Difference between Optical bandpass filter and normal/electrical bandpass filter

[thavamaran]

Hi there, I am very new to optical domain, and I am only used to the common bandpass filter that we use on circuits, those RLC and OP-AMP based bandpass filter.

I went through few books and can't really find the difference, do anyone have any link regarding this issue or idea about this.

Thank you.

 

[Antiphon]

Optical bandpass filters are multilayer thin films. You can treat this like a set of short transmission line sections of alternating admittance. This is more like a microwave circuit than a lumped element filter. It has an infinite number of poles unlike a circuit-based filter which has one pole per lumped element.

 

[thavamaran]

Optical bandpass filters are multilayer thin films. You can treat this like a set of short transmission line sections of alternating admittance. This is more like a microwave circuit than a lumped element filter. It has an infinite number of poles unlike a circuit-based filter which has one pole per lumped element.

Im still in the stage of simulating the filter, the problem here is, i need the transfer function of a optical bandpass filter in order to simulate the entire channel of optical network.

But then, normal circuity filters transfer function could be derived by considering capacitors and resistors and inductors. But in optical case, how do we get them. I went through a few papers, but yet i can't get the idea.

 

[sophiecentaur]

There are many hits on Google if you type in "interference filters". They are a form of transversal filter, working essentially in the time domain and the characteristic is not too hard to determine - see the more learned of the links for the formulae.

 

[Antiphon]

There are many hits on Google if you type in "interference filters". They are a form of transversal filter, working essentially in the time domain and the characteristic is not too hard to determine - see the more learned of the links for the formulae.

This is not quite right. Optical filters do not work in thetooke domain any more than circuits do.

They are microwave circuits, not lumped circuits. That means they are sections of cascaded transmission lines, not capacitors and inductors.

 

[sophiecentaur]

This is not quite right. Optical filters do not work in thetooke domain any more than circuits do.

They are microwave circuits, not lumped circuits. That means they are sections of cascaded transmission lines, not capacitors and inductors.

Yes, they are (like?) some microwave circuits.
I should, perhaps, have said that, as they are transversal filters, they are easier to analyse (and certainly easier to visualise) in the time domain - as are many microwave transmission filters. I realize it all boils down to the same thing in the end but it is often easier to look at some things from one point of view than another.
If you were to try to appreciate the behaviour of an oil film in the frequency domain, you would have difficulty. As a simple example of interference, based on path length difference (i.e. time difference), it is very easy to comprehend.

 

[Antiphon]

Yes, they are (like?) some microwave circuits.
I should, perhaps, have said that, as they are transversal filters, they are easier to analyse (and certainly easier to visualise) in the time domain - as are many microwave transmission filters. I realize it all boils down to the same thing in the end but it is often easier to look at some things from one point of view than another.
If you were to try to appreciate the behaviour of an oil film in the frequency domain, you would have difficulty. As a simple example of interference, based on path length difference (i.e. time difference), it is very easy to comprehend.

Sometimes I have to be very abbreviated when I'm on a tiny mobile keyboard. You are right, they are like microwave circuits. Properly speaking they are one-dimnsional optical waveguides. By calling them microwave circuits I meant they can be understood (and analyzed) mathematically as sections of transmission lines cascaded together and these are the same equations used to design micrwave circuit filters. In these types of filters, there are no inductors or capacitors. The inductance and capacitance is disrtibuted along the length of the transmission line section. Over a narrow frequency range you can make a mathematical transformation that makes them look like inductors and capacitors. You can then use some standard circuit filter ideas and then transform back to get the necessary transmission line parameters like length and characteristic impedance.


Where the multilayer filters are different than the usual microwave designs is that in almost all microwave circuits you are free to choose the impedance and length of the casscaded sections. Usually the sections are all the same electrical length (1/4 or 1/2 the center wavelength) but have a varying impedance profile. There are design tables for creating any type of bandpass profile you want to by connecting these transmission line sections.

By contrast the thin-film optical filters are limited to a certain set of materials (for economical designs anyway) and the only free design parameter is the thickness of each section. Most filter are produced using only two types of material, usually TiO2 and Sio2. This would be like designing filters using only two types of transmission lines but allowing the lengths to vary. While it can be done, there are very few design rules available for how to synthesize the filter. Most designs are numerically optimized with no analytical foundation whatever.

As for the time domain, I must disagree with you. How does the time domain help you visualize that the oil film will look say green and not red? The frequncy domain is much more nautral for this because you can envision green light being a wavelength match to the sizes of the layers and passing through while other wavelengths are reflected. The time-domain thinking allows you to see time delays and reflections very well, but not spectra which is how colors are best thought about.

FGor the OP, you need the transfer function of a transmission line. It involves the tangent function and can be either capacitive or inductive depending on the electrical size. If you really want to know how to do this, I can find you some references. The math is involved but not really difficult.

 

[sophiecentaur]

Interference filters work on differences in transit times. When the time / path differences are appropriate, you get cancellation or addition of reflections. If you have a quarter wave thick layer the delay both ways will cause a canceled reflection. This, of course, corresponds to a particular frequency but the thing works due to time delays. A lumped component circuit doesn't work in a way that is so obviously 'temporal'. That is all I was getting at.