Modeling of multi-mode fiber coupled Receivers and De-Multiplexers in OptiSystem

[Ahmad Mustafa]

Hello all,

I am simulating a DWDM system in OptiSystem and I want to have a multi-mode fiber coupled De-Multiplexer and a MMF Receiver. However, OptiSystem does not have such MM De-MUX and Rx. Can I simulate/model the multi-mode environment in such case?

Regards,
Ahmad

[Tech Support]

As you noticed, we don’t have a multimode demultiplexer in Optisystem, we will investigate setting up this type of capability but we have not determined the target release.

Are you looking for an idealized model to separate out the modes or a physically based MM Demux?

For the idealized model, one possibility would be to perform some Matlab Cosimulation:

We have a component called “Mode Selector” in our “Receivers Library/Multimode” directory from which you could extract out each transverse mode. Knowing each mode profile, you could then process this with Matlab to perhaps convert this to a non-multimode optical signal and then you could use our standard demuxes.

If you are looking for a physically based model, our FDTD product manager mentioned that there are some designs that could be simulated using OptiFTDDTD. For example: http://www.opticsinfobase.org/oe/abstract.cfm?uri=oe-21-15-17904

[Ahmad Mustafa]

Hi,

I can briefly explain you my application as follows:

I have a single mode transmitter with 2x40Gbps single mode Multiplexer. The two transmitters are centered at the same frequency i.e. 191.1THz but are orthogonal to each other in terms of polarization. The 80Gbps Polarization Division Multiplexed signal is transmitted through Free-Space over a distance of xkm and is received at a receiver which has a Multimode De-MUX. Here, is where the received signal experiences multimode effects. I have introduced a Linear Multimode Fiber block to add multimode effects before the signal is passed on to the WDM Demux. The two signals are separated on the first arm of the Demux and are to be converted into electrical signals by a MM optical receiver and later BER and Eye Diagram are observed on the Visualizer.

Now, I do not know what do you mean by an idealized model to separate out the modes and by a physically based MM Demux as my knowledge on working with multimode components and fibers is limited. However, I have attached my OptiSystem file for your consideration. I see an error when OptiSystem calculates Spatial Optical Receiver block as the input signal is invalid because it does not have spatial transverse modes(rightly so). I can introduce spatial transverse modes by using Spatial Optical Transmitter but in the laboratory experiment, I have a single mode transmitter so I want to avoid this solution to keep my simulation as close to the lab experiment as possible. What can you suggest in this scenario?

Feel free to ask me if you need more clarification about the setup.

Cheers,
Ahmad

Attachments:

1. Pol-Div-Multimode.rar

[Damian Marek]

Thanks for the detailed description! There are two changes I suggest.

Firstly, I would use the Spatial CW Laser. You can set the output of the transmitter to single mode operation, in fact I believe it is set by default. I think this would be a closer match to the lab experiment. It has all of the details that come with the CW Laser Source with the added benefit of supporting spatial modes.

You should also switch the Linear Multimode Fiber for the Measured-Index Multimode Fiber. The linear multimode fiber does not have an effect on the spatial profile. It is more concerned with applying modal delay, modal dispersion etc., but then again it is much faster for simulating! In the Measured-Index model you can quickly define a simple multimode fiber by core radius, cladding thickness and their refractive indices. As the name suggests, you can also import a refractive index profile for more complicated fiber designs.

Let me know if this helps!

[Ahmad Mustafa]

Thanks Damian. I will work on your suggestions and will get back to you later.

[Ahmad Mustafa]

Hi Damian and Tech Support,

I worked on your suggestions and now have better understanding of how multimode components work in OptiSystem. I have come up with some more questions though. Attached is my .osd simulation file for understanding the questions better.

I have a single channel transceiver where the Spatial CW Laser centered at 193.1 THZ has only one mode as can be seen from the single Value of Power Ratio Array and Pol. X m,n index array to represent the single mode laser that I have in the lab. It can be seen on the Spatial Visualizer where I only see one mode. The length and Attenuation of Parabolic-Index Multimode Fiber are set to 0.4m and 0.61dB/km respectively according to my lab setup. The Min. Signal Power is set to -19dBm. Finally, the signal is received in the Spatial Optical Receiver and I get perfect Eye Diagram and 0 BER on the BER Analyzer. Following are my queries that I want to understand:

1) The spatial CW Laser generates only one mode but I see multiple modes on Spatial Visualizer after Parabolic-Index Multimode Fiber. Should I not see only the fundamental mode as the length of the multimode fiber and the attenuation value is really low to generate multiple modes in the fiber?

2) How can I measure the power of the individual mode after the MM fiber? I assume the power in the modes besides the fundamental mode is really less that they can be ignored in this case.

3) The spot size after MM fiber is bigger than the spot size right after Spatial CW laser. I believe it is because of Modal dispersion. Yes?

4) I have increased the Min. Signal Power to -19dBm from the default value which is -100dBm to reduce the number of modes and I can further reduce the number of modes by further increasing this min. signal power value but in my lab, I cannot simply do this because of the hardware limits. Can you make any suggestion how to reduce the number of unwanted modes coming out of the Parabolic-Index Multimode fiber because by increasing the power of the input laser and length of the MM fiber along with the higher attenuation value, the generation of higher number of modes and with that mode coupling effects would make the BER and Eye Diagram go bad as per my understanding.

Please let me know if anything is not clear regarding the questions I posed.

Many thanks

Attachments:

1. Single_Channel_Multimode-Transceiver.rar

[Damian Marek]

Hi Ahmad,

These are great questions!

1) You will see multiple modes excited at the output of the fiber, because of mode coupling at the input of the multimode fiber. You can check the Help on the Parabolic-Index Multimode Fiber to see how this coupling is modeled. The input to the fiber is a single mode, but obviously it is not exactly the same as the fundamental mode of the fiber. This mismatch between the modes can cause higher order modes to be excited along with the fundamental fiber mode. This is an interesting topic and if you want to learn more I would suggest looking at this link:

http://www.rp-photonics.com/mode_matching.html

2) The power of higher order modes will usually be less than the fundamental mode, but this really depends on how your input mode(s) compare to the modes supported by a fiber. To determine the power you can use the Mode Selector component and a power meter. I already made a project file, that could be easily adapted for this application, in a different forum post:

https://optiwave.com/forums/topic/co-simulation-in-optisystem/

3) The spot size at the output is simply due to the fact that the fundamental mode of the fiber has a larger spot size than the mode supported by the Spatial CW Laser. Since these modes are propagating in a guided medium, theoretically there would be no beam waist expanding.

4) The best way to try and insure single mode behavior in a multimode fiber, is to match the incident laser mode to the fundamental mode of the fiber. Depending on the laser you are using in the lab, you could use a lens to focus the light into a smaller area or expand the beam into a larger area. OptiSystem has components to model this behavior. Of course, in a real optical fiber, bends and imperfections will, over longer distances, couple power away from the fundamental mode and into higher order modes.

Hope this helps!

[Thang Hoang]

Thanks guys, I am working on the same problem.

[Ahmad Mustafa]

Hi Damian,

These are great answers 🙂 They have helped alot. I have few more questions if you could help me understanding them:

1) How to know the fiber modes? I can select the number of fiber modes from the parameter LP(m,n) max but how to see the fundamental mode of the fiber itself? What I can see on the Spatial Visualizer is the result of mode mixing between the input modes and the fiber modes. Is there a way to see the fiber modes only?

2) As you say in the reply of point Nr. 3 that the spot size at the output is increased because of the larger spot size of the fundamental mode so how can I change the size of the fundamental mode of fiber? I do not see any parameter to do that. Does it depend on fiber core size?

3) you wrote in reply to the mode matching: “you could use a lens to focus the light into a smaller area or expand the beam into a larger area. OptiSystem has components to model this behavior”. Are you referring to Thin, Vortex lens and Spatial Aperture in OptiSystem?

Many thanks

[Damian Marek]

You’re welcome!

1) I’m not exactly sure what your question is here. You want to see which modes are supported in an optical fiber? The modes at the output of the fiber actually only consist of the fiber modes. The input mode spatial data is used to determine the different power levels (coupling coefficients) attributed to each fiber mode. If you are interested in solving for the different modes of a radially symmetric fiber, I would really suggest looking at OptiFiber. It is an extremely easy tool to use and will calculate the fiber modes, as well as give pertinent information. It can also be used with OptiSystem.

2) The spatial distribution of the fundamental mode (or any mode) is a complicated problem, which can sometimes be solved analytically. It depends on the refractive index distribution of your waveguide and is usually calculated by a mode solver. But yes, generally the spot size is determined (if the fiber is a step-index fiber) by the core radius.

3) Exactly!

Hope this helps!

[Author]

Posts