1. Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.
    Dismiss Notice

Multimode optical fiber bandwidth

Discussion in 'Mathematics and Physics' started by neptune, Oct 10, 2014.

  1. neptune

    neptune Member

    Joined:
    Jun 14, 2009
    Messages:
    466
    Likes:
    5
    Location:
    New delhi, India
    I have always been confused on this topic.
    According to the diagram there many modes (Spatial multiplexing) in a multimode fiber, and each of the mode can carry light of different frequencies (Wavelength division multiplexing), if it is so then why the bandwidth of multimode fiber is less then Single mode fiber where there is only one mode?
    more bandwidth meaning more data transmission rate in multimode fiber then single mode fiber.
    I am only asking for 10-100 meter of distance.
     

    Attached Files:

    Last edited: Oct 10, 2014
  2. steveB

    steveB Well-Known Member Most Helpful Member

    Joined:
    Jan 16, 2009
    Messages:
    1,297
    Likes:
    629
    There is no practical way to access the individual spatial modes to use them as separate channels. When you couple light into a multimode fiber, the light power gets distributed in all modes. Also, imperfections in the fiber and bends/stresses on the fiber causes the modes to cross couple and mix together as the light propagates down the fiber. Now the problem with many modes is that each mode travels at a different speed which causes multimode dispersion and that limits bitrate and distance severely.

    Singlemode fibers do not have this problem and can have much higher bandwidth.

    Wavelength division multiplexing is a different thing entirely. That is using different frequecies of light to allow separate channels. This can be done with either multimode or singlemode fiber.
     
    Last edited: Oct 11, 2014
  3. neptune

    neptune Member

    Joined:
    Jun 14, 2009
    Messages:
    466
    Likes:
    5
    Location:
    New delhi, India
    If there is no practical way to detect individual spatial modes to be used as separate channels then why do we use them ?
     
  4. dave

    Dave New Member

    Joined:
    Jan 12, 1997
    Messages:
    -
    Likes:
    0


     
  5. steveB

    steveB Well-Known Member Most Helpful Member

    Joined:
    Jan 16, 2009
    Messages:
    1,297
    Likes:
    629

    I didn't say there is no practical way to detect spatial modes. They are there and they can be detected. The problem is using the modes as individual exclusive communications channels, which includes coupling into the fiber, transmission through the fiber and then detection at the receiver. The detection problem might be the easiest to deal with, but the complexities of doing so don't make sense compared to other approaches. However, coupling into the fiber and selecting the individual modes and then maintaining the light in those modes without crosstalk between the other modes is very difficult. I would not say it is impossible, but it is very difficult and other ways are so easy, that no one tries to do this. I have seen research papers which try to do this with low modes fibers (e.g. bimodal), but it just is not practical from a cost point of view.

    But, why do we use them? We use them because they are there, and like many things in technology, there are pros and cons. There is also a historical aspect to this, because initially single mode fibers were not feasible and multimode fibers had to be used, and the inherent problems were dealt with as a matter of necessity.

    The benefit of having many modes is that coupling light into the fiber is much easier than single mode fiber. In the early days, only LEDs were practical, and LEDS can couple effectively into the multimode fiber. The problem with multimode fiber is the large modal dispersion which limits bitrate and distance.

    The advent and prolific use of singlemode fiber is basically, "not using" the modes. So often we don't use them.
     
    • Informative Informative x 1
  6. neptune

    neptune Member

    Joined:
    Jun 14, 2009
    Messages:
    466
    Likes:
    5
    Location:
    New delhi, India
    So if it is not impossible to use them as channels then would they surpass the bandwidth limit of single mode fiber ?Complexity and length not being the issue.
     
  7. steveB

    steveB Well-Known Member Most Helpful Member

    Joined:
    Jan 16, 2009
    Messages:
    1,297
    Likes:
    629
    In principle, I would say yes. There are also two polarization modes for each of the spatial modes. In principle, even these modes could be used as separate channels, which would double the bandwidth even further. Before this could be practical, better optical fibers would needed, which is asking a lot, since they are already amazingly good. Every imperfection in the fiber would need to be removed. Then, special cabling that puts no stress on the fiber would be needed. That is asking a lot too. If these steps are not taken, then the modes will cross couple, and crosstalk will be the death of a working system. Then, there are practical issues with coupling light into the spatial modes. Polarization is easier because lasers are polarized, but coupling to different spatial patterns is not really practical with today's technology.

    Now compare all that with use of wavelength division multiplexing on single mode fibers. The single mode fiber allows many GHz bit rates per channel, and greater than 50 channels is possible. If you need more, simply put more fibers in the cable. The optical fiber takes up very little space compared to the other cable components, such as strength members, sheathing layers, electrical power wires, ... etc.

    So, practically, what you suggest is not feasible, but theoretically what you suggest is completely reasonable. Perhaps someday the technology will develop to make this feasible.

    This reminds me of another idea that was once considered for increasing usable bandwidth. People have tried to use coherent detection to improve receiver sensitivity and to allow ultra-tight frequency division multiplexing. This is basically trying to use a laser, much the same way we use RF signals. The light becomes a carrier that is modulated with high fidelity using all of the same RF techniques (frequency modulation and phase modulation for example). Heterodyne and homodyne detection can then be done. But, even though it is possible, and very impressive demonstrations have been done, it is not practical compared to high speed channels using wavelength division multiplexing on single mode fiber.
     
    Last edited: Oct 12, 2014
    • Like Like x 1
  8. steveB

    steveB Well-Known Member Most Helpful Member

    Joined:
    Jan 16, 2009
    Messages:
    1,297
    Likes:
    629
    I should mention that there is another aspect of using spatial modes that would mitigate its benefits, even if it could be made practical. Once modal dispersion is removed by using only one mode for the channel, that channel has bandwidth limits set by the chromatic dispersion. Chromatic dispersion has two components: material dispersion and waveguide dispersion. The material dispersion is due to the glass itself, and the waveguide dispersion is due to the waveguide structure (i.e. the fiber geometry). The wavelength of zero material dispersion is 1310 nm, which is why that wavelength is popular. The 1550 nm wavelength is popular because that is where the fiber loss is the lowest. One trick is to design fibers such that the waveguide dispersion moves the zero dispersion wavelength to 1500 nm, which provides the benefit of zero dispersion and lowest loss. Then, people can even design dispersion flattened fibers that have near zero dispersion over a broad wavelength band near 1550 nm. In this way, wavelength division multiplexing can have super high speed channels, and many channel with very low attenuation for all channels.

    Now, if you try to use spatial modes, each mode has different wave-guide dispersion. Hence, it would be all but impossible to make zero dispersion at 1550 nm over a broad range, for each and every spatial mode. Hence, special dispersion compensators would be needed to correct dispersion in each spatial mode, which is another technical difficulty, but still not impossible in principle.
     
    • Informative Informative x 1
  9. neptune

    neptune Member

    Joined:
    Jun 14, 2009
    Messages:
    466
    Likes:
    5
    Location:
    New delhi, India
    both modal dispersion and chromatic dispersion are function of length of the fiber, if we reduce the fiber length to 5 meter then it wont be a factor, that fiber will work as fast link for copying one memory to another.
    So in short distances (5 m) WDM can also be used in multimode fibers.
    "The wavelength of zero material dispersion is 1310 nm, which is why that wavelength is popular. The 1550 nm wavelength is popular because that is where the fiber loss is the lowest."
    if only single frequency is used then how do they achieve WDM ?
     
  10. steveB

    steveB Well-Known Member Most Helpful Member

    Joined:
    Jan 16, 2009
    Messages:
    1,297
    Likes:
    629
    It's not just length that is a factor, but bitrate is a factor too. People like to talk about the BL product, which is bitrate times length. Dispersion limits the BL product. So, yes, if you keep the length short, you can operate at high speed and still be OK. But, you have to calculate to see what the actual limit is, and whether you are violating that limit.

    They, dont use WDM with that method. This idea is to have one channel at super high speed. Speeds over 40 GHz have been shown. If you try to use other channels that are off the zero dispersion wavelength, then speed is limited. This is why dispersion flattened fibers are of great interest for WDM long distance systems.
     
    • Like Like x 1
  11. neptune

    neptune Member

    Joined:
    Jun 14, 2009
    Messages:
    466
    Likes:
    5
    Location:
    New delhi, India
    40 GHz without using WDM ...how ?
    "If you try to use other channels that are off the zero dispersion wavelength, then speed is limited"
    if it is zero dispersion at that wavelength then why is speed limited ?
     
  12. steveB

    steveB Well-Known Member Most Helpful Member

    Joined:
    Jan 16, 2009
    Messages:
    1,297
    Likes:
    629
    Actually I don't know the details of how they get 40 GHz. It's amazing, isn't it! There are various research papers on this, which you can read if you are interested.

    I said that if you are OFF the zero dispersion wavelegth, speed is limited. If you are ON the zero dispersion wavelength then the limit is much higher.
     
  13. neptune

    neptune Member

    Joined:
    Jun 14, 2009
    Messages:
    466
    Likes:
    5
    Location:
    New delhi, India
    ok nice, another question what is the maximum limit of modulation for both LED and Laser OR what's the speed in Hz by which these things transmit data.
    and which one is the fastest modulation technique used in optical fiber network
     
  14. steveB

    steveB Well-Known Member Most Helpful Member

    Joined:
    Jan 16, 2009
    Messages:
    1,297
    Likes:
    629
    I can't answer that with certainty because I stopped working in that field over 10 years ago. Technology changes and memories fade. You simple need to look at the available devices on the market to find this out.

    However, it is helpful to know a few basic things when you do the research. Typically, LEDs and ELEDs etc. have the lowest modulation rate, exceeding 100 MHz. Then, diode lasers of various types (particularly the DFB laser) can be directly modulated at substantial speeds (perhaps exceeding 1 GHz). Then there are external modulators, such as lithium niobate based devices which can go well beyond 1 GHz. Then, I imagine there are various permutations of these devices and new inventions that might surprise us. For example, I saw an article for a plasmonic LED that does many GHz modulation rates. I don't even know what a plasmonic LED is.
     
    • Informative Informative x 1
  15. neptune

    neptune Member

    Joined:
    Jun 14, 2009
    Messages:
    466
    Likes:
    5
    Location:
    New delhi, India
    alrightie...
    how do I give you a like ?
     
    Last edited: Oct 14, 2014

Share This Page