RoF with QPSK doesn’t work

[Maciej Nowacki]

Hi,

I have made RoF with QPSK modulation but it doesn’t work. Could you help me ? OscilloscopeVisualizer_2 doesn’t show anything but Electrical Constellation Visualizer_1 gives nice results.

Best Regards

[Abhishek Shrama]

Hello maicej
Which version of optisystem you are using? It is better to use new version of optisystem beceause there is DSP component in new Qpsk system ,which is useful for signal non linearity and dispersion compensation.Here i have given a link for better understanding of QPSK DSP component system.Also some discussion on this topic already done,i am also giving links of forum topics

100 Gbps DP-QPSK System with Digital Signal Processing

DSP

PDM QPSK transmitter and receiver

[Maciej Nowacki]

Hello
I’m using Optiwave 13 but I’m doing the simplest version of ROF with QPSK. It’s only for seminar at university.
Could you help me how to figure out problems with my example ?

[Damian Marek]

Hello Maciej,

There were a couple small problems with your project.

1) The low pass filter for your OFDM Demod was not set correctly, so I changed the cutoff frequency to the symbol rate.
2) The symbol rate was also incorrect for your simulation so I changed it to Bit rate/2
3) The decision component should be set to a threshold of only 0 and values of -1 and 1. After changing these options I got good results.

Attached is the augmented project. Cheers!

[Maciej Nowacki]

Hello,

Thanks Damian! You have big beer on me. But I have one more question. I have been trying to measure BER but my system doesn’t work. I got very wird results. Could you help me ? I have been trying change place of BER analyzer to everywhere but it doesn’t work.

Cheers

[Damian Marek]

Using higher modulation formats it is not advised to use the BER Analyzer to calculate the BER. Your main options are to use the BER Test Set to test a long sequence of bits through your system, or to use the constellation visualization, which can give you an error vector magnitude and symbol error rate.

Regards

[Alessandro Festa]

Hi Damian, just one curiosity: so you suggest to use BER Analyzer for simple modulation formats (like standard PAM NRZ, RZ) and BER Test tests for higher modulation formats (QAM, PSK…)?

[Damian Marek]

Yes! The BER Analyzer calculates the BER from the generated eye diagram of the system. Higher modulation formats like QPSK and 64 QAM do not have traditional eye diagrams because the amplitude and phase are being modulated. In this case the analog would be using the Constellation Diagrams to calculate the error vector magnitude and symbol error rate.

The BER Test Set can still be used to find the straight BER though since it does a simple comparison between input and output.

Regards

[Alessandro Festa]

Thanks Damian for clarifing this point!

[Ajay Vyas]

Yes Damaian you right.
Thanks to correct our error. I have same mistake in my project . Is there any relation between the constellation diagram and eye diagram.
It is necessary if we get the good constellation diagram it mean that the eye diagram should be clear?
Please clear it.

[Maciej Nowacki]

Hello

Do you have any samples using BER test set(analog system) ? I have used BER test in my project but It doesn’t work. Could you look into my project ?

[Damian Marek]

If you check the OFDM samples given along with OptiSystem they should all have BER Test Sets. Just look in the Advanced Modulation folder of examples.

[Maciej Nowacki]

My BER test set still doesn’t work. Could you give me a hint where the problem is ? Whole system is working but BER doesn’t show anything.

[Maciej Nowacki]

I have made it but my system works up to 1,2 km. Am I right that it is too short ? Dispersion destroys my signal very much.

[Alessandro Festa]

Hi Maciej,

first of all, your sequence length is too short in my opinion (32bit only, I would increase to 1024 for more reliable BER results). If you increase the sequence length, you will notice that BER is not good even at shorter length…

[Maciej Nowacki]

Do you know where should I look for problem ? Previously I had BER analyzer but I got BER about e-30 for 200km link. Should I look for problem in BER test or in system ?

[Alessandro Festa]

So you are saying that the system works with BER Analyzer, but does not work with BER Test Set?

[Maciej Nowacki]

Not exactly. I think that BER Analyzer gives too good results. I got BER about e-30 for 200 km long link.

[Damian Marek]

This is why I mentioned above that you should not use the BER Analyzer for higher modulation formats. Generally, people make the mistake of converting the signal to binary and attaching a BER Analyzer, which will not work. This is because at one point you will need some type of threshold detector to convert back to binary and this will recreate a “perfect” binary representation that the BER Analyzer will interpret as a “perfect” eye diagram, even if the threshold detector incorrectly detected a symbol.

[Alessandro Festa]

Ok now I understand. I think that if you post both the layouts in the same file (BER Analyzer & BER Tests sets) we could help.

[Maciej Nowacki]

I have posted file with 3 layouts. All different layouts have problem with BER analyzer.

Thanks very much!

[Alessandro Festa]

I am seeing that BER is bad on layouts 1 and 2, but it is 1E-36 on layout 3…is this the same result you are seeing?

[Maciej Nowacki]

Yes but for layout 3 and optical fiber lenght of 220 km. I have BER around 2.30 e^-45. I think that it is too good BER.Do you share my opinion ?

[Alessandro Festa]

I understand and verified your point with 220km. But please note that you have a very limited bitrate in your system (only 19Mb/s), so I would not exclude a good BER at such a low bitrate!

[Author]

Posts