be designed, yes, by an amateur radio technical hobbyist, probably yes. Can a college student design it and build it as a graduate thesis, I would say yes. Is there any point to building it, any use for it, or any point to this conversation, NO.

I agree with what you said.

For instance, to make money from a novel idea, there will be a need to think of novel applications for it as well.
Unfortunately, I am not rich to start novel commercial projects that take advantage of this simple low-cost topology.

[KerimF]

Those who are still enthusiastic in electronics basics may like to comment technically on the uploaded example. Its Vcc=10V so that CD4046 can be used at 2*fc [910 kHz +/- 155 kHz].

I also uploaded screen shots to those who don’t have LTspice, to run the files in 'FM_DSB-SC_455K_10V_v1.zip'.



Any question is welcomed.
Please read my few previous posts to avoid repeating things.

[geggi1]

If you want to simulate the modulation and demodulation you can look into GNU-radio that is a freeware. Together with a SDR tv dongle you can demodulate almost anything.
As a replacement for Matlab have a look into Scilab that is a freeware equivalant.

[KerimF]

FCC licences are given for 4kHz bandwidth in general. This is good to transmit voice by SSB.

It may be possible to transmit voice by using DSB-SC in 4 kHz bandwidth if we apply the technique of 'Narrow Band Voice Transmission'.
But I am not sure if the added complexity at the baseband side is compensated by simplifying other parts in the transmitter and receiver in order to take advantage of the frequency lock.

[KerimF]

If you want to simulate the modulation and demodulation you can look into GNU-radio that is a freeware. Together with a SDR tv dongle you can demodulate almost anything.
As a replacement for Matlab have a look into Scilab that is a freeware equivalant.

Thank you for your valuable suggestions. Unfortunately, from where I live, I can't access anymore (since around 15 years ago) any technical tool offered, for free or not, on the internet... due to the world's sanctions applied on the people among whom I was born and live.

Would you please notify me if any of these tools can help you design the receiver on post #1 (just its DSB-SC demodulator, when the suppressed carrier is frequency modulated)?


 

[radiolistener]

Thank you for your valuable suggestions. Unfortunately, from where I live, I can't access anymore (since around 15 years ago) any technical tool offered, for free or not, on the internet... due to the world's sanctions applied on the people among whom I was born and live.

I'm using GNU Octave which is a free alternative for Matlab. It uses the same syntax and libraries with the same functions, so you can share your code with Matlab (with some limitations). This is complete free and open tool and is not limited by any kind of "sanctions". You can get its source code and use it for free, the only limitation is to keep all derivative products also free and open with providing its source code.

If you want to analyze your modulation scheme, you first need to express it mathematically. This will allow you to analyze and validate your ideas in a mathematical form, as is customary in modern science. Matlab/Octave simply provides a way to formalize and test your ideas in mathematical form.

For instance, do you have LTspice?

Before doing it in LTspice, we're needs to analyze it in math form, it allows to check if LTSpice circuit is implemented correctly.

As you said, the above formulas are for the usual AM signal when the carrier is not suppressed, that is 0<m<1. In this case, the polarity of the carrier doesn't reverse.

If m>1, the carrier polarity reverses at zero crossing of the modulating signal.

I don't have Matlab (or the like). I guess you know that, in case the carrier is suppressed, A=0 and m=infinity. And let us assume you will write a code for this case (A=0), will Matlab tell us how to recover the missing carrier (frequency and phase) to recover the modulating signal?

DSB signal can be expressed in math with just removing "1 +" field from expression.

Here is example for both AM and DSB modulation for 7000 Hz carrier and 1 Hz message:

Code: [Select]

Fs = 384000;
N = 2^16;

A = 1;       % Carrier amplitude
M = 0.8;     % Message amplitude
m = M / A;   % AM Modulation index m = M / A

carrier = A * cos(7000 * 2*pi*(0:N-1)/Fs);
message = M * sin(1    * 2*pi*(0:N-1)/Fs);
   
% AM modulation:
signal_am = (1 + message/A) .* carrier;

% DSB modulation:
signal_dsb = message .* carrier;

As you can see this "1+" field in AM modulation is multiplied by carrier and so adding carrier to AM modulation output, but it is missing from DSB output. This explains why carrier is missing from DSB modulation in math form.

Here is spectrum of AM modulated signal from example above:


And here is spectrum of DSB modulated signal from example above:


Can you explain your modulation in math form, so we can analyze and test it before implementation?

[KerimF]

@radiolistener
Sorry, I don’t know your syntax. I use Excel only for my calculations.

The time domain equation of the received signal v(t) is:

v(t) = Ac*sin(W1*t)*cos[Wc*t + (dF/F2)*sin(W2*t)]

where:
Ac = 1.8 V
W1 = 2*pi*4000 rad/sec [in case the AM modulating signal is 4 kHz]
W2 = 2*pi*300 rad/sec [in case the FM modulating signal is 300 Hz]
Wc = 2*pi*455000 rad/sec [in case the carrier frequency is IF]
dF = 30000 Hz [the maximum frequency deviation, 30 kHz]
F2 = 300 Hz [see W2 above]

I hope this helps.

[KerimF]

If a carrier (say 100 kHz) is amplitude modulated by a sinewave (say 10 kHz), its spectrum bandwidth is...
If a carrier (say 100 kHz) is frequency modulated by a sinewave (say 10 kHz) with a small deviation (say 100 Hz), its spectrum bandwidth is...
If the 100 kHz carrier is amplitude modulated by a 10 kHz and frequency modulated by a 10 kHz sinewave with a 100 Hz deviation, its spectrum bandwidth is...

[mawyatt]

v(t) = Ac*sin(W1*t)*cos[Wc*t + (dF/F2)*sin(W2*t)]

where:
Ac = 1.8 V
W1 = 2*pi*4000 rad/sec [in case the AM modulating signal is 4 kHz]
W2 = 2*pi*300 rad/sec [in case the FM modulating signal is 300 Hz]
Wc = 2*pi*455000 rad/sec [in case the carrier frequency is IF]
dF = 30000 Hz [the maximum frequency deviation, 30 kHz]
F2 = 300 Hz [see W2 above]

I hope this helps.

The term "(dF/F2)*sin(W2*t)" apparently is not in radians per your definitions? The maximum absolute value of such is +-dF/F2 or +-100, not sure if this is what was intended?

Best

[KerimF]

@mawyatt

Let us assume that B (beta) is the maximum deviation of the carrier phase:

v(t) = Ac*cos[Wc*t + B*sin(Wa*t)]

We get the instantaneous frequency by the derivative of [Wc*t + B*sin(Wa*t)]:

F = Wc/(2*pi) + B*Wa/(2*pi)*cos(Wa*t)
F= Fc + B*Fa*cos(Wa*t)

Therefore :
dF = B*Fa [dF=the maximum frequency deviation and Fa=the frequency of the FM modulating signal]]
B = dF / Fa

The above equation becomes:
v(t) = Ac*cos[Wc*t + dF/Fa*sin(Wa*t)] [corrected]

[mawyatt]

@mawyatt

Let us assume that B (beta) is the maximum deviation of the carrier phase:

v(t) = Ac*cos[Wc*t + B*sin(Wa*t)]

We get the instantaneous frequency by the derivative of [Wc*t + B*sin(Wa*t)]:

F = Wc/(2*pi) + B*Wa/(2*pi)*cos(Wa*t)
F= Fc + B*Fa*cos(Wa*t)

Therefore :
dF = B*Fa [dF=the maximum frequency deviation and Fa=the frequency of the FM modulating signal]]
B = dF / Fa

The above equation becomes:
v(t) = Ac*cos[Wc*t + dF / Fa (Wa*t)]

Those 2 equations are not the same tho!!

Example at T=1 the 1st equations cos argument is {Wc + Sin(Wa)}, while the second equations cos argument is {Wc + (dF/Fa)Wa}!!

Best

[KerimF]

@mawyatt
So, I wonder what your right equation of FM modulation is... v(t)=...

I see... I missed the 'sin', sorry

[mawyatt]

Well all we know is that mathematically in reference to above the max argument for the cos is Wc + B and the 2nd equation cos argument is Wc + (dF/Fa)*Wa. These cos arguments aren't equal @ T=1 unless B = (dF/Fa)*Wa and Sin(Wa) =1!!

Best

[KerimF]

It seems you didn't notice yet that I corrected the second (final) equation on my post #17, thanks to you:

v(t) = Ac*cos[Wc*t + dF/Fa*sin(Wa*t)] [corrected]

Best

[mawyatt]

Didn't note your correction, thanks!!

Best

[KerimF]

I'm using GNU Octave which is a free alternative for Matlab. It uses the same syntax and libraries with the same functions, so you can share your code with Matlab (with some limitations). This is complete free and open tool and is not limited by any kind of "sanctions". You can get its source code and use it for free, the only limitation is to keep all derivative products also free and open with providing its source code.

I wish you are right. But facts are facts.
On our dear planet, Earth, humans have been forced, since always in human history, to live in different, if not divided, worlds (after all, this is how the world is designed, and no one can change it).

[RoGeorge]

Octave is like Matlab, just that Octave is free.  If you can't download Octave from where you are, try using Tor Browser (not torrents, though a torrent client might work too), and visit the Octave download page using the Tor Browser instead of whatever browser you may be using now.  Tor behaves like a VPN, it accesses the Internet as if you were from somewhere else.

https://www.torproject.org/download/
https://octave.org/download