A Simple Reliable Topology to Demodulate an AM DSB-SC Signal

[KerimF]

Hello,

The two well-known topologies for demodulating an AM double sideband suppressed carrier signal (DSB-SC) are of Costas Loop and the Squaring Method.

In this thread, I present a third topology which combines those two known topologies, while being simpler and more reliable.

It is simple because it doesn't need I-Q signals (Costas Loop) or an LC selective filter at 2*fc (Squaring Method). And the values of its Rs and Cs are not critical, also its 4 ICs are low-cost standard ones (CD4046B, CD4013B, CD4066B and 74HC86 in the attached example, for fc=455 KHz).

It is reliable and practical because it is of synchronous detection, and it locks to the suppressed carrier in a relatively wide spectrum. So, even if the frequency of the suppressed carrier varies at the transmitter for a certain reason (because of temperature for example) the AM modulating signal is recovered properly. In case it varies deliberately (transmitting AM and FM simultaneously) the FM modulating signal can be recovered too, by adding another LPF (at VCO_in).


I attached an example as LTspice files to simulate it (AM-FM-DSB-SC_455K_10V_v1.zip).
In this example, I added the FM option to show how the demodulator locks to the suppressed carrier even if its frequency varies in a relatively wide bandwidth (here, fc varies between 455-30 kHz and 455+30 kHz, modulated by a 300 Hz sinewave).

To those who don’t have LTspice, I also attached two screen shots:
AM-FM-DSB-SC_455K_10V_v1_asc.png
AM-FM-DSB-SC_455K_10V_v1_plt.png

Since this topology is not known universally at any university (though I had the chance to build its basic version, as an MS thesis, at the American University of Beirut, AUB, in spring 1979), it needs a name. Let us call it 'Harmonic Loop' (because it locks to 2*fc), since calling it 'Kerim Loop' is out of question in these days.

Cheers,
Kerim

Edited: In the schematic (+BPF...) is (+LPF...).

Edited: I attached the zip file, but it is removed. For the future, does this mean that attaching zip files is not allowed in this site or its content which was LTspice files? Thank you.

 

[poormystic]

Hi
I'm amazed, and will nominate you for the Inner Circle if you can make a real one work.

 

[KerimF]

In the 80's, I used it in my short-range RF voice links (around 3 km, air distance) between my house and workplace. And when a MW receiver was tuned to my channel (say around 1 MHz), the listener would hear sort of noisy interference while a voice signal was transmitted (or nothing during silent periods) since the frequency of the suppressed carrier was varying continuously over about 6 channels (60 KHz). At that time, its frequency was modulated by 6 Hz sinewave instead of 300 Hz (in the attached example). It was an easy way to scramble the transmission.

But I didn't like to keep scrambling my RF links using this noisy solution. After a few months, I decided to use it on FM band.
The frequency of the suppressed carrier was 32768 Hz instead of 38 KHz (see stereo FM protocol) which was modulated by the voice signal. I didn't need to add a pilot (a stereo FM transmitter has to also transmit a 19 KHz pilot to help the FM receiver recover the suppressed 38 KHz). Naturally, I let the baseband (R+L) of the FM channel be empty. While an FM receiver was tuned to my channel (say around 100 MHz), the listener would hear just a silent channel. I used this quiet scrambling method for many years till I got a 'bell' phone line at home (which was located in a rather poor neighborhood, far from the city center).

 

[poormystic]

I think you might do well to study the theory of modulation.
It's been a good while for me but I am sure if this - a "parallel" AM signal is inevitably generated in the process of frequency modulation. Filters are employed to clean such things up.
... BTW I'd better declare that I have not fully analysed and comprehended your circuit. I'm still surprised it works.

 

[KerimF]

I think you might do well to study the theory of modulation.
It's been a good while for me but I am sure if this - a "parallel" AM signal is inevitably generated in the process of frequency modulation. Filters are employed to clean such things up.
... BTW I'd better declare that I have not fully analysed and comprehended your circuit. I'm still surprised it works.

AF: Audio Frequency
MC1496 can generate a DSB-SC signal (say at Fc=1 MHz, modulated by AF_AM).
If the frequency of the carrier is first modulated by a sine wave AF_FM (say of 6 Hz or 300 Hz) so that the frequency deviation (delta F) is relatively wide (say +/- 30 KHz), we get an AM-FM signal whose carrier is suppressed.
The effective bandwidth of such a signal is 2*[BW of AF_AM + (FM delta F + BW of AF_FM)].
If:
BW of AF_AM = 4 KHz
BW of AF_FM = 300 Hz
FM delta F = 30 KHz

BW of AM-FM = 2*[4 + (0.3+30)] = 68.6 KHz

 

[KerimF]

... BTW I'd better declare that I have not fully analysed and comprehended your circuit. I'm still surprised it works.

Did you read the comments on the schematic?

The simple trick that lets PLL (whose VCO mid-frequency is set at 2*Fc) be locked, despite the polarity reversal of the carrier (Fc), is the frequency doubler (see U2 and U3, XOR gates). Everything else is well-known.

 

[KerimF]

May I add that, obviously, such a demodulator works also for any amplitude modulation index (AM DSB signal), from m=0 (no AF modulating signal) to m=infinity (no carrier). And, being of synchronous type, the transmitted audio bandwidth is limited by the allowed channel bandwidth only (in case of transmitting hi-fi music, for example).

Although I knew this very practical topology and used it in the 80's in real, it is still assumed being non-existent at all faculties of RF communications around the world.
And I will likely not live long enough to hear its given name when it will be taught at the universities besides the two well-known ones.

Kerim

 

[KerimF]

Hello again,

I like to add that I had the chance to talk about the demodulator above with some professors at faculties of RF communications.
Their common reply is that there is no reason to teach it in these days because, in general, DSB-SC is no more used.
But, at the same time, they don’t mind teaching the two less practical solutions of Costas Loop and the Squaring Method (for demodulating a DSB-SC signal).
Why they have to contradict themselves; God only knows... and some of you here perhaps.

Cheers,

Kerim

 

[Delta Prime]

But, at the same time, they don’t mind teaching the two less practical solutions of Costas Loop and the Squaring Method (for demodulating a DSB-SC signal).

As a 14 year-old middle school student in the United States, it was part of the curriculum .

 

[poormystic]

As a 14 year-old middle school student in the United States, it was part of the curriculum .

Wow
No SSB at our school... no modulation at all. (Sniff)
Quite buhulious, in hindsight.

 

[KerimF]

By the way, do you know a reference that shows how to recover two modulating signals of a DSB-SC one, in case the frequency of the suppressed carrier is also modulated, as on post #1? Thank you.
(Please note that I, unlike you, am not allowed to access any AI site).

 

[Delta Prime]

(Please note that I, unlike you, am not allowed to access any AI site).

Noted; young man.