- This topic has 25 replies, 6 voices, and was last updated 8 years, 9 months ago by alistu.
-
AuthorPosts
-
-
June 13, 2015 at 4:39 pm #21670alistuParticipant
Hi all,
In an OFDM system, what is the reason for symbol rotation which can be seen in the attached file named image1? And what kind of equalization do you suggest to overcome this problem?
P.S. In my simulations of WDM-OFDM systems, the rotation degree is different for different channels.Thank you in advance.
-
June 14, 2015 at 9:58 am #21686Ashu vermaParticipant
Hello Alisthu
Can you please tell,hoq many WDM channels you are using?You can do one thing set the LO(Local oscillator) phase to keep the symbols at the right position.For long distance communication you can use matlab to compensate the effects of non-linearities .Optical phase conjugation is best way to do this. -
June 14, 2015 at 1:15 pm #21692alistuParticipant
Hi Sam Sung,
I have done the simulation with 16 and 32 WDM channels. Changing LO phase was the first thing I tried for the single user OFDM, but in order to find the exact angle sometimes you need to repeat the simulation process for a couple of times. Besides it’s not sensible to do such a thing for every channel of a WDM system. I will consider Optical Phase Conjugation. Meanwhile, I will be grateful to hear about other compensation methods.
Thanks for your reply. -
June 14, 2015 at 11:45 pm #21695RavilParticipant
Hi Alistu,
According to your question about the compensation methods, I agree with Sam Sung’s suggestion of using matlab compensation: several types of equalizers are available in DSP. Please, see the following link: https://optiwave.com/applications/100-gbps-dp-qpsk-system-with-digital-signal-processing/. The main compensation methods for given system and fiber medium are coherent detection and digital back-propagation. See the following links about it:
http://www.ncbi.nlm.nih.gov/pubmed/21935052; http://ee.stanford.edu/~jmk/pubs/digital.backpropagation.jlt.10-08.pdf -
June 14, 2015 at 11:49 pm #21696RavilParticipant
According to matlab usage, DSP Components are designed such a way that you can either use build-in algorithms or insert any any others (i.e., your own) at each stage of the compensation process. However, the restriction is that you should use downsampling from optiSystem. As an example, you can take a look at the section “Inserting your own MATLAB algorithm” in the help documentation of the “DSP for 16QAM”.
-
June 15, 2015 at 2:19 am #21706alistuParticipant
Hi Ravil,
Thanks for your reply and useful links. I would also like to know you opinion about the reason for the symbol rotation and why it is different in different OFDM channels. Others seem to think it’s because of nonlinearities. And I’m looking for some type of equalization that has the same effect on my system as changing the initial phase of LO laser does, which is preventing symbol rotation and not necessarily giving a more clear constellation diagram.-
June 15, 2015 at 9:35 am #21712Damian MarekParticipant
The rotation of a constellation is due to a change in phase of all symbols. This could mean propagating down the length of a fiber or due to dispersion. The smearing of the individual symbols is also a phase effect, which without looking at your design I would assume is from dispersion as well or possible SPM. The CW Laser also introduced phase error from its linewidth.
The OFDM demodulator component is able to efficiently correct this issue with carrier phase estimation from the pilot tones, but the older OptiSystem 12 components do not have this functionality.
Regards
-
-
June 15, 2015 at 11:31 am #21717alistuParticipant
Thank you Damian. Since I’m using Optisystem 12, what do you suggest that I do in order to correct it? Should I first study about pilot tones and then create some equalizer with the same functionality?
-
June 15, 2015 at 2:16 pm #21720alistuParticipant
Still I can’t understand why in a WDM-OFDM system rotation in different channels is different while the launched power in each channel is the same as the others and all channels pass through the same line. Anyone has any ideas about this?
-
June 15, 2015 at 3:44 pm #21731Damian MarekParticipant
The difficulty in changing the delay of the OFDM Demodulator is one of the reasons that led us to design the new OFDM components. Have you tried changing the delay in the OFDM Demodulator?
It would make sense if the different channels have different delays as they experience different dispersion levels. Unless your fiber is set to constant dispersion for all wavelengths.
Regards
-
August 18, 2015 at 4:56 am #23787Ashu vermaParticipant
Hello chee,interesting question.you can observe it by your own.In fso channel i dont think there is any option of dispersion however propagation delay tab is there.As far as i know you need to change the delay inside ofdm demodulator to get good results and constellation .But dispersion is not there in fso channel i think so
-
August 18, 2015 at 5:11 am #23790cheeParticipant
Hi Sam Sung, i recovered the constellation diagram by changing the phase in the laser. does this mean dispersion do not affect FSO channel? the rotation of constellation is caused by delay?
also, my aim is to used the simulation BER results to compare with practical FSO model. i wonder if the result can still be trusted now that i have change the setting in the laser.-
August 18, 2015 at 5:15 am #23791cheeParticipant
i tried to change the delay in OFDM demodulator but did not have any effect on the constellation diagram.
-
August 18, 2015 at 6:29 am #23797alistuParticipant
Hi Chee, I think the results obtained by changing laser phase can still be trusted as long as you report this as a type of equalization. This is of course very basic, compared to some other equalization types used in OFDM. And symbol rotation can be avoided by using matlab component and a program that would calculate the rotation and rotate the symbols back, so you wouldn’t have to change laser phase each time.
-
February 24, 2016 at 4:01 pm #30136Mohamed Ben zegalamParticipant
Hi Alistu,
Please may I know what do you mean about “using matlab component and a program that would calculate the rotation and rotate the symbols back”
As far as I understand from this discussion that there is no such away to calculate the rotation. if I am wrong, what is the formula to calculate the rotation?, and if there is any paper discuss the rotation symbol in optical communication system please tell me.thanks
-
February 24, 2016 at 4:37 pm #30137alistuParticipant
Hi Mohamed,
The method used for this purpose is channel estimation using pilot signals. And Matlab can be used for this purpose for sure. Surely this can be helpful if you are interested to combat this problem in OptiSystem 12. However, in OptiSystem 13 OFDM components, the capibility of using pilot signals is available.
Regards
-
February 24, 2016 at 4:46 pm #30138Mohamed Ben zegalamParticipant
Hi Alistu,
thanks for your reply. actually I am trying to make OFDM 12 work with phase compensate instead of pilot and channel estimation. do you know any paper study the phase shift in the optical transmission line
-
February 24, 2016 at 7:48 pm #30139alistuParticipant
My pleasure. Actually, I was looking for solutions of that type while I came across the channel estimation methods. I cannot think of any phase rotation compensation method that does not make use of some type of channel estimation.
I myself came up with some method using input data information for phase rotation compensation (where the channel estimation is not used), but my supervisor thought it was not practical since the input info is not available at receiver. Anyway, I only wanted to replace it with manually having to change symbol rotations.
Regards
-
-
-
August 18, 2015 at 5:22 am #23792Ashu vermaParticipant
Attach your system here so that it will easy to check .Laser phase effect your consellation but every time you can change the phase at different distance or change due to power.Optisystem v 12 has this delay adjustment issue.I think there is no specific way to calculate tge delay which you can pot at denodulator for different channels.Thing you can do is ,try to change the delay vakus by using iteration.
-
August 19, 2015 at 3:33 am #23828Ashu vermaParticipant
see the file
-
August 19, 2015 at 3:51 am #23830cheeParticipant
sorry, i cant open the file, i am using ver 12
-
August 19, 2015 at 4:40 am #23834alistuParticipant
Hi Chee, If you wish to install version 13, be careful every file you open and save just once with the new system, you will not be able to open it again with the oldest versions. So I suggest creating a backup of your files if you consider doing this. And there is sometimes dispersion in FSO, depending on the channel(eg in atmosphere).
-
-
-
August 19, 2015 at 3:55 am #23831Ashu vermaParticipant
Oops,this file from optisystem 13.you cant open it.Have you tried trial version? Which is free of cost?you can download it and use fir 30 days.If you have that download already then you can use the file i have attached.However i have just modified the sample from optisystem v11 file in v13 using fso and it works fine and covers almost 2-3 Km .Also work for 100Gbps and more with error free communication.
-
October 16, 2015 at 10:27 am #26666alistuParticipant
Hi all,
I was trying to read the technical background on OFDM component, but some parts are illegible as the font is really small. And zooming using m=windows magnifier doesn’t help. Can anyone please tell me whether I can get these explanations from somewhere else, for example from the Optiwave website? Thank you.
-
-
AuthorPosts
- You must be logged in to reply to this topic.