This is actually how non-interrupting fiber locators and fiber taps can work. They use a tight deflection, below maximum curve radius but still pretty tight, and they capture the light bleeding from the bend. Unless you have very high end transceivers on the terminating side the loss will barely register. The most common use of this is with line technicians. When you’re going through a 128 strand bundle you often are required to double check any strands for activity prior to starting a splice. Its not too difficult to accidentally pull the wrong strand or be on the wrong bundle (especially when people sometimes leave hand scrawled tags or it’s a random mid-line repair and you’re standing in a cramped cold service box).
Also with locating beacon light (a very bright beam used for finding a specific strand several times stronger than normal transmission power) you can actually visibly see the leakage on bends (even through outer plastic sheathing). You don’t want to ever shine those in your eyes however…
@@LogicalNiko It's visible light, so its easy to keep that away from you. The silent killer is those +20 dBm EDFA outputs. They'll burn plastic on an unfocused beam
@@DilemmaCS yep. Most standard drivers in visible light don’t have the power to cause any issue really. Illuminating drivers can cause temporary blindness in areas of your retina, but they usually recover quickly unless you repeatedly hit them with high intensity light. But yes, as you said once you go outside your perception range it gets a lot more dangerous. I’m sure many veteran techs have a few random extra blind spots. Luckily our brains are great at dealing with that…but still.
The amount of energy that remains in each channel is determined by the coupling rate per unit distance, the coupled mode mixture, waveguide thickness, radiation wavelength, distance between the guides and the distance of the coupling region. You should be able to find parameters in the single mode case that makes the energy alternate between the guides as they traverse the coupling region and the remaining energy in each guide at that point is determined by the distance of the coupling region being a certain multiple of the distance traversed per unit coupling cycle.
I did my final year project for my physics bsc simulating transverse modes in frequency domain, inside optical fibres, did some simulations of coupling and how the effective refractive indicies change as a function of wavelength (transfering the mode from primarily excited in one to the other), seeing the coupling in the time domain is just fucking beautiful, awesome stuff.
It's fantastic to see the effects of impedence mismatch at the source exiting the bulk material and entering the fiber/space region, then again when it enters on the other side...
my guess for the leakage at the boundaries is that outside it is rectangular waveguide with perfectly conducting walls, and in the middle it is actually guided by index contrast, hence the abrupt change introduces leakage.
I plan to go into interesting things like these kinds of obscure physics when I eventually make it to a good university, and simulations like this remind me of what I'm fighting for. 💗
Thank you so very much for creating such an engaging explanation of waveguide physics, underscored by a composition that is the perfect complement to your video.
@@violaanderson175 I've watched over 100 other videos on this topic already by experts in this field. It's weird what people call nonsense when they haven't done any research on the matter at all, too funny actually.
The light in the tune looks *ALMOST* exactly like the auras I see before I get a migraine. It's so close, and so vivid, my head hurt just looking at the screen.
@Adam-ru3km as someone with absolutely no clue, my first guess would be they're different models describing the same interaction, but emphasizing the dominant element. In inductive coupling the magnetic field would dominate, in capacitive the electric field would dominate, in this I'd imagine both are sufficiently weak (or perhaps canceled near a resonant point) that it's instead dominated by fringe em wave effects. That said, I don't know what I'm talking about haha.
I used these 2x2 fused fibre couplers in my PhD setups all the time. They are exactly equivalent to a mirror in free space optics (at least as far as I was concerned). However I also used a 3x3 coupler at one point which was nothing like a mirror. It was made in the same way as the 2x2 couplers, just with gently melting 3 fibres together instead of 2. (One input and output was to a large ring with pumps and gain and an output etc, one input and output to a small ring that acted as an overlayed interferometer with piezo length control. The final input was not used whilst the final output was used for an error signal as part of locking the fibre laser to a single frequency / single mode).
Nice! It looks like the wave that have propagated far enough in the secondary fiber is about 90 degrees out of phase.... I say to may self when I stop start the video :).
Straight forward enough that even though I don't know much if anything about fiber optics, I can tell there must be some signal leak when the cable bends and that leaked signal is picked up by the parallel cable. The how or why or math of it I have no idea about but I think I get a small grasp of the concept
Very nice! I would comment, though, that the "multi-mode" width shown is still pretty darn single mode by fiber industry standards. Consider a 125/9 um optical fiber, standard for 1550nm light (about 1000nm, or 1um, when you take the 1.5 refractive index into account) That permits 9 full wavelengths in the cross section of the channel, versus the 2.5 or so you show in your simulation, so you still get a very dominant transverse mode in the sim. OM2 multimode fiber has a 50um core, effectively 20 times wider than shown here. Your true single-mode channel makes me wish 1um core fiber was available, though. It might be interesting to see the difference between this multimode and a much wider channel.
Thanks! The finite-difference method I'm using here makes it hard to simulate wavelengths that are too small. So perhaps this simulation would be more realistic for EM waves somewhere between infrared and microwaves.
It's not just about the core size though, but also the refractive index difference of core and cladding. In a SM fiber the refractive index difference is tiny, so despite having a 9um core, it is truly single mode. Anything other than a Gaussian is not guided and decays quickly. I've used fibers below their cutoff wavelength and you can see when they go multimode by the output not being Gaussian anymore. You can get fibers with core diameters of 3um, but they are absolutely terrible at 1550nm. Their NA is big, their losses are terrible, and coupling is horrendous because even small misalignments cause large losses. Also if you use normal glasses, at some point the core is too small (NA too big) to even guide the Gaussian. With integrated photonics (optical chips) the issue is usually the opposite, the refractive index differences are large. So you need super tiny waveguides for single mode, which in turn means you need to be able to manufacture very small features and it's hard to couple into a fiber due to the mode field diameter differences.
I'm aware the IoR is quite large. For smaller values, there would be much more leakage for the geometry used here. Real fibers have several layers of different IoR to prevent that.
I'm curious. While the wave propagation displayed is obviously two-dimensional, are the calculations in three dimensions or are those also two-dimensional? Also, I'm even more curious if you could use the same basic LightWave propagation concept to simulate shining a light in a four-dimensional environment, and using a lens in that environment to project some of the reflected light onto a two-dimensional surface, and show that surface as an image of the four dimensional environment, perhaps with a few colored four-dimensional polytopes in it.
This simulation uses two dimensions. You can define the wave equation in any dimension. However, simulating it with a finite elements methods, as used here, becomes quickly more time-consuming. For a simulation lattice with 1000 points on each side of an n-dimensional cube, the computation time scales like 1000 to the power n.
what i would be really interested to see is what CWDM/LWDM look like when theres several different wavelengths being used over a single strand of singlemode...
Very nice simulation. We can go deeper into detail, if you wish so, but I don't want you to take this as a request. Fiber optics have different layers with different drifraction values to get it to bend internally before reaching the end of the cable so the light beam reaches the reflective surface at a shallower angle, to reduce losses along the path.
my hungover brain right now doesn't even want to know why some parts of the cable bleed light more than others, but I still found it mesmerizing to look at
How does this process differ from RF/IR electromagnetic signaling? I know distance is the most limiting significant factor for inductive coupling. What makes these phenomena so completely different?
What similation program are you using here? Could this be used for audio frequencies? Im designing transmission lines for low frequency infra sound and need a way to simulate enclosures. Thanks ahead 😊
Oh no, if you would have run the simulation just a tiny bit longer you would have seen the reflection of the impedance discontinuety returning in the second fiber. You see the beginning of it returning when the wave starts to wobble in the second path (thats the standing wave expanding). I would like to see more simulation on that phenomena. But I like your visualization verry much of how it does explain how nearby waveguides exchange fields and how a signal can be split even without directly contacting it. Keep on exploring and uploading your results, they are very good quality.
So could you use an adjacent optical cable as a signal amplifier? Or even to scramble? Forgive my ignorance, I am a normal person. Also, what's up with the big leakage at the walls? At first I thought it was because of the bend, but it seems not.
Uses include beam splitters, sensors, and powering devices wirelessly - see for instance en.wikipedia.org/wiki/Evanescent_field#Evanescent-wave_coupling . The leakage is due to the abrupt change of the exterior medium, from completely reflecting to merely damping with a lower index of refraction.
if you use an S-matrix formulation you can solve this quicker and also then scale lengths and widths and recalculate instantly. Curious - the bending into the coupling regime is non-adiabatic?
I think the S-matrix formulation is only exact for very specific geometries, when you know the eigenelements of your system. Otherwise it is at best approximate.
@@NilsBerglund You're correct that it's for specific geometries but you can "couple" those geometries together and account for the scattering into each eigenmode of each geometry. So in that sense it's no more an approximation than FDTD / FEM. it's then a question of how efficient that model is. I used to work in photonics
Are photons actually able to tunnel to the other side of the insulation sheath? This would mean you have signal loss even in ideal conditions, and you must apply significant physical separation of fibers in all situations.
The is always some probability of tunneling. But it decreases exponentially with the width of the sheath, so it becomes extremely small when the sheath is wide enough.
This demonstration does a good job of showing how it is possible. That being said: In industry: The individual fibers aren't bundled up against each other. Each one is wrapped in light blocking protective material.
It's probably a bit of both. When the fiber emerges from the wall, it's the first opportunity for the light propagating at a too large angle to the interface to escape. The bend tends to increase that effect.
interesting, I know nothing about this subject but I noticed that the effect that the coupling has is sort of an "extraction" of the leaked information from the live fiber. Could this be used or is this ever used to an advantage such as; reading the leaked light wave information for error correction or even using this property as a way to condition the resulting light wave?
Apparently, it is used in beam splitters, sensors, and in powering devices wirelessly - see for instance en.wikipedia.org/wiki/Evanescent_field#Evanescent-wave_coupling . There may be other uses.
@@NilsBerglund i think they appear by simply having a glass silver interface with incident light ray in the glass. I will check. Most papers about this are with nano particles but they exist on very simple interface cases
Yes. SPP is just a solution of maxwell equations into simple dielectric functions. If you can simulate évanescent waves at total reflection angles, you can simulate a SPP. it’s just a specific case where the ray is trapped at the interface because something similar to total reflection occurs in both ways, creating a double évanescent wave propagating in the interface plane. The simplest way is using a silver thin film on a prism. To couple the incident ray with the SPP wave, you need a particle, defect, roughness or grating region. I’m pretty sure a simple Drude Ag model will do the job. Only thing is finding the proper coupling feature and incident angle
So is the coupling between the fibers like an induction of the same signal in the first fiber to the second even though the second isn't connected to the source?
One can describe it like that. Except that rather than a only a constant magnetic field that one would have for copper wires, here we have an oscillating electromagnetic field.
A possible explanation for the left-hand boundary is that this is the first chance for waves that propagate at a too large angle from the boundary to escape. Once they have, waves have to disperse before they are able to leak. This does not really explain the leakage at the right-hand boundary, though.
@@NilsBerglund you could have some reflection at the right-hand boundary if you have a substantial effective refractive index mismatch between the two sections of the fiber.
Depends on what you mean by overlap. What matters the most is the distance between the fibers. The evanescent wave decays exponentially on a scale given by the wave length. Thus the distance between fibers should be at most comparable to the wavelength. The length of the section where the fibers are close can have an effect on the amount of energy transmitted.
It cycles as well. Once all the energy has coupled into the second fibre you're effectively back at the starting point again. At least that's how I remember it from 25 years ago😁.
2 Fibers singing: F1: *“Wake me up”* F2: *“Wake me up inside”* F1: *“Can't wake up”* F2: *“Wake me up inside”* F1: *“Save me....”* F2: *“Call my name and save me from the dark”* 😂
So... Is this how you tap into a fiber optic line without breaking the the connection? Also could this be used to eaves drop on a quantum entanglement secured optical communications?
That is one way of eavesdropping. I'm not sure about the quantum situation - I guess if it is properly encrypted that offers an extra level of protection. Also, such a tap may be easier to detect in that case.
Yes this can be used for eavesdropping. I'm summarising; but with quantum key distribution you only send single photons and the eavesdropper doesn't know what the polarisation of the 'stolen' photon was so they can't 're-inject' it back down the line.
@@richardforster1239 consider that you get light leakage from fiberoptic lines if they bend and also absorption from the material, how does such a quantum key system deal with that? Surely they have to send multiple photons of the same polarity, otherwise any leaks/absorption would cause a false positive indicating someone trying to spy on the line.
@@NilsBerglund yes encryption would be a problem, but obtaining the data is half the battle, once you have it you can work on decryption at your leisure.
@@jonmcentire You're on the right lines. The intended recipient doesn't know the correct polarisation state to measure half the time anyway and so doesn't get a trustworthy answer (but they don't know it yet, you don't know if you get a photon through because your polariser was at the right angle or whether it was at 45 deg one way or the other and you also got lucky). So many more photons than necessary are sent and later on a public discussion occurs about which photons to use to form the key, and this process will detect man in the middle tapping. The wikipedia page explains it far better than I can. Look for BB84.
This is a purely classical simulation, of an effect used for instance in optoelectronic components such as beam splitters. I'm not an expert, but there is likely a quantum analogue of this.
Is it weird that I can tell the wires are following the curve y= +/- e^(-1/x^2)? Because I'm like 90% sure that's the exact curve used. This has nothing to do with the video, it's a great visual, I just noticed it really quick.
@@NilsBerglund Ooo, I graphed this out and I can see how I confused it. The exponential looks less sharp compared to this, weirdly enough, but they are very similar shapes. Thanks for sharing!
i like how this is an entirely serious video but half of the comments are either referencing the lobotomy meme or Evanescence, the band. both the video and the comments are just brilliant to me, but for different reasons.
Hm, I don't know how that works. But I have a couple of simulations of seismic waves, see ruclips.net/video/Zv5uL0Xk0nE/видео.html and ruclips.net/video/AiITAG8Kr_8/видео.html
Is the wave propagation from the point at the start - where it crosses out of the reflective boundary - due to the curve in the medium or because the outside boundaries more reflective than the medium? Just curious to know if the light coming off is doing so all the time or just when there’s a curve in the medium that is sheer to the waveform.
I think it's mainly that the medium outside the fiber changes from being completely reflecting (wave speed zero) to having a low index of refraction (higher wave speed than in the fiber).
What about modulated light sources? Like CW modulation and amplitude modulation for a time far longer than wave period. How will the modulation signal be at the end of the coupled fiber? Could it be totally replicated?
Here are some examples with amplitude-modulated signals: ruclips.net/video/nxhrIs7TGOc/видео.html ruclips.net/video/7jOpLqNcj4o/видео.html ruclips.net/video/aHkAPXTYaAs/видео.html There is quite a lot of deformation, but the fibers I simulate here are of rather poor quality compared to industrial fibers.
@@NilsBerglund Thanks! the deformation and noise might be due to the pulse nature of the modulation, passing from 0 to 1 in no time, and such its Fourier transform of infinite components. If the pulse is longer with respect to wave period, it might be recognizable at the end.
This is a purely classical simulation, while entanglement is a quantum concept. Perhaps entanglement can play a role in the quantum analogue of this situation, but this is not what is illustrated here.
It is a classical effect. Even when ray optics predicts total internal reflection, with wave optics there is a so-called evanescent wave leaking from the fiber. Its intensity decreases exponentially with the distance, but if the other fiber is close enough, a wave can be excited in that fiber.
Cool sims. I would add that modern fiber is graded, which limits the external fields and improves transmission losses-it also makes evanescent leakage very small. The bends you are showing are greatly exaggerated, and you seem to have a problem at your boundaries that causes a lot of leakage.
Thanks. This is really a quite simple set-up I came up with, that uses only 3 different regions, with different IoR and dissipations. I'm sure one can do much better, for instance by including more layers.
By putting two optical fibers next to each other, one of them can pick up some information going through the other one. You can use that for eavesdropping, or as a beam splitter.
Also is this model showing electricity or whatever is being modeled gets transfered from a cable that is active to an inactive cable simply by touching it?
The simulation solves the 1d wave equation. It could represent the electric field in an electromagnetic wave, but it could also apply to sound waves, for instance.
That's what I was thinking, fibre optic communications are supposed to be secure as you can't tap into them without breaking the connection first and alerting them to your presence. But with this way you could read the data without damaging the connection
In the leftmost and rightmost parts, the outside of the fibers is completely reflecting. In the middle mart, the outside has a lower index of refraction and positive damping. Some waves with a too large angle with respect to the interface can escape there.
This is actually how non-interrupting fiber locators and fiber taps can work. They use a tight deflection, below maximum curve radius but still pretty tight, and they capture the light bleeding from the bend. Unless you have very high end transceivers on the terminating side the loss will barely register.
The most common use of this is with line technicians. When you’re going through a 128 strand bundle you often are required to double check any strands for activity prior to starting a splice. Its not too difficult to accidentally pull the wrong strand or be on the wrong bundle (especially when people sometimes leave hand scrawled tags or it’s a random mid-line repair and you’re standing in a cramped cold service box).
Smart
Still you shouldnt trust devices that arent yours :) - no matter if prism is big or small
Also with locating beacon light (a very bright beam used for finding a specific strand several times stronger than normal transmission power) you can actually visibly see the leakage on bends (even through outer plastic sheathing).
You don’t want to ever shine those in your eyes however…
@@LogicalNiko It's visible light, so its easy to keep that away from you. The silent killer is those +20 dBm EDFA outputs. They'll burn plastic on an unfocused beam
@@DilemmaCS yep. Most standard drivers in visible light don’t have the power to cause any issue really. Illuminating drivers can cause temporary blindness in areas of your retina, but they usually recover quickly unless you repeatedly hit them with high intensity light. But yes, as you said once you go outside your perception range it gets a lot more dangerous.
I’m sure many veteran techs have a few random extra blind spots. Luckily our brains are great at dealing with that…but still.
The amount of energy that remains in each channel is determined by the coupling rate per unit distance, the coupled mode mixture, waveguide thickness, radiation wavelength, distance between the guides and the distance of the coupling region.
You should be able to find parameters in the single mode case that makes the energy alternate between the guides as they traverse the coupling region and the remaining energy in each guide at that point is determined by the distance of the coupling region being a certain multiple of the distance traversed per unit coupling cycle.
Thanks!
Shouldn’t they have optical insulation around them in real life?
Not if you’re the one who removed the insulation for surveillance purposes
I did my final year project for my physics bsc simulating transverse modes in frequency domain, inside optical fibres, did some simulations of coupling and how the effective refractive indicies change as a function of wavelength (transfering the mode from primarily excited in one to the other), seeing the coupling in the time domain is just fucking beautiful, awesome stuff.
Interesting. Feel free to share any good parameter values that you might know of.
Could you please share me your bachelor's thesis?😀
First wire: WAKE ME UP
Second wire: WAKE ME UP INSIDE!
It's so dumb but I'm laughing like a dork at it xD
CALL MY NAME AND SAVE ME FROM THE DARK
"Jr.. did you let the optical fibers touch? How many times have I told you not to let your optic fibers touch!! Your grounded young man."
-some dad
And this is how the flux capacitor was invented.
It's fantastic to see the effects of impedence mismatch at the source exiting the bulk material and entering the fiber/space region, then again when it enters on the other side...
Wow the coupling brings it to life!
How many people will search "Evanescence" and wind up here unexpectedly :D
Hi
Six.
Oops, I did a clever marketing :)
Huh... "bring me to life"?
BRING ME TO LIGHT
This video really brings me to life!!
Oh, I had a Fiber Optics course in the university three years ago. And your videos are excellent visual explanation about optical phenomena.
Awesome, thank you!
my guess for the leakage at the boundaries is that outside it is rectangular waveguide with perfectly conducting walls, and in the middle it is actually guided by index contrast, hence the abrupt change introduces leakage.
maybe try replacing it with a gradual change over a couple wavelengths?
That might work. Real optical fibers have several layers with gradually changing IoR.
I plan to go into interesting things like these kinds of obscure physics when I eventually make it to a good university, and simulations like this remind me of what I'm fighting for. 💗
Best of luck!
Thank you so very much for creating such an engaging explanation of waveguide physics, underscored by a composition that is the perfect complement to your video.
This looks like the future of ultra-fast consistent computing that everybody can easily afford... quite revolutionary.
Do you see this future everywhere or here only? You are talking nonsense
@@violaanderson175 I've watched over 100 other videos on this topic already by experts in this field. It's weird what people call nonsense when they haven't done any research on the matter at all, too funny actually.
The light in the tune looks *ALMOST* exactly like the auras I see before I get a migraine. It's so close, and so vivid, my head hurt just looking at the screen.
Thanks, this is useful when explaining directional RF couplers to beginners.
How does it differ from inductive coupling?
@Adam-ru3km as someone with absolutely no clue, my first guess would be they're different models describing the same interaction, but emphasizing the dominant element. In inductive coupling the magnetic field would dominate, in capacitive the electric field would dominate, in this I'd imagine both are sufficiently weak (or perhaps canceled near a resonant point) that it's instead dominated by fringe em wave effects. That said, I don't know what I'm talking about haha.
I used these 2x2 fused fibre couplers in my PhD setups all the time. They are exactly equivalent to a mirror in free space optics (at least as far as I was concerned). However I also used a 3x3 coupler at one point which was nothing like a mirror. It was made in the same way as the 2x2 couplers, just with gently melting 3 fibres together instead of 2. (One input and output was to a large ring with pumps and gain and an output etc, one input and output to a small ring that acted as an overlayed interferometer with piezo length control. The final input was not used whilst the final output was used for an error signal as part of locking the fibre laser to a single frequency / single mode).
Nice, thanks for sharing!
Now this is fiber optic cable core
Nice! It looks like the wave that have propagated far enough in the secondary fiber is about 90 degrees out of phase.... I say to may self when I stop start the video :).
great example of why you should shield your optical fibers!
They’re shielded out of the factory. The real lesson is not to bend them too sharply.
Awesome and beautiful. Great job Nils!!
My father installs fiber optic extensimeters for a living, he will love this simulation
I love this channel and follow and enjoy every video. but I have very little idea what im looking at - It's just a joy
Straight forward enough that even though I don't know much if anything about fiber optics, I can tell there must be some signal leak when the cable bends and that leaked signal is picked up by the parallel cable. The how or why or math of it I have no idea about but I think I get a small grasp of the concept
Very nice!
I would comment, though, that the "multi-mode" width shown is still pretty darn single mode by fiber industry standards.
Consider a 125/9 um optical fiber, standard for 1550nm light (about 1000nm, or 1um, when you take the 1.5 refractive index into account)
That permits 9 full wavelengths in the cross section of the channel, versus the 2.5 or so you show in your simulation, so you still get a very dominant transverse mode in the sim.
OM2 multimode fiber has a 50um core, effectively 20 times wider than shown here.
Your true single-mode channel makes me wish 1um core fiber was available, though.
It might be interesting to see the difference between this multimode and a much wider channel.
Thanks! The finite-difference method I'm using here makes it hard to simulate wavelengths that are too small. So perhaps this simulation would be more realistic for EM waves somewhere between infrared and microwaves.
It's not just about the core size though, but also the refractive index difference of core and cladding. In a SM fiber the refractive index difference is tiny, so despite having a 9um core, it is truly single mode. Anything other than a Gaussian is not guided and decays quickly. I've used fibers below their cutoff wavelength and you can see when they go multimode by the output not being Gaussian anymore.
You can get fibers with core diameters of 3um, but they are absolutely terrible at 1550nm. Their NA is big, their losses are terrible, and coupling is horrendous because even small misalignments cause large losses. Also if you use normal glasses, at some point the core is too small (NA too big) to even guide the Gaussian.
With integrated photonics (optical chips) the issue is usually the opposite, the refractive index differences are large. So you need super tiny waveguides for single mode, which in turn means you need to be able to manufacture very small features and it's hard to couple into a fiber due to the mode field diameter differences.
Teenage me in the early 2000s would I've been very excited at the use of this word.. okay, who am kidding, I still am 😅😂
Wake me up!
The index of refraction in a single-mode fiber is about 1.47; your index is relatively high.
I'm aware the IoR is quite large. For smaller values, there would be much more leakage for the geometry used here. Real fibers have several layers of different IoR to prevent that.
I'm curious. While the wave propagation displayed is obviously two-dimensional, are the calculations in three dimensions or are those also two-dimensional? Also, I'm even more curious if you could use the same basic LightWave propagation concept to simulate shining a light in a four-dimensional environment, and using a lens in that environment to project some of the reflected light onto a two-dimensional surface, and show that surface as an image of the four dimensional environment, perhaps with a few colored four-dimensional polytopes in it.
This simulation uses two dimensions. You can define the wave equation in any dimension. However, simulating it with a finite elements methods, as used here, becomes quickly more time-consuming. For a simulation lattice with 1000 points on each side of an n-dimensional cube, the computation time scales like 1000 to the power n.
what i would be really interested to see is what CWDM/LWDM look like when theres several different wavelengths being used over a single strand of singlemode...
Very nice simulation.
We can go deeper into detail, if you wish so, but I don't want you to take this as a request.
Fiber optics have different layers with different drifraction values to get it to bend internally before reaching the end of the cable so the light beam reaches the reflective surface at a shallower angle, to reduce losses along the path.
Thanks! I'm considering it.
I don’t understand what the two vertical lines express
Waves cannot propagate outside the fibers at the left and right of the vertical lines, while they can in between.
Termination points of the fibre optic lines?
There it is. The first time I've we've heard the word Evanescence used outside of the band name.
Now this is the optic fiber cable core im here for
my hungover brain right now doesn't even want to know why some parts of the cable bleed light more than others, but I still found it mesmerizing to look at
have you ever done a simulation of mutual inductance and ghost voltage in electrical conductors? that would be cool.
No, not yet. I'm not sure this is doable with my codes, but if yes, I may get there.
@@NilsBerglund regardless, thank you for all the videos so far.
How does this process differ from RF/IR electromagnetic signaling? I know distance is the most limiting significant factor for inductive coupling. What makes these phenomena so completely different?
@@Adam-ru3km Signal frequency/amplitude, permeability, and a lot of other factors I'd guess.
That is the most gorgeous leaking of…anything….that I have ever seen.
Oooh this must be how you can see into my eyes like open doors
Excellent video! You can also show how energy reflected in main fiber that is pickuped by coupled line and traveled to begin of coupled fiber.
What similation program are you using here? Could this be used for audio frequencies? Im designing transmission lines for low frequency infra sound and need a way to simulate enclosures. Thanks ahead 😊
I'm using my own code, available on GitHub. Feel free to use it. It should work for sound waves as long as the wavelengths are in a suitable range.
how can you see into my eyes
like open doors?
leading you down into my core
where i've become so numb
Why haven't we had optical computing using this mechanism?
I think this exists, look for photonics and optoelectronics. Perhaps it has not yet reached a stadium where it is used in commercial electronics.
You do have this but this is a linear operation (if you want logic gates you need nonlinear)
I've seen data diodes that work on a similar principle to this.
is this fiber optic cable core? 😮
It is a fiber with only one layer, so I guess one could call it that.
Oh no, if you would have run the simulation just a tiny bit longer you would have seen the reflection of the impedance discontinuety returning in the second fiber. You see the beginning of it returning when the wave starts to wobble in the second path (thats the standing wave expanding). I would like to see more simulation on that phenomena. But I like your visualization verry much of how it does explain how nearby waveguides exchange fields and how a signal can be split even without directly contacting it. Keep on exploring and uploading your results, they are very good quality.
Thanks! I think you can see such a reflection in the simulation ruclips.net/video/nxhrIs7TGOc/видео.html . There will be more, however.
optical fiber cable core in its truest form
So could you use an adjacent optical cable as a signal amplifier? Or even to scramble? Forgive my ignorance, I am a normal person.
Also, what's up with the big leakage at the walls? At first I thought it was because of the bend, but it seems not.
Uses include beam splitters, sensors, and powering devices wirelessly - see for instance en.wikipedia.org/wiki/Evanescent_field#Evanescent-wave_coupling .
The leakage is due to the abrupt change of the exterior medium, from completely reflecting to merely damping with a lower index of refraction.
very cool and clear demonstration!
if you use an S-matrix formulation you can solve this quicker and also then scale lengths and widths and recalculate instantly. Curious - the bending into the coupling regime is non-adiabatic?
I think the S-matrix formulation is only exact for very specific geometries, when you know the eigenelements of your system. Otherwise it is at best approximate.
@@NilsBerglund You're correct that it's for specific geometries but you can "couple" those geometries together and account for the scattering into each eigenmode of each geometry. So in that sense it's no more an approximation than FDTD / FEM. it's then a question of how efficient that model is. I used to work in photonics
This brings me to life.
So an optic fiber can gain energy from just being near another energized fiber?
Yes, indeed. This is called coupling of fibers.
@@NilsBerglund very interesting. Thanks for the info
Are photons actually able to tunnel to the other side of the insulation sheath? This would mean you have signal loss even in ideal conditions, and you must apply significant physical separation of fibers in all situations.
The is always some probability of tunneling. But it decreases exponentially with the width of the sheath, so it becomes extremely small when the sheath is wide enough.
This demonstration does a good job of showing how it is possible. That being said: In industry: The individual fibers aren't bundled up against each other. Each one is wrapped in light blocking protective material.
Is this what splicing does,or does splicing somehow continue the initial fiber?
Splicing means "gluing" two fibers end to end, usually with an electric arc.
Is the light leaking from the corner of the bend, or is it something with the interaction with the wall the fiber is emerging from?
It's probably a bit of both. When the fiber emerges from the wall, it's the first opportunity for the light propagating at a too large angle to the interface to escape. The bend tends to increase that effect.
interesting, I know nothing about this subject but I noticed that the effect that the coupling has is sort of an "extraction" of the leaked information from the live fiber. Could this be used or is this ever used to an advantage such as; reading the leaked light wave information for error correction or even using this property as a way to condition the resulting light wave?
Apparently, it is used in beam splitters, sensors, and in powering devices wirelessly - see for instance en.wikipedia.org/wiki/Evanescent_field#Evanescent-wave_coupling . There may be other uses.
Could you show surface plasmons? It’s a somewhat difficult concept to explain and a video like this would be extremely helpful !!
That still seems to be out of reach of what I am able to simulate, but perhaps I can get there later on.
@@NilsBerglund i think they appear by simply having a glass silver interface with incident light ray in the glass. I will check. Most papers about this are with nano particles but they exist on very simple interface cases
Yes. SPP is just a solution of maxwell equations into simple dielectric functions. If you can simulate évanescent waves at total reflection angles, you can simulate a SPP. it’s just a specific case where the ray is trapped at the interface because something similar to total reflection occurs in both ways, creating a double évanescent wave propagating in the interface plane.
The simplest way is using a silver thin film on a prism. To couple the incident ray with the SPP wave, you need a particle, defect, roughness or grating region. I’m pretty sure a simple Drude Ag model will do the job. Only thing is finding the proper coupling feature and incident angle
Fiber optics waking up inside
So is the coupling between the fibers like an induction of the same signal in the first fiber to the second even though the second isn't connected to the source?
One can describe it like that. Except that rather than a only a constant magnetic field that one would have for copper wires, here we have an oscillating electromagnetic field.
Hmm, does anyone have any idea of why there is point source like leakage when the fiber crosses the media boundaries?
A possible explanation for the left-hand boundary is that this is the first chance for waves that propagate at a too large angle from the boundary to escape. Once they have, waves have to disperse before they are able to leak. This does not really explain the leakage at the right-hand boundary, though.
It seems to only happen on the obtuse side of the angle as well
@@NilsBerglund Would you bet it reflects an actual phenomenon or could it be blamed on the simulation?
@@NilsBerglund you could have some reflection at the right-hand boundary if you have a substantial effective refractive index mismatch between the two sections of the fiber.
@mr_rede_de_stone916 I think it is an actual phenomenon, except that real optical fibers have several layers that help reducing this effect.
Nice! The overlap length is in order of wavelengths correct? The longer the overlap the better the transmission i guess?
Depends on what you mean by overlap. What matters the most is the distance between the fibers. The evanescent wave decays exponentially on a scale given by the wave length. Thus the distance between fibers should be at most comparable to the wavelength. The length of the section where the fibers are close can have an effect on the amount of energy transmitted.
It cycles as well. Once all the energy has coupled into the second fibre you're effectively back at the starting point again. At least that's how I remember it from 25 years ago😁.
Good for them
This is beautiful. Could you do one where a multimodal input is coupled to many single mode outputs? E.g., as in a photonic lantern
Thank! Maybe, I'll have to check.
You might say it's like the one fiber is waking the other up inside
The secondary fiber seems like it takes on a much cleaner signal, is this effect used to de-noise optical cables?
I'm not sure the signal is really cleaner, it may be an effect of the color gradient used.
Cool but, why does it remind me of bacteria kinda, or just micro organisms in general?
how thin can a optic fiber be before light of given wavelength no longer conduct inside ?
Th fiber stops being conducting when its width is comparable to the wavelength. So in effect, the width should be a bit larger than the wavelength.
Useful to see a comparison with MEEP
Can you make a video where full transfer happens in the second waveguide?
I don't know if this is possible. Maybe one has to choose very specific parameter values (refractive index, size of fibers).
@@NilsBerglund It is possible, just the propagation constant and straight section in the video for coupling matters
2 Fibers singing:
F1: *“Wake me up”*
F2: *“Wake me up inside”*
F1: *“Can't wake up”*
F2: *“Wake me up inside”*
F1: *“Save me....”*
F2: *“Call my name and save me from the dark”*
😂
So... Is this how you tap into a fiber optic line without breaking the the connection? Also could this be used to eaves drop on a quantum entanglement secured optical communications?
That is one way of eavesdropping. I'm not sure about the quantum situation - I guess if it is properly encrypted that offers an extra level of protection. Also, such a tap may be easier to detect in that case.
Yes this can be used for eavesdropping. I'm summarising; but with quantum key distribution you only send single photons and the eavesdropper doesn't know what the polarisation of the 'stolen' photon was so they can't 're-inject' it back down the line.
@@richardforster1239 consider that you get light leakage from fiberoptic lines if they bend and also absorption from the material, how does such a quantum key system deal with that? Surely they have to send multiple photons of the same polarity, otherwise any leaks/absorption would cause a false positive indicating someone trying to spy on the line.
@@NilsBerglund yes encryption would be a problem, but obtaining the data is half the battle, once you have it you can work on decryption at your leisure.
@@jonmcentire You're on the right lines. The intended recipient doesn't know the correct polarisation state to measure half the time anyway and so doesn't get a trustworthy answer (but they don't know it yet, you don't know if you get a photon through because your polariser was
at the right angle or whether it was at 45 deg one way or the other and
you also got lucky). So many more photons than necessary are sent and later on a public discussion occurs about which photons to use to form the key, and this process will detect man in the middle tapping. The wikipedia page explains it far better than I can. Look for BB84.
Awesome!! Is this the same principle used in optical quantum computers ?
This is a purely classical simulation, of an effect used for instance in optoelectronic components such as beam splitters. I'm not an expert, but there is likely a quantum analogue of this.
Is it weird that I can tell the wires are following the curve y= +/- e^(-1/x^2)? Because I'm like 90% sure that's the exact curve used.
This has nothing to do with the video, it's a great visual, I just noticed it really quick.
Actually, the function is a + b*(1+2*x^6)/(1+x^6) for suitable a and b. An exponential would have been steeper, I think.
@@NilsBerglund Ooo, I graphed this out and I can see how I confused it. The exponential looks less sharp compared to this, weirdly enough, but they are very similar shapes. Thanks for sharing!
i like how this is an entirely serious video but half of the comments are either referencing the lobotomy meme or Evanescence, the band. both the video and the comments are just brilliant to me, but for different reasons.
I wish I had time to study again a lot of subjects from graduation that I didn't understand very well. We have to learn kinda fast.
I understand that some scientists are using the signals in fiber optic cables as seismometers. Would you be able to model that somehow?
Hm, I don't know how that works. But I have a couple of simulations of seismic waves, see ruclips.net/video/Zv5uL0Xk0nE/видео.html and ruclips.net/video/AiITAG8Kr_8/видео.html
Must be common knowledge in the spy game :) Beautiful simulation, as always!
Many thanks!
"How can you see into my eyes?"
By coupling two optical fibers apparently
Problem solved!
Is the wave propagation from the point at the start - where it crosses out of the reflective boundary - due to the curve in the medium or because the outside boundaries more reflective than the medium?
Just curious to know if the light coming off is doing so all the time or just when there’s a curve in the medium that is sheer to the waveform.
I think it's mainly that the medium outside the fiber changes from being completely reflecting (wave speed zero) to having a low index of refraction (higher wave speed than in the fiber).
That is one way to tap all communications without anyone being able to notice.
My takeaway is that you could build a passive minimally invasive optical coupler and sneak out some data
This is indeed one application of this.
Why is it that light is only transmitted to the left-hand side of the second fibre?
Wonder if this could be used in light based circuit design somehow
It is used. Search for fiber-optic splitters and directional couplers if you want to find out more.
What about modulated light sources? Like CW modulation and amplitude modulation for a time far longer than wave period. How will the modulation signal be at the end of the coupled fiber? Could it be totally replicated?
Here are some examples with amplitude-modulated signals:
ruclips.net/video/nxhrIs7TGOc/видео.html
ruclips.net/video/7jOpLqNcj4o/видео.html
ruclips.net/video/aHkAPXTYaAs/видео.html
There is quite a lot of deformation, but the fibers I simulate here are of rather poor quality compared to industrial fibers.
@@NilsBerglund Thanks! the deformation and noise might be due to the pulse nature of the modulation, passing from 0 to 1 in no time, and such its Fourier transform of infinite components. If the pulse is longer with respect to wave period, it might be recognizable at the end.
it's so beautiful!
I have zero idea what is going on right now but it looks cool
...does this mean we'll one day need twisted pair fiber optics now?
Is this also some sort of photon-entanglement?
This is a purely classical simulation, while entanglement is a quantum concept. Perhaps entanglement can play a role in the quantum analogue of this situation, but this is not what is illustrated here.
I don't understand. The fibers are isolated? The propagation is due the quantistic mechanical effect?
It is a classical effect. Even when ray optics predicts total internal reflection, with wave optics there is a so-called evanescent wave leaking from the fiber. Its intensity decreases exponentially with the distance, but if the other fiber is close enough, a wave can be excited in that fiber.
@@NilsBerglund Thanks
Cool sims. I would add that modern fiber is graded, which limits the external fields and improves transmission losses-it also makes evanescent leakage very small. The bends you are showing are greatly exaggerated, and you seem to have a problem at your boundaries that causes a lot of leakage.
Thanks. This is really a quite simple set-up I came up with, that uses only 3 different regions, with different IoR and dissipations. I'm sure one can do much better, for instance by including more layers.
@@NilsBerglund most absorbing boundaries have limitations-they work better if the fields are propagating normal to the boundary.
I understand now. My life is complete.
Fiber optic cable core 😊
like a big worm going through a cave, and that s where the pebbles vibrate... what is this for exactly? can someone explain in layman s terms?
By putting two optical fibers next to each other, one of them can pick up some information going through the other one. You can use that for eavesdropping, or as a beam splitter.
is not a reader of some kind better for this? i imagine...@@NilsBerglund
Is this a model of electric flow 2 fiber optics under certain conditions are some other model?
Also is this model showing electricity or whatever is being modeled gets transfered from a cable that is active to an inactive cable simply by touching it?
The simulation solves the 1d wave equation. It could represent the electric field in an electromagnetic wave, but it could also apply to sound waves, for instance.
how much was the dB loss?
This looks like a photonic transistor.
Idk if it is or if its not
So is this how you can spy on communication lines without cutting it open?
I guess Van Eck phreaking is a related phenomenon, though I'm not sure it relies particularly on evanescent fields.
That's what I was thinking, fibre optic communications are supposed to be secure as you can't tap into them without breaking the connection first and alerting them to your presence. But with this way you could read the data without damaging the connection
@@Osama-Bon-Jovi-01 or inject
Question: In case of single mode; are two outputs in phase? It seems so but I'd like to see the actual phase comparison.
This may depend on the distance between the fibers, and the length of the segment where they are close to each other.
What's the boundary causing amplitude loss on each side?
In the leftmost and rightmost parts, the outside of the fibers is completely reflecting. In the middle mart, the outside has a lower index of refraction and positive damping. Some waves with a too large angle with respect to the interface can escape there.
What? Is real ? Separate fiber interferencia. How?
So grateful for you. Thanks
I got a braingasm from this
Could you also simulate this but with two fibers that have different diameter. And another simulation where the lower fiber gets smaller diameter.
What do you expect to happen then?