Error during simulation

[chee]

i am designing a system of coherent detection using 16-QAM OFDM. my design is based on the example provided by Optiwave. the differences between my system and the example are the parameter settings of OFDM and using FSi instead of fiber. there is an error in the OFDM component every time i try to simulate the said system. i was unable to identify what is wrong. Kindly hope somebody can help me here. thanks in advanced.

Attachments:

1. Project15.osd

[MHD NA]

Hi Chee; I think you didn’t choose the components variables correctly. Try with another values for each component in diagram.

[alistu]

Hi Chee,

The only thing that I particularly found being set wrong was “the number of subcarriers per port”, which was not the same as the number of subcarriers in the “subcarriers location”. Even though I changed it, the error was still there. I increased your sequence length(to 4096), and now I am having the error for OFDM demodulation component!

[chee]

yes alistu, i have the same problem too. but i noticed that if i follow the exact value from the example, then it have no problem. i wonder what is the problem. here is my system after i try to change it back to my project’s parameter. my trial license left 17 days only. please help me.

Attachments:

1. Project1.osd

[alistu]

I am also very curious about this problem, but unfortunately I haven’t managed to solve it. I think the problem has to do with the number of subcarriers, though. When the maximum number of subcarriers is reduced, the simulation is run smoothly. So this might be a technical problem.

[chee]

may i know what is the parameter that affecting the BER. i tried increasing the optical amplifier gain but to no avail, the BER rate stil in the range of 0.5. is there any component that can be used to dramatically increase the BER to the acceptable range?
i noticed that in the example file OFDM Coherent Detection Single Port Single Polarization 16QAM, the BER is 0.00779. is this value acceptable?

[alistu]

Depending on their noise figure, optical amplifiers add noise to the signal also, so the excessive use of them might not help further from just compensating for the attenuation caused by the line. Is it possible that you upload your simulation file so that we would be able to talk on a specific case? Otherwise there are many reasons so as to why this might happen, such as the line being too long or the bit rate being to much, or incorrect system settings.

[chee]

sure thing. because i am using FSO model in Southeast Asia region, the rain rate is very high, which is why the attenuation is in the range of 50-70db\km. i would like to improve the BER to the acceptable range. if i am not mistaken. BER should be at least 10^-9 right?

Attachments:

1. Project2.osd

[alistu]

You’re right about the acceptable BER range (10^-9 or 10^-12 sometimes). I don’t know why even a back-to-back system in your implementation has 0.5 error, but I used your FSO channel in Optisystem sample (as attached) and then by using amplifiers in between the 1 km distance, the way you can see in my implementation), you can get better results. I did it this way because of the high attenuation of your line.

Attachments:

1. project2-modified.osd

[Damian Marek]

Hi Chee,

I found that your MZMs were not biased properly at the null point. Biasing both of them at 2 and -2 for each arm results in a more linear transfer function for small electrical signals. In addition increase the samples per bit to at least 16 samples per bit.

There is still some error (and I disabled the FSO component for now) because the extinction ratio of the MZMs shifts the constellation diagram a bit. I do not know what MZMs you are limited to using but a larger extinction ratio is better. Also properly choosing the gain is important to balance the symbol separation and the distortion of the transfer function.

Try out the example attached.

Attachments:

1. Project2-Reuploaded.osd

[alistu]

Hi Damian,

I cannot open your attached file. When I try to open it, the error “Failed to load document from storage.” is given by the software. I am using the version 13.0.3 though and I can open other osd files. Is there something wrong with it? Thank you.

[Damian Marek]

It should work fine. Try downloading the file and then opening it from OptiSystem instead of double clicking the actual file.

[chee]

Dear Damian,i am too unable to open the file as i am using trial version. my project does not have any limitation regarding MZM modulator. what is the range of extinction ratio that i can set? is there going to be other effect on the signal after i alter the extinction ratio?

[alistu]

Thank you. It seems I am still having the problem, though. The exact same error is being given when I try to open it by the software itself.

[Damian Marek]

Hey guys,

I changed the attachment please try again! I was using an internal version of OptiSystem.

[alistu]

I managed to open the new file, thank you. I lowered the bit rate of this system to 10Gbps and now I am facing a BER of around 0.4, so there still seems to be problem (I didn’t get the part you mentioned gain, so I don’t know if I have to manipulate gain). I had not worked with the new OFDM components. There is another problem addressed by Chee in another forum, which occurs when he set maximum number of subcarriers to 1024. Could you please make a comment on that. Thank you again.

[chee]

Dear alistu, by only changing the extinction ratio of the mzm to 50dB, with the FSO deactivated, i managed to get 0 BER. the samples per bit is already set at 32.

[Damian Marek]

When you change the bit rate and the number of subcarriers, you need to change the bit rate in the BER Test Set to reflect this. Look at the examples in the technical background of the OFDM Modulations. The bit sequence generators (e.g. BER Test Set) need to take the Bit rate value of Global bit rate * number of sub carriers / maximum possible subcarriers. This will correctly “fill” exactly all of the carriers with bits before the OFDM Modulation creates the OFDM symbol and transmits it.

Regards

[chee]

Dear Damian, i have made the changes on the BER test set accordingly, but still the error persists. attached is my osd file. thanks.

Attachments:

1. Project.osd

[Damian Marek]

Chee,

Start from the project file I uploaded previously. In your attached file the bias point is the original incorrect value, the samples per bit is also too low. If you get the error at the OFDM Demodolation component set the sequence length to a longer value. Try the attached.

Alistu,

You are correct about the bit rate updating automatically, but I imagine you forgot to change the Global Symbol rate to the new value of 2.5 Gbps. This has an effect on the filters and the OFDM Demodulation.

Attachments:

1. Project-Chee.osd

[alistu]

You’re right! I had set it to 5 instead of 2.5. Thank you.

[alistu]

Thank you Chee for the remark. And thank you Damian for your notification on changing subcarriers in BER test set. That was a careless mistake I made. But when I set Bit rate as 10Gbps in your implementation without changing the number of subcarriers or anything else, I get BER=0.4 and since the Bit rate is calculated in script mode in BER Test Set, there’s no need to change it manually. So I just changed the bit rate, I don’t know what else to change.

[MHD NA]

Hi Chee;
There are many factors affect on BER and Q-factor also, such as the power of transmitted signal, SNR, attenuation, dispersion in the fiber. You should control with those factors to adjust BER value which must be at least e-09 and the corresponding Q-factor equals 6.
Try that and Good Luck

[alistu]

Hi Mohammad, Thank you for the numerating the important factors. Apparently, non of these factors now improve the performance of the system implemented by Chee. Please, have a look at my reply on the new topic Chee has started and state your opinion about that if you have the time: optiwave.com/forums/topic/low-ber-in-coherent-ofdm-simulation/ . Thank you.

[Author]

Posts