DSB and SSB Radio: Explaining the Confusion

[Weightofananvil]

Hey,
I'm learning about DSB and SSB in my communications classes and I'm a little confused with the concept.

So, We know that in a AM signal there is the Modulating signal (Information) and the Carrier signal (Transmission signal.) We also knows LSB and USB are created when the signals are combined

in DSB signals it is said that the carrier is suppressed and only 1 side band remains (SSB) It also says that at the "min" peak in the cycles there is a phase shift.

This confuses me a little bit because in my textbook they are only speaking of the transmission section of the am signals. Why would you want to eliminate the carrier if the signal hasn't been sent? It also says one of the main fall backs of SSB and DSB is that on the receiving end it is hard to remove the carrier signal. This all seems counter-intuitive of each other...

How I picture AM in my head is that the envelope of the carrier frequency is reshaped by the modulating signal. Due to my backround in audio I tend to think of the differences in sidebands, harmonics or anything of that nature in terms of the sound frequency spectrum and then just realize where it actually is on the spectrum. I understand that if the signal is single, you only need 1 side band which saves power but I don't understand what exactly is happening with the carrier.

Typically in textbooks they always show things with sin waves, but if USB and LSB are different frequencies then how is the modulated signal shown with just a single frequency? (aside from showing the signals before they are joined)

Hopefully these aren't all super silly,
but I've been waiting for it to click together.
Thanks in advance!

 

[davenn]

So, We know that in a AM signal there is the Modulating signal (Information) and the Carrier signal (Transmission signal.) We also knows LSB and USB are created when the signals are combined

no that is not quite correct

in DSB signals it is said that the carrier is suppressed and only 1 side band remains (SSB) It also says that at the "min" peak in the cycles there is a phase shift.

just to clarify that for you. You would generally have DSB with suppressed carried
(DSB with carrier is basically an AM signal)

SSB as it suggests is just one of the sidebands upper or lower with a suppressed carrier

This confuses me a little bit because in my textbook they are only speaking of the transmission section of the am signals. Why would you want to eliminate the carrier if the signal hasn't been sent?

the carrier isn't eliminated but rather suppressed it allows for a much narrower bandwidth of transmission, which is a much more efficient use of the available spectrum in a particular frequency band. It's also much more efficient use of the transmitter as it doesn't have to consume power producing a continuous carrier as AM does.
EDIT: Ohhh and I forgot to mention, SSB offers a much better SNR ( Signal to Noise Ratio) than AM

have a read of this ...
https://en.wikipedia.org/wiki/Single-sideband_modulation

and this
https://en.wikipedia.org/wiki/Double-sideband_suppressed-carrier_transmission

and this
http://www.radio-electronics.com/in.../single-sideband-suppressed-carrier-ssbsc.php

and finally this
http://electronics.stackexchange.com/questions/79939/am-and-suppressed-carrier-ssbthere's a lot of serious reading for you Dave

 

[TurtleMeister]

Hey,
So, We know that in a AM signal there is the Modulating signal (Information) and the Carrier signal (Transmission signal.) We also knows LSB and USB are created when the signals are combined

no that is not quite correct

What is not correct about it?

 

[Averagesupernova]

What is not correct about it?

The process of amplitude modulation is more than a combination of signals. The carrier is multiplied with the modulating signal. This gives us 2 new frequencies plus the originals. The 2 new signals are the upper and lower sidebands.

 

[TurtleMeister]

The process of amplitude modulation is more than a combination of signals. The carrier is multiplied with the modulating signal. This gives us 2 new frequencies plus the originals. The 2 new signals are the upper and lower sidebands.

Yes, I agree. But how does it make the op's statement wrong? He didn't use the word multiply? Or, maybe he used the word combined instead of multiply? Personally, I probably would have used the word mixed. But anyway, I think the op probably knows what's happening in this particular instance. Just a poor choice of words maybe.

 

[TurtleMeister]

How I picture AM in my head is that the envelope of the carrier frequency is reshaped by the modulating signal.

Well, that's the way it would appear on an oscilloscope. Try viewing it with a spectrum analyzer. You can probably find some examples with Google image search. Try "amplitude modulation spectrum".

 

[Averagesupernova]

Yes, I agree. But how does it make the op's statement wrong? He didn't use the word multiply? Or, maybe he used the word combined instead of multiply? Personally, I probably would have used the word mixed. But anyway, I think the op probably knows what's happening in this particular instance. Just a poor choice of words maybe.

This is a case where there is a lot of confusion due to audio mixing which is summing. Frequency mixing is a completely different animal which of course you know. I just choose to avoid the confusion.

 

[Weightofananvil]

Thanks both of you, I probably would have assumed it was similar to audio mixing. I'm going to read those articles, if I'm still foggy on stuff Ill report back.

Since each sideband is a different signal does that mean in Audio transmission (say radio) which is a stereo signal L and R would each be in a sideband?

 

[davenn]

Since each sideband is a different signal does that mean in Audio transmission (say radio) which is a stereo signal L and R would each be in a sideband?

No, each sideband has the same signal info within it

and this is the exact reason you can get rid of the carrier and one sideband ...
because you still have all the info ( voice signal) in the one remaining sideband

 

[AlexCaledin]

...Since each sideband is a different signal does that mean in Audio transmission (say radio) which is a stereo signal L and R would each be in a sideband?

Yes, there was such a system of AM stereo!

https://en.wikipedia.org/wiki/AM_stereo#Kahn-Hazeltine

 

[Weightofananvil]

No, each sideband has the same signal info within it

and this is the exact reason you can get rid of the carrier and one sideband ...
because you still have all the info ( voice signal) in the one remaining sideband

Yes, That makes sense.

Yes, there was such a system of AM stereo!

https://en.wikipedia.org/wiki/AM_stereo#Kahn-Hazeltine

But it also makes sense that it could be done.
I imagine the modulating circuit is combined with the carrier separately L and then R and then later combined. There would still be harmonics I imagine. Probably not the best thing in the world.

 

[davenn]

Yes, there was such a system of AM stereo!

https://en.wikipedia.org/wiki/AM_stereo#Kahn-Hazeltine

Kahn-Hazeltine[edit]
The Kahn-Hazeltine system also called ISB was developed by American engineer Leonard R. Kahn and the Hazeltine Corporation. This system used an entirely different principle—using independently modulated upper and lower sidebands.

hadn't heard of that one before

 

[jim hardy]

Hmmmmistaken post , hereby erased...

as i said, corrections welcome =

thanks Alex !

old jim

 

[jim hardy]

Okay, thanks Alex ! post deleted ...

 

[Weightofananvil]

Thanks a lot Jim!

So when you have a modulation signal and you are designing the circuitry for modulation and demodulation is there principles/rules in which you choose a proper carrier frequency in relation to the modulation frequency (in many cases a modulation frequency range?)

Thanks for the all the replies. This is helping a great amount.

 

[davenn]

So when you have a modulation signal and you are designing the circuitry for modulation and demodulation is there principles/rules in which you choose a proper carrier frequency in relation to the modulation frequency

The two are not related ... eg the modulation freq is commonly the voice audio range, 300Hz to ~ 3 kHz
The carrier freq can be kHz to 100's of GHz and anything in between

one of my amateur radio transceivers has TX capabilities on a bunch of bands between 1.8 and 30 MHz, then 50 - 54 MHz, then 144 - 149MHz, 430 - 440MHz
and finally 1250 - 1300MHz all in one box. It can do AM, FM, SSB and digital modes across all those bands.Dave

 

[jim hardy]

What Dave said...

 

[Averagesupernova]

The two are not related ... eg the modulation freq is commonly the voice audio range, 300Hz to ~ 3 kHz
The carrier freq can be kHz to 100's of GHz and anything in between

one of my amateur radio transceivers has TX capabilities on a bunch of bands between 1.8 and 30 MHz, then 50 - 54 MHz, then 144 - 149MHz, 430 - 440MHz
and finally 1250 - 1300MHz all in one box. It can do AM, FM, SSB and digital modes across all those bands.Dave

Not me. I have old(er) school stuff.

 

[davenn]

The Kenwood TX2000X is a really nice bit of kit
It replaced 4 separate radios in the shack

 

[Merlin3189]

My maths is probably worse than Jim's, but: it starts ok, but if you do go further with the "mishmash" it just reverts to sin(ω_c)sin(ω_a) plain and simple.

I agree that in practice you may get all sorts of other bits and pieces, IMO because the multipliers are not perfect and the original carrier may not be pure. So you do filtering as you say. But the maths says you could get a pair of perfect sidebands.

I agree totally with the voila, 2 sidebands. But this is DSB. Notice, there is no carrier present, just the two sidebands. No need to suppress nor eliminate a carrier. It simply is not there, IF you can accurately multiply the carrier by the modulation.
For SSB there is also a similar expression which generates a single sideband, without carrier nor other sideband. It too relies on accurate multiplication, but also on being able to generate quadrature signals for both carrier and modulation, which was very difficult before DSP.
(In case anyone is interested, LSB = cos(ω_c)cos(ω_a) + sin(ω_c)sin(ω_a)
and USB = cos(ω_c)cos(ω_a) - sin(ω_c)sin(ω_a) )

The expression for AM is (carrier) x (1 + modulation) or sin(ω_c )(1+sin(ω_a))
This is easy for old guys to remember, if you think of the old amplitude modulation class A circuit. It was done by modulating the HT line to the PA. Since this can't go negative, you add the modulation to the standing HT, so that it swings (up to) from near zero to near double.
With no modulation, you get full carrier;
With modulation amplitude equal to the HT, you get 100% modulation with full carrier and two sidebands, carrying between them power equal to that of the carrier;
With greater modulation the valve cuts off during modulation peaks and you get splatter (mishmash?) and become very unpopular;
With modulation amplitude less than the HT, you get undermodulation, with full carrier and two sidebands carrying less power than the carrier. This is what is happening most of the time, because only peaks can be allowed to reach 100% modulation

A bit long winded, but OP seems to know about this signal. What you notice here is that in every case you are putting at least half your power into the carrier and this carrier is exactly the same whatever the modulation, if any. So it tells the receiver nothing by itself. The information about the modulation is all in the sidebands. But, as OP says, you can easily see the modulation in the envelope of the combined signal. And it is ridiculously easy to get the modulation back, just using a diode (or crystal and cats whisker.)

So one motivation for DSB is to not waste power on transmitting carrier. If you can transmit 100W, instead of sending 50W of carrier and 50W of sidebands, send just the sidebands at the full 100W, then let the receiver generate a carrier signal and add it to the sidebands it receives. That will regenerate that simple AM signal in the receiver, but at twice the power.
Doing this (in the receiver) was not trivial, which is one reason why DSB was not that common. You can work out what the carrier frequency is, because it is midway between the sidebands. And you can manually adjust your local oscillator to sit at the right frequency. But it is much more critical than tuning a simple AM signal.
Since the two sidebands are like mirror images of each other, which both contain exactly the same information just arranged in a different way (if the whiz kids will allow that fuzzy desciption), you only need one of them, hence Single Side Band.
One is sin(ω_c +ω_m) the other sin(ω_c -ω_m)
Again they don't look nor sound like the original modulation, but if you reinject the carrier sinewave at the receiver, then you get the modulation back.
Eg. Take the USB sin(ω_c +ω_m) and multiply by sin(ω_c)
you get 1/2[cos(ω_m) - cos(2ω_c + ω_m) ] which looks messy, until you realize the left cos is the modulation and the right cos is RF at nearly twice the carrier frequency, so can be filtered out to leave the modulation.
Again you need to know what the carrier frequency is and it is a bit more difficult than with DSB. If the receiver knows what frequency is used at the transmitter it has to set it's oscillator to that frequency (and keep it there, used to be the difficult part!) One (common) technique was simply to have a matching crystal in both transmitter and receiver. Even without knowing the exact frequency, as with DSB, a skilled listener can tune the oscillator until the signal becomes intelligible when you get close enough to the original carrier.

It also says one of the main fall backs of SSB and DSB is that on the receiving end it is hard to remove the carrier signal.

I think you mean "drawbacks" and "on the receiving end it is hard to reinstate the carrier".
Yes. As said above, you need the carrier to make sense of an SSB or DSB signal. Recreating that was the problem. Crystal controlled oscillators was one solution. Phase locked loops was another. Reduced carrier, rather than suppressed carrier was a halfway house.
Removing the carrier at the transmitter was also a problem, mainly met by crystal filters and balanced mixers until DSP came along.

Typically in textbooks they always show things with sin waves, but if USB and LSB are different frequencies then how is the modulated signal shown with just a single frequency? (aside from showing the signals before they are joined)

The modulated signal is generally not a single frequency (could be for SSB), and I don't think anyone would show a modulated signal as a simple sinewave.

First using sinewaves in books, makes the maths easy to follow. They can write the formulae for any modulating signal - it just gets messier.
Secondly, about 200yrs ago Fourier showed that a more complex signal is mathematically identical to a collection of sinewaves. You can work out what happens to any signal waveform, by working out what happens to each of those sinewaves and adding the results together.

USB & LSB, are bands of frequencies.
A single sinewave modulation produces an upper side frequency and a lower side frequency. Eg. 300Hz audio modulated onto a 1MHz carrier, gives 1.0003 MHz and 0.9997 MHz, spot frequencies. If you know that the carrier was 1MHz and you get either of these frequencies, you can work out what the audio frequency was. Try this: if you receive 1.0005 MHz, what was the audio frequency?
When a radio station broadcasts a pop record, they want to send a complicated waveform which is a mixture of frequencies (at varying levels). So for each frequency you get an upper and lower side frequency. Together these make the two bands of frequencies, called sidebands. Converting these side bands of frequencies back to their original sinewaves and adding them together gives us the original complicated waveform.
In the eg. above, if you had a sound containing a 300Hz tone and a 500Hz tone of twice the amplitude, it wouldn't look like a simple sinewave, but you'd get 1.0003MHz with one amplitude and 1.0005MHz with twice the amplitude. (and for DSB or AM, 0.9997 and 0.9995MHz) This modulated signal would not look like (nor be) a simple sinewave. When you added back the carrier frequency and recovered the 300Hz and 500Hz , you would get twice as much 500Hz as 300. When you added them together, they would again make the same waveform as the original mixture.

 

[Weightofananvil]

Whoa, Thanks tons Merlin.
Now give me a minute to swallow all the information I've gotten and ill be back.
(Probably with more questions)
:-)

 

[jim hardy]

My maths is probably worse than Jim's, but: it starts ok, but if you do go further with the "mishmash" it just reverts to sin(ωc)sin(ωa) plain and simple.

Thanks Merlin

you're right of course.

I am trying to go back fifty years to figure out where i picked up the products mis-information.
Somehow it got reinforced about fifteen years ago in a Fourier treatment of modulation... must've been something else,.

As Mark Twain said -
"The older i get
the more things i can remember
the more i can imagine
and the less clear the difference."

sorry for the red herring.

 

[TurtleMeister]

@Merlin3189
I think at 100% modulation the power in the side bands is 50% the power in the carrier (single tone transmission). In several places in your post #20 you seem to indicate otherwise. Or maybe I'm miss reading it.

 

[Averagesupernova]

Both sidebands together will be the same power as the carrier at 100% modulation.

 

[TurtleMeister]

http://www.radio-electronics.com/in.../am-amplitude-modulation/theory-equations.php

It can be seen that for a case where there is 100% modulation, i.e. M = 1, and where the carrier is not suppressed, i.e. A = 1, then the sidebands have half the value of the carrier, i.e. a quarter of the power each.

 

[Merlin3189]

I agree with the last 2 posts.
It's something that often trips me up, forgetting to square the amplitude of waves when talking about power, but I said, "100% modulation with full carrier and two sidebands, carrying between them power equal to that of the carrier;" (Carrier = 1/2, one SB = 1/2 carrier = 1/4, Two SBs =1/4 + 1/4 = 1/2 = carrier)

 

[Averagesupernova]

Here is how I remember it: Watch the power meter with no modulation, compare to 100% modulation. Power reading will about double.

 

[TurtleMeister]

Here's another excerpt from the same website I referenced previously:
http://www.radio-electronics.com/info/rf-technology-design/am-amplitude-modulation/efficiency.php

When the carrier is fully modulated i.e. 100% the amplitude of the modulation is equal to half that of the main carrier, i.e. the sum of the powers of the sidebands is equal to half that of the carrier. This means that each sideband is just a quarter of the total power. In other words for a transmitter with a 100 watt carrier, the total sideband power would be 50 watts and each individual sideband would be 25 watts. During the modulation process the carrier power remains constant. It is only needed as a reference during the demodulation process. This means that the sideband power is the useful section of the signal, and this corresponds to (50 / 150) x 100%, or only 33% of the total power transmitted.

I did a little homebrewing back in the 60's and 70's and one of the things I remember is that when selecting the modulation tubes and transformer, they need to be rated at least to 1/2 the DC power input to the final PA for 100% modulation.

 

[Averagesupernova]

That looks contradictory to me. I didn't follow your link, but from what you pasted in: This means that each sideband is just a quarter of the total power. In other words for a transmitter with a 100 watt carrier, the total sideband power would be 50 watts and each individual sideband would be 25 watts.
-
100 watt carrier would put each sideband at more than 25 watts if there is a quarter of the total power in each sideband.

 

[TurtleMeister]

That looks contradictory to me. I didn't follow your link, but from what you pasted in: This means that each sideband is just a quarter of the total power. In other words for a transmitter with a 100 watt carrier, the total sideband power would be 50 watts and each individual sideband would be 25 watts.
-
100 watt carrier would put each sideband at more than 25 watts if there is a quarter of the total power in each sideband.

You seem to have a misconception. The modulator power is added to the carrier power. As stated in the quote, if the carrier power is 100 watts and you modulate it with a single tone at 100%, the total output power will be 150 watts. 100 watts in the carrier and 25 watts in each sideband.

 

[Averagesupernova]

So how is 25 watts a quarter of 150 watts? THAT is my point.

 

[TurtleMeister]

So how is 25 watts a quarter of 150 watts? THAT is my point.

It's not a quarter of the total power, it's a quarter of the carrier power.

 

[Averagesupernova]

Maybe so, I am not above a mistake in that area. Where I AM above a mistake is this quote: This means that each sideband is just a quarter of the total power. In other words for a transmitter with a 100 watt carrier, the total sideband power would be 50 watts and each individual sideband would be 25 watts.
TOTAL power and CARRIER power are not the same thing. For the record I would not believe much off a site that gets that wrong.

 

[TurtleMeister]

Maybe so, I am not above a mistake in that area. Where I AM above a mistake is this quote: This means that each sideband is just a quarter of the total power. In other words for a transmitter with a 100 watt carrier, the total sideband power would be 50 watts and each individual sideband would be 25 watts.
TOTAL power and CARRIER power are not the same thing. For the record I would not believe much off a site that gets that wrong.

Wow, that is an error. Not mine, but from the web site. I didn't even notice that. It should read "This means that each sideband is just a quarter of the carrier power. In other words for a transmitter with a 100 watt carrier, the total sideband power would be 50 watts and each individual sideband would be 25 watts." Apologies, I see why you were confused.

 

[Averagesupernova]

I can't take the time to do the math right now but I was REALLY CERTAIN that the total power doubles from 0% to 100%. The link does not support that no matter how you look at it. I will check in here later.