- This topic has 22 replies, 5 voices, and was last updated 8 years, 9 months ago by
Gokul Boro ..
-
AuthorPosts
-
-
December 2, 2015 at 1:47 am #28295
Gokul Boro .
ParticipantHi everybody,
I would like to know the reason why the constellation visualizer shows messy constellation while the BER Test set shows BER=0.
The snapshot is attached herein.
Thank you. -
December 2, 2015 at 1:50 am #28296
Gokul Boro .
ParticipantSorry, the snapshot could not be attached in the topic. The snapshot is attached in this comment.
-
December 2, 2015 at 8:51 am #28318
Damian Marek
ParticipantI remember seeing this a while back when I was doing something wrong! I don’t remember exactly what it was, but I think it was a setting in the Constellation Visualizer, right-click and go to Component Properties. It might not have the proper bit rate or symbol rate, or the constellation type is not correct.
Try making sure all those settings are correct. I tend to always trust the results of the BER Test Set since it is a fairly simple component, so the error must be somewhere else.
Cheers!
-
-
December 2, 2015 at 2:18 am #28298
Abdallah Ismail
ParticipantI think that I faced that problem with PM-QPSK. check Polarization De-Multiplexing using DSP.
BER is zero because all the samples are in the right corners. -
December 2, 2015 at 4:29 am #28308
Gokul Boro .
ParticipantHi Abdallah thank you for your reply.
do you mean that I need to add DSP block to my system? -
December 2, 2015 at 4:45 am #28311
Abdallah Ismail
ParticipantSure
-
December 2, 2015 at 4:47 am #28312
Abdallah Ismail
ParticipantIt is a very important part in your system.
Also at first remove all impairments like noise from the system. -
December 2, 2015 at 5:42 am #28313
Gokul Boro .
ParticipantThank yor Abdallah.
But sir I am trying to implement the system going through a journal paper. In that paper there is nothing mentioned about DSP block. It seems that the result if fine without using this block. Also I have seen the constellation to be messy only at the Visualizer. When I checked the graphs of the OFDM block at the project browser, the constellation is fine. Is this a problem with the visualizer or design itself. -
December 2, 2015 at 6:02 am #28314
Abdallah Ismail
ParticipantI’m not sure.
If you are working on back to back system, you may not need to DSP. Otherwise, the DSP may be important as the fiber length increase. -
December 2, 2015 at 6:39 am #28315
Gokul Boro .
ParticipantYes, I am working on back to back system at the initial stage.
-
December 2, 2015 at 9:14 am #28319
Gokul Boro .
ParticipantHi Damien thanks for replying and what you suggested worked.
As you have mentioned I have made the reference bit rate similar to that of the BER Test Set and it worked. Now the constellation points are so much clear that it completely matches with the constellation in the Tx section. It proves that the BER Test Set is showing a reliable reading. Now can a back to back optical communication system have such response?-
December 2, 2015 at 9:46 am #28320
Damian Marek
ParticipantGreat! In a simulation it can, but realistically of course there is always a source of distortion or noise. There can be distortion from the modulators which will give an unevenly spaced constellation because of the non-linearity of the mach zehnder modulators. There can also be noise introduced by the PIN photo-diodes at the receiver side etc..
-
-
December 3, 2015 at 12:40 am #28340
Gokul Boro .
ParticipantThank you once more damien.
i would like to know one more thing related to constellation diagram. There is difference between the constellation diagram in the constellation visualizer and that of the OFDM block constellation viewed from the project browser. Which one is to be considered now among these two? Or both are correct? The snapshot is attached.-
December 3, 2015 at 1:03 am #28346
alistu
ParticipantHi Gokul,
OptiSystem 13 OFDM components have the capability of decision making based on decision boundaries in signal space. So in any case, the output signal of the OFDM demodulator will be aligned the way shown in your right image. However, you can see the signal constellation before decision making step in project browser.
Regards
-
-
December 3, 2015 at 1:24 am #28347
Gokul Boro .
ParticipantThank you Alistu. By signal constellation before decision making, did you mean the constellation before channel estimation? Now what I wanted to know was that, why the constellation in the visualizer is different from that of OFDM seen from project browser? After Pilot symbols and carrier phase estimation block, there is Decision, EVM and SER computation block. But does that block cause the signal to be so perfect?
-
December 3, 2015 at 1:58 am #28350
alistu
ParticipantYou’re welcome Gokul. No, channel estimation is another stage imposed before the decision stage. The decision making that I am referring to is just to decide which of the 16 constellation points (in your case) is the position of the signal (after channel estimation and so on), based on which position the symbol is closest to. Your left image is the stage after channel estimations and so on, and the right one the one after decision making.
-
-
December 3, 2015 at 2:24 am #28354
Gokul Boro .
ParticipantThank you Alistu. But, How do I do that?
-
December 3, 2015 at 2:52 am #28356
alistu
ParticipantYou’re welcome. You have been using OptiSystem OFDM component to get the result, so there is nothing left for you to do since the block does all the stages (the OFDM blocks in OptiSystem 13). Using constellation visualizer at the output of OptiSystem 13 OFDM demodulator does not give much information, and it is better to see it in project browser like you did.
-
-
December 3, 2015 at 4:17 am #28361
Gokul Boro .
ParticipantOk. Thank you Alistu. So I can rely on the results of the Constellation diagram that can be seen from the Project Browser (OFDM Component).
-
December 4, 2015 at 8:50 am #28381
Mohamed Ben zegalam
ParticipantMay I know the different between the three OFDM mod/demod components id Optisystem 13
-
December 4, 2015 at 9:53 am #28418
alistu
ParticipantHi Mohamed, Please see my answer in your topic related to the question; https://optiwave.com/forums/topic/ofdm-pon/#post-28417 .
-
-
December 4, 2015 at 11:44 pm #28421
Gokul Boro .
ParticipantIn earlier version, OFDM component was incapable of incorporating pilots and training symbols. But Optisystem 13 version is capable of doing this. And these parameters are very important for better signal transmission.
-
-
AuthorPosts
- You must be logged in to reply to this topic.