Question about return path in optical network | docsis.org

You are here

Question about return path in optical network

4 posts / 0 new
Last post
snedim
Question about return path in optical network

Hello everyone, I know that this is not a "docsis" question but I hope so that you can help me about this dilemma. I'm new in one cable tv company in BiH , and we have many problems with some optical nodes that are joined together, and to time to time we have noise or modems just goes off. Problem is that for example 3 nodes have the same return path wavelength(1310nm)with fp laser and there are joined with optical spliter 1/3 (33%) ,and send back to optical receiver with one fiber.
I think that this type of joining is problem and that we must use some sort of dwdm or cwdm for return path and that return path wavelength must be different , but my boss keeps telling me that problem is not in that , and problem is something else.

Does anyone have experiance with this.
Thank's in advance

mbowe
For the forward path, it

For the forward path, it would be common enough to have 1 x laser per mac domain. Connect the laser to optical splitter (up to 6 way or so), and shoot the output to all nodes in this mac domain.

For the return path, it would NOT be common to have an optical splitter. I think the only time this would be appropriate would be when you had dual/diverse fibres from each node and you were sending the same signal via both paths.

If you wanted to combine the return from multiple nodes (each with their own wavelength), then you would use a WDM MUX, not a splitter.

Are you sure your 3 way optical splitter is being used on the return and not the forward path?

OldDocsisGuy
Question about return path in optical network

DON'T combine optical returns unless you use different wavelength's. My company tried it in the late 90's to try to save money, and it didn't work.
The wavelengths will vary just a bit due to temperature swings, build quality of the lasers, etc, and will eventually beat together.
We have time lapse pics from a spectrum analyzer showing the noise floor going from a -35db to a +7db. Obviously, this won't work. You need to mux different light frequencies together and demux them at the receive end.

snedim
Thanks mbowe and OldDocsisGuy

Thanks mbowe and OldDocsisGuy for quick replay ,
Mbowe I'm sure this is the way how they are joined unfortunatly. I have the same opinion like you two but I needed opinion from people with more experiance in this area. I'm new in this company so I need to make project for new part of CATV network and I'm stuck with this because i think this is the wrong way. The main reason why my boss keeps telling me that this works because some part of network are made on this way and it works.

Log in or register to post comments