Sac ocdma with MD code

I believe the spacing between the frequencies depends on the overall bandwidth you want to use. The less spacing between the wavelengths, the better (as long as you get the desired bit-error-rate). So why don’t you simulate the system for different frequency spacings and then choose the most suitable one?

As for what you have asked about the type of light source to use, the coherent light from laser would definitely be the best choice. However, because SAC-OCDMA systems are used in some applications such as Fiber-to-the-Home as a low-cost solution, sometimes LEDs are used as a cheaper light source along with FEC schemes to reduce the costs. The paper “A Comparison of Optical Sources for Spectral
Amplitude Coding OCDMA” which could be found easily by google makes comparisons between the two, but I guess the above explanation would suffice.

Basic difference of using the LED rather than laser is only cost of the system while laser wikl provide you better results due to more directinality,coherence and monochomatic nature.However LED along with multimode fibers are preferably used to decrease cost of the system.

As I mentioned before, if I were you I would have tried different bandwidth values to find the optimum one, which would be the one to give good system performance (with desired bit-error-rate) and yet have the least amount possible. I cannot think of any number right now, but I would do as what I said to find the proper value.

You’re welcome Gurpreet. On the receiver side, the signal first passes through the same FBGs as it has passed through on the transmitter side before being launched into the channel, and then through the photodetector. And the number of FBG equals the number of wavelengths of the number of weights used. So if W=3, then you should use three FBGs. I suggest that you take a look at optisystem SAC OCDMA sample.

I think you have to attach the FBGs in series the way shown in the attached Optisystem file. I have not seen otherwise in the papers but that doesn’t mean such a way would be wrong. The thing is, I didn’t get exactly the other way you are suggesting than the series attachment, but using FBGs in series works for sure.

You’re welcome. I recommend that you search into review papers to get more accurate answers for questions like this. I haven’t worked in free-space-optics field, but I know one thing for sure: The reach depends on the atmospheric situation of the channel. For example, in free space you can have longer links with the desied reliability than other FSO channels. 2Km link is enough, if not too much. You might want to increase the number of the users instead.

Since you are working in free-space-optic field, your choice of weather condition is related to the application of the system you are simulating. Is it used in inter-building communication or inter-satellite communication? I suggest that instead of increasing the distance, try to increase the number of users or the transmission bit rate because 2km is enough for many purposes, unless you know exactly about the application of your work.

You’re welcome Gurpreet. Glad to be of any help.

It seems to be working OK, but may I ask why you haven’t used fiber bragg gratings on the transmission side as well as on the reception side of the FSO channel? As I pointed out, lasers are considered expensive components and this way you are using several of them. By doing as depicted in optisystem sample osd file, You only need to use one laser for each user on the transmission side.

I saw them on the reception side. I think you can easily do it yourself. Change the design so that on the transmitter, the same FBGs are used (after the laser and before launching the signal into the line) as in the receiver of each user. Please refer to the SAC OCDMA sample osd file that I have mentioned several times for more clarification.

I haven’t worked with SAC OCDMA (simulation work) but I think in case of MD codes, you just implement MD codes (instead of Hadamard codes) with FBGs and that should be it. There is a paper named “Development of a new code family based on SAC-OCDMA system with large
cardinality for OCDMA network” that I refer you to, which has not used FBG I think.

In that paper, the number of users is 5 and the code weight is 4. Having these numbers in mind, try to simulate the scheme depicted in figure 6 of the paper. Try reading the simulation part of the paper before implementing it. here’s another picture from another paper, with W=4 and N=2 which is close to your own simulation. FBGs are uniform. Take a look at this and its paper “Simulation of a SAC-OCDMA 10 User ×15 Gb/s System Using MD Code”, which expalains the same stuff again. try to make something out of these two.

You’re welcome!
Yes, so have you in your own simulation whose osd file you have attached in this forum. However, I suggested reading the two papers just to understand how the system is supposed to work in order to learn how to implement your own. Try to find answers to these questions in these papers.

I never told you to go for LED. If you disagree, please give me the number of the reply in which I have stated such a suggestion. I think they have used demultiplexer to distinguish between different wavelengths. But I am not sure what is exactly going on. Right now I am busy but I will be back in a few hours with an answer hopefully.

I strongly suggest that you go through the whole paper, and then try to implement your own system in Optisystem like the one simulated in the paper. If FBGs are used in the paper, use them in the same way. Just the pictures and screenshots from the paper are not enough. You’ll need to read the text as well.

You’re welcome. Is your problem hopefully solved now?

You can generate signals with arbitrary bandwidth in the base band. For example, you generate a baseband OFDM signal using the OFDM modulator with a bandwidth of 20nm. Then by using the laser you will be able to upconvert it and change its center frequency to the laser frequency. When this happens, the spectrum has 20nm bandwidth in the laser frequency.

Hello gurpreet
If you want to broad the spectrum of laser signals you can do the following things
1) use hnlf highly non linear fiber to broad spectrum over a freq. Band.using spm and xpm effects
2) you can use SOA as a non linear medium to broad the spectrum and distribute signal over different dreq. Signals
3) use EAM with the carving high freq. To upconvert or downconvert your signal
I think these suggedtion help you.if you need help to implement these systems,i can help you

I think choosing between power splitters/combiners and multiplexers/demultiplexers depends on your system. A multiplxer adds signals and a demultiplxer divides the signals again. But in a power combiner, the signals are added so that in each branch of the splitter, we have an attenuated version of the whole signal. In a MUX there is no attenuation, unless you change the value of insertion loss from zero.

Hi Sam Sung, Yes, I am well aware of spread spectrum and CDMA technics and so on. But in those cases, the signal is spread pseudo-randomly and not completely randomly. For example in CDMA, each user’s signal is multiplied by a code which is known at the corresponding receiver. But I didn’t know about spreading the spectrum using XPM.

Hi Gurpreet, Here are some useful links that I have found, which talk about the basic concepts:

http://www.optics.rochester.edu/workgroups/agrawal/publications/papers/paper_1989_10.pdf
http://core.ac.uk/download/pdf/11533984.pdf

And this one is more practical than theoretical:
http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6813584

I suggest that you at least take a look at the last one. I hope these help.

Thank you very much Sam Sung for the useful links. However, I think we have somehow deviated from the main problem Gurpreet stated. I think it would help him and me much more than a hundred links if you would very kindly make an osd file, demonstrating how to get multiple wavelengths from the same laser. Because you have read the papers you have introduced and sure know how to do that more than we do. I also have Gurpreet’s problem and it would be a real help if you do this for us. Thank you in advance.

In that example, another parameter is defined in the global parameters. that parameter is named “SignalBitrate” and this is the parameter that the bitrate of each user in the PRBS is referred to. So as far as I can tell, the global bitrate parameter is in practice replaced by this parameter in this example.

Thank you very much Sam Sung for the explanations and the file attached. I think this is some of what me and Gurpreet were curious to know about the issue, aside from the papers and all the theoretical aspect of it. And we are waiting for more explanations plus the attached osd file after you find the time.

Hi Gurpreet, Please attach the osd file if possible. Thank you.