Why Do We Use Mixers? - Explained

[CraigH]

I understand that mixers are used to convert a signal from one frequency to another, but why would we want to do this?
If the signal needs to be transmitted at a certain frequency then why not just use amplitude modulation or frequency modulation to put that signal onto a carrier signal of that certain frequency, isn't that how it's normally done?
Thanks

 

[f95toli]

But mixers are neccesary to build circuis that can do AM or FM modulation...

 

[CraigH]

I thought operational amplifiers were used for AM and FM modulation?
Are mixers just another way of achieving modulation?

 

[Studiot]

Mixers do exactly what it says on the tin.

They combine two or more (analog) signals in some way.

Modulation is a form of signal combination, but there are non analog forms of modulation eg pulse width modulation or pulse height modulation to name but a few.

Mixers are a circuit configuration used to achieve analog modulation. They may incorporate amplifiers, operational or otherwise.

(Operational) amplifiers therefore form more fundamental building block, just as transistors form an even more fundamental one.

Mostly with amplifiers we require linearity. That is the output is a constant times the input, with the constant the same at all frequencies.

Modulation is a non linear process so we require additional circuitry to configure an amplifer as a modulator.

One simple form of modulation occurs when we perform multiplication of one signal by another.
This is the kind of modulation you have already asked about in another thread.

 

[vk6kro]

I understand that mixers are used to convert a signal from one frequency to another, but why would we want to do this?
If the signal needs to be transmitted at a certain frequency then why not just use amplitude modulation or frequency modulation to put that signal onto a carrier signal of that certain frequency, isn't that how it's normally done?
Thanks

Amplitude modulation can be achieved in the way you describe, although large transmitters would usually be modulated at a lower level and this could be amplified.

Frequency modulation can be produced by varying the frequency of an oscillator. This frequency modulated signal can be passed through frequency multipliers that increase the frequency (and the amount of frequency modulation) with each multiplication.

While this was how it was done in the past, it had some serious disadvantages.
Oscillator drift tended to get magnified, so signals became more unstable, and you needed to generate each frequency individually. This usually meant buying a crystal for each frequency that was to be used.

What is commonly done now, is to generate a signal from a phase locked loop, frequency modulate it by varying a control voltage and then use a mixer to convert this to the required output frequency.
This way, you could generate many possible output frequencies, (but generate them at a low frequency where sophisticated integrated circuits can be used), just by controlling a programmable frequency divider with a switch.
Drift is not magnified because a mixer is used and not frequency multipliers.

It is also possible to generate AM signals this way, although this would not be done much, if at all, now.

More importantly, single sideband signals (ie an AM signal with the carrier and one sideband removed) can be generated with a crystal filter on one frequency, (say 9 MHz) but converted to many possible frequencies by the use of a mixer and a local oscillator generated by phase locked loop techniques.

 

[rbj]

I thought operational amplifiers were used for AM and FM modulation?

in AM, you're mutliplying two voltage signals; $x_m(t) x_c(t)$ simple op-amp circuits can multiply a voltage signal by a constant. what multiplies one time-variant signal with another one?

Are mixers just another way of achieving modulation?

yeah. they do it by adding the two signals together, then applying that summed signal to a non-linearity. this non-linearity at least will have a 2nd-order term

$$(x_m(t) + x_c(t))^2$$

blast that out and you can see a term with the product of the two signals.

 

[Averagesupernova]

What is commonly done now, is to generate a signal from a phase locked loop, frequency modulate it by varying a control voltage and then use a mixer to convert this to the required output frequency.
This way, you could generate many possible output frequencies, (but generate them at a low frequency where sophisticated integrated circuits can be used), just by controlling a programmable frequency divider with a switch.
Drift is not magnified because a mixer is used and not frequency multipliers.

Actually, drift is magnified in this way as well. In effect, the master oscillator that provides a reference is multiplied up. Not by direct frequency doublers/triplers, but since the VCO is compared to the reference by the frequency divider(s) the end result is the same. Mixers are not necessarily used especially when using dual modulus prescalers. That is not to say that mixers are never used in frequency synthesis.

 

[vk6kro]

The reference oscillator and the vco are both divided down before comparison. So the effect depends on the start frequency for each of these oscillators. However it is not comparable with the old style FM multiplier chains which could multiply all the way from 8 MHz to 500 MHz.
These transmitters had great difficulty with out-of-band emissions.

Commercial VHF transmitters I have seen invariably use mixers to convert a range of low frequency channels to VHF.
Doing this makes it easier to produce a transceiver using much of the same circuitry that was used in the transmitter for the receiver.

 

[Averagesupernova]

Yes, the VCO and reference are typically divided before the phase detector. BUT, the reference is typically a lot lower to start with so my point still stands. In the past, (NTSC) a TV tuner typically used a dual modulous prescaler and not a mixer before the main frequency divider. If the reference oscillator was 6.4 MHz and we wanted to run the VCO at 640 MHz then the error on the reference would be X100 on the VCO. It doesn't matter how much the reference is divided down on the way to the phase detector since the VCO will be divided down to the same frequency. However, if a mixer is involved, you are correct that the multiplication effect will be eliminated, but since not all PLLs use a mixer scheme, it is not safe to generalize by saying PLL eliminates a multiplication error. Also remember that crystal oscillators have a stability in PPM not absolute hertz so raising the frequency of the reference will also raise the +/- spec.

 

[vk6kro]

The mixer is not part of the Phase Locked Loop.

I'll try again.

A Frequency Modulated multi-channel signal is generated at a low frequency, possibly by a PLL.

This FM signal, which can be switched in frequency at a low frequency, is then converted in a mixer to the final output frequency and amplified before being transmitted. This is done by mixing this signal and a stable oscillator signal, so that the sum of the two frequencies is equal to the transmit frequency of the transmitter.

If a PLL is used to generate the signal, it is because clever chips that can do this are mainly available for use at low frequencies. At these frequencies, very stable crystal oscillators can be used as a reference for generating the signals.

The poster's question was not about PLLs. He asked why we use mixers in transmitters rather than apply modulation directly to the final amplifier of the transmitter.
In the case of FM, this is not reasonably possible and FM is derived either from a mixer or applied to the early stages of a frequency multiplier chain that drives the transmitting amplifier.
Using a mixer has some outstanding advantages, as I have mentioned.

 

[Averagesupernova]

The mixer is not part of the Phase Locked Loop.

In some cases it is. Quit confining yourself to commercial two-way and 2 meter transceivers and you will see what I mean. I have seen a lot of PLL schemes in my day and still continue to see things I have not seen before.

 

[sophiecentaur]

I understand that mixers are used to convert a signal from one frequency to another, but why would we want to do this?
If the signal needs to be transmitted at a certain frequency then why not just use amplitude modulation or frequency modulation to put that signal onto a carrier signal of that certain frequency, isn't that how it's normally done?
Thanks

Imagine you wanted to make a comms FM transmitter to operate at a number of frequencies. Would it be easier to make a special oscillator / modulator at each frequency you needed or to produce one high quality modulator at a convenient Intermediate Frequency (i.f.) and mix this up to anyone of the required operating frequencies? Using an i.f., followed by a mixer, is particularly useful when you need to produce a well behaved, 'tight' channel filter (either for receiving or transmitting). You only need to do this once for all operating frequencies if you can use a relatively simple up converter or down converter (mixer).

Look up "Superhetrodyne". Lovely old fashioned term but it describes how mixing is used in almost every receiver you will find. (I am sure people will give a list of exceptions, such as TRF but they're not really relevant).