x

JDSU Parades Picky Part

It may seem small, but JDS Uniphase Inc. (Nasdaq: JDSU; Toronto: JDU) says its new two-port reconfigurable Wavelength Blocker (WB) is the first concrete evidence of a breakthrough technology it unveiled last week (see JDSU: MEMS Aren't Just Memories).

The WB, announced this afternoon, will be shown at the upcoming Optical Fiber Communication Conference and Exhibit (OFC). It's already shipping to customers, JDSU says, although names can't be given.

The breakthrough JDSU touts is that the WB can be used by OEMs to pick and choose wavelengths on command, a feature that's been largely missing from optical components used in DWDM gear.

It works as follows: The device takes in any wavelength in the C-band of DWDM frequencies (1520 to 1570 nanometers) or in the extended L-band (1570 to 1620 nm). Wavelengths can be accepted with spacing of either 50 GHz or 100 GHz. While not a switch, the WB blocks or attenuates any of the incoming channels, sending only one signal among many through its output port.

"The WB will save costs, replacing a lot of [optical-electrical-optical] conversions," says Dave Danagher, VP of switching and routing at JDSU. The part will help OEMs create equipment that allows optical network adjustments to be made automatically, via software, eliminating the infamous "truck rolls" that signal manual intervention. And that, Danagher says, will lead to dramatic cost savings for carriers.

The WB appears to be the latest in a groundswell of reconfigurable parts set to debut at OFC. Earlier this week, for instance, SpectraSwitch Inc. announced it will be showing a multiport, reconfigurable optical component at the conference (see Optical Switching Gets Flexible).

Like the WB, SpectraSwitch's product is aimed for use in reconfigurable optical add/drop multiplexers. It's also implemented in liquid crystal technology, like the WB. But SpectraSwitch's product has eight ports, and the WB has just two.

Not to worry, JDSU says. It's got multiport follow-ons planned. And those will be based on the MEMS (micro-electro-mechanical system) technique the vendor touted in its announcement last week. "Liquid crystal is a better fit for wavelength blocking," Danagher says. But MEMS will be needed to give the additional flexibility required to support the more complex functions in higher-end switching equipment, he maintains.

JDSU spokespeople say there's no contradiction here. The same new technology is being used to select wavelengths, whether the component is implemented with liquid crystal or MEMS, they say. But there are tradeoffs of cost and efficiency in using one or the other.

Just how these reconfigurable parts perform, and how the different implementations stack up against one another and against the offerings from other vendors, remains to be seen. Pricing isn't being given out by any of the vendors involved, nor is information on who will trial and test the goods. Perhaps the OFC will provide a starting point for real-world tirekicking and comparisons.

— Mary Jander, Senior Editor, Light Reading
http://www.lightreading.com For more information on OFC 2002, please visit: www.nottheofc.com

Physical_Layer 12/4/2012 | 10:50:40 PM
re: JDSU Parades Picky Part Can anyone comment - isn't this device basically the same as the company's LC based dynamic gain equalizer...except that instead of attenuating, the liquid crystal mechanism will simply block the wavelength totally?

Can anyone elaborate on why you would want to block wavelengths totally?
Pauline Rigby 12/4/2012 | 10:50:38 PM
re: JDSU Parades Picky Part You can use the DGE as the basis of an optical switch. I'm going to do an update shortly that explains why. So please return to this article.

[email protected]
let-there-be-light 12/4/2012 | 10:50:38 PM
re: JDSU Parades Picky Part I'm just guessing, but it might be useful to block wavelengths totally in sensitive OADM applications where you want to be sure that there is no crosstalk between the dropped and added channels.

Any other ideas out there in cyberland?
Physical_Layer 12/4/2012 | 10:50:38 PM
re: JDSU Parades Picky Part "I'm just guessing, but it might be useful to block wavelengths totally in sensitive OADM applications where you want to be sure that there is no crosstalk between the dropped and added channels."
_________

I was thinking the same thing, but sort of concluded that this would add significant cost...you'd think the equipment guys would just expect better performance on the add/drop component and not require a WB - but I can't be sure. I wonder if there is some obvious application we're just not thinking of yet.
Petabit 12/4/2012 | 10:50:37 PM
re: JDSU Parades Picky Part Pauline,

The DGE is the Optical Blocker. You use a 10% tap to do a passive drop of the channels, use the wavelength blocker, and then a 10% tap to add new channels back in.

Why? Because the day 1 cost is very low (no mux / demux filters), you can change the add / drop channels using filters that are not in the express path (no traffic interruption), and they thing actually works.

All DGEs can be used as wavelength blockers in this way (and vice versa).

P.
eewhiz 12/4/2012 | 10:50:37 PM
re: JDSU Parades Picky Part I believe that Petabit is on target on the intended application, and to realize that it is configurable, so that any optical carrier can be add/dropped. TO do this with AWG's or thin film filters and a boat-load of optical switches or shutters would cost alot more than this WB...
eewhiz
alloeo 12/4/2012 | 10:50:18 PM
re: JDSU Parades Picky Part Petabit, in general I agree with you. However, regarding actual applications, I still have a few doubts:

>You use a 10% tap to do a passive drop of the channels, use the wavelength blocker, and then a 10% tap to add new channels back in.

So for a channle to be added, the 10/90 combiner introduces 10dB loss. For the same channle to be dropped, a 10/90 splitter gives another 10dB loss. This is 20dB loss just out of the box, not counting insertion loss, etc. That means amplifiers, and noise, etc.

Second, if I want to drop more than one wavelength, I have to split 10% for each wavelength I want to drop, that's 10% less power for the bypassing wavelengths.

Third, this is more of a concern here: if I only tap off 10% of the light I want to drop, there will be 90% that keeps traveling on, how can I get rid of this residue wavelength and add a fresh new one on the same wavelength?

Also, I believe the same function can be achieved by using a or more tunable FBG and a circulator.

This is a nice step forward from JDSU, but my guts tells me it's still far from real applicaiton, and is arguably part of the anual OFC hype.

HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE