frequency summer

[dr.power]

Hello guys,

Please somebody enlights me regarding this:
**broken link removed**

After looking at the above pic I thought why we need to use SSB modulation to generate a single sideband signal? (Supposing we need the USB) why not just SUM to frequencies by a frequency summer circuit if any? Then we do not need a very sharp filter to filter out one sideband (hard to be achieved).

Thanks

 

[Nigel Goodwin]

Perhaps you would care to explain what you mean by a 'frequency summer'?.

 

[dr.power]

Perhaps you would care to explain what you mean by a 'frequency summer'?.

Thanks for your input Nigel,

Well, for SSB (say using the USB) the upper sideband is Fc+Fm (which Fc is the carrier frequency and the Fm is the massage frequency). Now suppose you have a "'frequency summer'" to sum Fc and Fm, then you'll get the Fc+Fm is frequency speqtrm, Spectrum?
Is it important that in the time domain the carrier amplitude does, or does not while at frequency spectrum you'll get Fc+Fm?

 

[alec_t]

Fm is the massage frequency

Are you sure you're in the right forum?

Alec

 

[dr.power]

Are you sure you're in the right forum?

Alec

LOl,
Sorry I meant ""message"".

 

[Nigel Goodwin]

Thanks for your input Nigel,

Well, for SSB (say using the USB) the upper sideband is Fc+Fm (which Fc is the carrier frequency and the Fm is the massage frequency). Now suppose you have a "'frequency summer'" to sum Fc and Fm, then you'll get the Fc+Fm is frequency speqtrm, Spectrum?
Is it important that in the time domain the carrier amplitude does, or does not while at frequency spectrum you'll get Fc+Fm?

You seem to be making up imaginary terms and functions?.

There's no such thing as a 'frequency summer'.

A frequency mixer (which is what does exist) outputs the sum, the difference, and both original frequencies - a double balanced one reduces this to just sum and difference.

The most common SSB generator uses double balancing to get rid of the carriers, and a sharp filter to get rid of the unwanted sideband.

 

[crutschow]

If you are referring to a "summer" such as an op amp circuit, it sums voltages not frequencies (you just get one frequency riding on the other with no change in the frequency spectrum). To get a frequency summer you need a multiplier or mixer which does output the sum (and difference) of two frequencies, as Nigel stated.

A simple circuit to just give the sum of two frequencies and no other frequencies is not possible. If it were it would already have been invented.

 

[MrAl]

Hello,

I think what he means is a signal 'adder', so my interpretation would be to add the two signals with a summer. The result would be the higher frequency riding on the lower frequency, which is not the same as modulation so it would not get detected correctly in a circuit that was designed to detect and demodulate a usb signal.

Just to note that strictly speaking, true modulation does not produce cos(2*pi*Fc*t)+cos(2*pi*Fm*t) usb, it produces cos(2*pi*Fc*t+2*pi*Fm*t) usb. These are very different kinds of signals. Note that Fc+Fm are two frequencies (even for single frequency Fm) while the true modulation signal (the single term cosine one) is only one single frequency.

What would happen with added Fc+Fm (as interpreted above) is the Fc alone would propagate while the Fm would not propagate very well, so only Fc would be detected instead of the required cosine signal above. Since Fc got neither amplitude nor frequency modulated there would be no message signal to detect at the receiver. Since there would probably be a hint of non linearity in the transmitter when the signals are mixed, there may be a very faint signal detected but it would be hardly worth using as the signal to noise ratio would be way too high to be useful.

Also just to note, there are other ways (multi-stage) to do the correct modulation that require less sharp filtering. You could do a search for modulation methods i guess and see what you can find. Of course there is digital too.

 

[Mr RB]

Sorry I couldn't resist;
Q; "Frequency summer"?
A; "Once a year!"

Now in a lame attempt to say something a little more on-topic; I guess you could make a "frequency summer" using something like a PIC and measuring the 2 input frequencies then generating the sum frequency. Maybe you could do it in analogue by 2 freq->voltage converters and summing the 2 control voltages into a VCO.

Whether or not it is actually useful is another matter...

 

[MrAl]

Sorry I couldn't resist;
Q; "Frequency summer"?
A; "Once a year!"

Now in a lame attempt to say something a little more on-topic; I guess you could make a "frequency summer" using something like a PIC and measuring the 2 input frequencies then generating the sum frequency. Maybe you could do it in analogue by 2 freq->voltage converters and summing the 2 control voltages into a VCO.

Whether or not it is actually useful is another matter...

Hi MrRB,

Yeah

If you think about what you just said in the context of this thread, what you are suggesting is a single frequency to voltage converter where we input the message signal and add a DC offset. That would produce a voltage that increases based on the frequency of the message signal. No need for a second frequency. The voltage then converted back to a frequency and that frequency would be some base frequency that increases by an amount equal to the message frequency.
We could analyze this to see if it is the same as USB but i dont think it would be either.

Usually one of the goals is to make the circuit as simple as possible too. If we werent worried about that then i would think digital would be the way to go.

 

[dr.power]

hanks for all your inputs.

Ok now I enplane my mean.
As you know in Am modulation 2 frequencies are multiplied to each other to create the upper along with the lower sidebands (DSB). Both the upper and of course the lower sidebands are created because they have been "multiplied" to each other. When I saw the link in my first post, I reach to this conclusion that for SSB we just need the summation of 2 frequencies (we had to multiple 2 frequencies to get DSB then filter one side band to get SSB), So I though if there is any CIRCUIT as frequency summer (MrRB, lol) so that sums the audio with the higher carrier so that create a DIRECT and straightforward SSB?
How it makes sens now?!

PS, Yes I know that an op-amp summing circuit just adds 2 amplitudes to each other, but my question was is there any direct frequency summing circuit ?

 

[MrAl]

Hi again,

Yes that makes sense and that's a reasonable question. You're asking if we are given the required signal, can we find a simpler way to generate it.

What i was trying to show was that with somewhat simple circuits there are various ways to get the desired signal, but some techniques just wont work simply because they wont produce the right signal. That's usually assuming we dont have control over the receiver, but sometimes we do so we might find other techniques if that end is open to us too, but there are other considerations apart from just generating a signal that we can transmit...namely, the power bandwidth where we want to use as much power as we can without wasting any.

So assuming we dont have any control over the receiver design, we have to generate a signal cos(w1*t+w2*t), and just to note that is also not the same as cos(avg(w1*t)+avg(w2*t)) which would result from averaging the two frequencies over time before creating the signal.
The most common way to do this is to multiply the two signals (not the frequencies per se') and filter out the lower side band. Is there any other way to do this?
The usual method comes from multiplying the two signals and then noting that we get both upper and lower sidebands when we use trigonometric identities to expand or reduce the result. We get one result we want and one we dont want, so we filter out the one we dont want.
But what if we work backwards?

If we start from the signal we want:
sig=cos(w1*t+w2*t)
and use trig identities to expand or reduce we find we can get an equivalent:
sig=cos(t*w1)*cos(t*w2)-sin(t*w1)*sin(t*w2)

And what exactly does this expression tell us? Well, the right term is the original modulation where the two signals are multiplied, but the left term is a little different...it's the two signals shifted by 90 degrees each and THEN multiplied, and of course the subtraction is just the analog subtraction of the two signals.
This basically means that if we shift the phase of the two signals and use these new signals in a modulator, then subtract the modulation of the two signals without phase shift we end up with the upper side band, which is the signal we were after, and we didnt use any filters.

How well would this work in practice? We'd first have to design a circuit to shift the phase of the message signal by a constant 90 degrees.
We'd then have to look at the effects of errors in the modulation or phase shifting circuit and see how they affect the result.

In any case, that's probably the only way get what you are after without using a digital technique which would probably be cheaper these days.

ADDED LATER:
I did a quick search on modulation techniques and found that the technique described above is known and has been used for quite some time already, since the days of vacuum tubes. It's known as "Hartley Modulation" and also has it's advantages and disadvantages.
From this we might note that there is no really simple way to do this except for a purely digital technique when that technology is allowed.