Wavelength Issues in the Downstream Direction James O. “Jim” Farmer Alan M. Brown Enablence Technologies (Wave7 Optics) IEEE 802.3av meeting, November 2008

Wavelength Issues in the Downstream Direction V-OLT OLT Upstream Broadcast (extended services) Downstream data Upstream data Optical subassembly Passive optical network (PON) down WDMWDM 1550 Must attenuate opposite signal >30 dB ONT

Wavelength Establishing View (Wavelength is to Scale) Wavelength (nm) 1 Gb/s upstream 10 Gb/s upstream Today’s 1 Gb/s downstream (1490 nm) Broadcast 1550 nm /-3 nm downstream 1590 nm down- stream 14 nm nominal transition Today’s 50 nm nominal transition (comfortable) nm nominal transition 1570

But Wait! It Gets Worse ) When we take tolerances and temperature drift into account, the situation gets worse 5 So we will make certain concessions to the real world 5 Broadcast (extended services) band 5 Was nm (existing transmitters) 5 Make it nm nm data down 5 Was and is /-3 nm nm data down (currently not in standard) 5 Was /-10 nm 5 Make it /-3 nm 5 Makes OLT somewhat more expensive, but maybe not prohibitively so. Makes ONT easier ) We shall make these assumptions and look at the implication for using broadcast with downstream data. We will not look at upstream data transmission at this time. Those messy, practical details

Assume we try to use 1577 nm with broadcast Wavelength (nm) Reduced broadcast wavelength 1. Specified 1577 nm data wavelength 2. Practical filter bandwidth after taking into account temperature drift and initial center frequency tolerance

Assume we try to use 1577 nm with broadcast Wavelength (nm) Reduced broadcast wavelength Practical filter Filter center frequency is low, and temperature forces it even lower 1 nm transition - Impossible!!!

Now try it with a downstream wavelength of 1590 nm Wavelength (nm) Reduced broadcast wavelength 1. Specified 1590 nm data wavelength (reduced bandwidth) 2. Practical filter bandwidth after taking into account temperature drift and initial center frequency

Wavelength (nm) Now try it with a downstream wavelength of 1590 nm Reduced broadcast wavelength Practical filter Filter center frequency is low, and temperature forces it even lower 14 nm transition. Tight, but much better than 1 nm

Conclusion ) Use of the broadcast overlay with a 1577 nm data carrier is impossible ) Use of the broadcast overlay with a 1590 nm data carrier is difficult, but should be feasible ) We do not object to use of 1577 nm for PR(X)30 ) We seek reinstatement of the 1590 nm wavelength in order to preserve use of the broadcast overlay, which we have agreed is still important