Difference between filter and amplifier

[simplsoft]

Hello

Can anyone tell me what is the difference between filter and amplifier.

If an active filter is designed using OP-AMP or OTA then we also have gain of active element. It can amplify as well as filter out irrelevant frequency component from the signal. Similar is in an amplifier if amplifier is designed then also has gain and it amplifies signal. It can also work upto a certain frequency limit lets say if the amplifier amplifies signal upto 100KHz then it can also be like a low pass filter that allows frequency to pass upto 100KHz.

So then whats the difference?

Thanx

 

[ronaldomponte]

What is the difference between water and ice?
We use water to quench the thirst and the ice to refresh a drink.
But if we eat ice we can also quench the thirst and drinking water on a hot day is also refreshing. <))
You define how you want to use both, the same with amplifiers and active filters.
When your application demands more on amplification rather than filtering, you design a "pure amplifier" paying more attention to parameters for meeting the GBW requirements.
Remember that in this case you only have the degree of freedom to create a low pass filter since at some point your amplifier meets cut-off frequency.
When your application demands a more accurate filtering of your signal, you employ your knowledge on active filters to design the most suitable transfer function for you:
Chebyshev, Butterworth, Linkwitz–Riley, Paynter, Bessel, Elliptic filter and so on.
But, essentially, everything can be seen as the same but with slightly different purposes: As the water is H20 regardless of its physical state of matter, amplifiers and active filters are comprised of active and passive components performing pieces of each task according to what you want to accomplish.

 

[c_mitra]

A filter processes input in the frequency domain and an amplifier processes the input in the time domain.

The corresponding transfer functions are different but related via a Fourier transformation.

You need to consider the ideal cases first and then apply the limitations.

 

[crutschow]

They are two separate functions (but both can occur in the same circuit).

An amplifier increases the level of a signal to give a higher voltage or power out then the input.
It can also be used to buffer a high impedance signal to give a low impedance output.

A filter limits the frequencies that can pass.
Typical types are low-pass which attenuates signals above a design frequency, high-pass which attenuates signals below a design frequency, and band-pass which attenuate signals both above and below a design frequency (can be looked as a combination of a low-pass in series with a high-pass filter).

 

[simplsoft]

Yes but still I think amplifier and low pass filter is same. If you use an amplifier having a frequency band upto 10KHz similary if you design a low pass filter for a cut-off frequency of 10KHz then both are same thing. Then it means filter and amplifier is same

- - - Updated - - -

thanx for your reply.

But still if any application requires a designing of low pass filter for let suppose 10KHz then I can use filter with cut-off of 10KHz. Similarly an amplifier can also have a frequency band upto 10KHz it means both are same.

Although amplifier has a gain. But if I design an amplifier use it in active filter then, the amplifier that is used in filter also has gain.

Regards

 

[ronaldomponte]

Yes but still I think amplifier and low pass filter is same. If you use an amplifier having a frequency band upto 10KHz similary if you design a low pass filter for a cut-off frequency of 10KHz then both are same thing. Then it means filter and amplifier is same
- - - Updated - - -

Yes, I understand your point. In essence, ice and water can be seen as the same, if you want, and I agree with you.
But the point the others are trying to make (which I think is also very important to consider) is basically in what you want to do.
For instance, consider you have a signal with enough amplitude coming from a particular source. So your equipment is capable of detecting this signal without any amplification process. You ideally can simply plug and play.
But for some reason, your signal has interference with another source at a frequency above 10 kHz so you need to do something to suppress this artifact. OK, good enough.
In this case, you can design a low pass filter focusing only on the cut-off frequency you need to suppress the artifact.
You don't care too much if now you have a gain of 1 or 10 because your signal is already strong enough to be detected by your measuring system.
So, if you want to save power you give 0 dB, otherwise you don't mind.
Now, consider the same thing, but your signal is very weak.
In this case, you lose a degree of freedom, and you need to meet gain and cut-off frequency.
Thus, before filtering the signal, you may want to amplify it first so you ensure that your active filter will attenuate the undesirable signals more accurately.
You may think: Wait a minute. But what impedes me to do both things in the active filter?
Well, sometimes your application allows it, sometimes it doesn't, so you need to adequate your design accordingly.
Mathematically speaking, they are different functions and serves for different purposes.
But, of course, since in an active filter you use an amplifier, you may argue that there might be some overlap in the functionalities - as in the water/ ice example.

 

[c_mitra]

Yes but still I think amplifier and low pass filter is same. If you use an amplifier having a frequency band upto 10KHz similary if you design a low pass filter for a cut-off frequency of 10KHz then both are same thing. Then it means filter and amplifier is same...s

But they are still not the same; just because one unit performs both the function simply means these two functions are put in one package.

An ideal amplifier has no frequency limit. An ideal filter has no amplitude limit.

All real life amplifiers have some filter action and all real life filters have some amplification (attenuation in general) function.

But try to think in terms of ideal devices. You can grasp the principles better and far more easily.

 

[ronaldomponte]

Yes, @c_mitra now wrapped up nicely.

 

[LvW]

A filter processes input in the frequency domain and an amplifier processes the input in the time domain.

The corresponding transfer functions are different but related via a Fourier transformation.

You need to consider the ideal cases first and then apply the limitations.

To me, this sounds rather confusing.
Of course, both units (amplifier and filter) are processing signals in the time domain (that is the only domain which really exists!).
And - if we want and/or if it makes sense (!) - we can describe the properties of both devices also in the frequency domain, using appropriate mathematical tools.

 

[c_mitra]

To me, this sounds rather confusing.
Of course, both units (amplifier and filter) are processing signals in the time domain (that is the only domain which really exists!).
And - if we want and/or if it makes sense (!) - we can describe the properties of both devices also in the frequency domain, using appropriate mathematical tools.

Confusing? Sure!

Consider a resistor; it responds to the voltage applied and the defining function (Ohms law) does not have time in it (it is not frequency sensitive).
Consider a capacitor; it responds to dV/dt and the definition (c.dV/dt=i OR c.dV=i.dt) depends on the time.
Similar considerations applies to an inductor.
You can make a voltage divider using resistors and a simple filter using a capacitor.
To solve complex filters (or amplifiers) equations, you often use Fourier or Laplace transformations. They are as real as the time domain study.
Some problems look simple in time domain and some in frequency domain. Frequency and time are conjugate variables (you will see omega.t appear in places where you do not expect it).

 

[FvM]

A filter processes input in the frequency domain and an amplifier processes the input in the time domain.

The distinction is literally wrong and leads to nothing.

As said, a signal or the characteristic of a device can be characterized both in time and frequency domain. You may prefer frequency domain specification for some filters and time domain for some amplifiers, but the opposite can be appropriate as well, depending on the application range.

I also appreciate LvW's comment that frequency domain is "not real", in other words an abstraction that ignores some parameters.

 

[KlausST]

Hi,

A resistor operats in time domain.
A capacitor, inductor, analog filter operates in time domain.
A digtial filter like FIR, IIR, biquad .. operates in time domain.

BUT you may doing signal processing in frequency domain:
* convert an analog signal into digital with an ADC (time domain)
* then perform an FFT (transforming informations from time domain into frequency domain)
* then modify the FFT results (here is where the data procassing is done in frequency domain)
* then perform an inverse FFT (transforming data back to time domain)
* then feed a DAC with the data to generate analog signals again. (time domain)



Klaus

 

[simplsoft]

But still the confusion is the difference between low pass filter and amplifier. Bot amplifier and filter can work in time and frequency domain the difference is not clear. Can you explain please.

 

[c_mitra]

But still the confusion is the difference between low pass filter and amplifier. Bot amplifier and filter can work in time and frequency domain the difference is not clear. Can you explain please.

To understand a filter and how it works, it is suggested that you consider a mechanical analogy. Consider a mass suspended from a spring.

That is having a characteristic frequency; the system absorbs energy and oscillates with higher amplitude with higher supply of energy.

But how this energy is supplied? The energy must be supplied at the characteristic frequency of the system. (remember electrons jumping from one energy level to another?)

But what happens if we supply energy at another frequency (slightly different from the characteristic frequency)- how this will react.

That is called forced vibration. It can still absorb energy but much less. The energy supplied at another frequency (a periodic force) has two parameters: amplitude and phase.

How the amplitude depends on the frequency? That is given by the absorption curve. How the phase is affected? That is described by the dispersion curve.

For the system (mass with a spring) has two parameters: characteristic frequency (resonance) and damping (dissipative part). Just like C and R in a simple filter.

Now you translate this to the electronics. The basic ideas do not change.

 

[KlausST]

Hi,

What explanation do you expect?

I'd say they are different things, but often they are combined.
You may use a passive filter, you may use an amplifier without filter, you may combine them to an active filter.
A passive filter will reduce the energy of a signal, the amplifier may increase the energy.

Maybe one can compare it with a pulley (filter) and a motor (amplifier)
The pulley may reduce force, but it increases the length...but it won't reduce the energy.
Let's say you want to lift a stone with 200kg in 1m of height. With an ideal 10:1 pulley the firce may be reduced to 20kg...but you need to pull 10m.
The motor helps to reduce the energy for you to move the stone.

Klaus

 

[ronaldomponte]

@simplsoft, as it was already mentioned, mathematically speaking, filters and amplifiers perform different functionalities.
Therefore, they are different in essence.
This is the first step you need to understand. If you don't see this mathematical difference, we can not make progress.
Active filters are a class of filters that use an amplifier mainly to avoid the passive filter network to be loaded and give signal amplification when needed.
So, when you get the transfer function in s-domain you will see something like H(s) = A/(s+wo), for a first-order low pass filter, being A the gain and wo the cut-off frequency.
Separating the wheat from the tare on the circuit might be tricky sometimes since they can be seen as one entity performing two things at the same time.
It is not that difficult to understand, but the more you stir the nature of things you may change from electronics to metaphysics and then is much easier to get confused.
That's why I have prompted the water/ice analogy, which I really hope it has not complicated things more than helping.
Cheers.