& cplSiteName &

Dispersion: New Version

Light Reading
News Analysis
Light Reading
4/29/2002

Component startup Accumux Technologies Inc. today announced its first product: a dispersion compensator module that it claims delivers better performance, in a much smaller package, than existing technologies (see Accumux Intros Dispersion Compensators).

The module, called Light Fixer, is the first dispersion compensator to make use of so-called etalon technology -- bouncing light back and forth between two mirrors. The same idea is already used in other components, such as lasers and wave lockers, but hasn't yet caught on for curing the smudging of light pulses that occurs when high data rates are carried over long distances.

This smudging happens because light pulses are made up of bunches of wavelengths that travel through fiber at slightly different speeds. As a result, the light pulses spread out when they travel long distances, eventually overlapping and becoming indistinguishable from one another (see Light Reading's Beginners Guide for the basics on Chromatic Dispersion and Polarization Mode Dispersion (PMD)).

Until now, two techniques have been used to tackle this problem.

The first is based on the use of special dispersion compensating fiber from vendors like Corning Inc. (NYSE: GLW). However, this solution is bulky, expensive, and can have high insertion loss, according to Ross Kocen, Accumux's VP of marketing. It's also fixed. The amount of dispersion compensation can't be adjusted, so this can't be used in networks that deploy optical switching.

The second solution uses Fiber Bragg Gratings (FBGs) from companies such as Alcatel Optronics (Nasdaq: ALAO; Paris: CGO.PA), Phaethon Communications, Sabeus Photonics Inc., Southampton Photonics Inc., TeraXion, and 3M Company, to name but a few. FBGs can be tuned to deliver different amounts of dispersion compensation, but there's another snag: They can't be used for multiple wavelengths. You need one FBG per wavelength, so this solution is expensive and cumbersome in networks using DWDM.

Accumux claims its etalon-based compensator has none of these drawbacks. It works by making some wavelengths of light bounce between its mirrors for longer than other wavelengths, in order to restore the original shape of the pulse. The upshot is compactness and the ability to compensate 80 channels at the same time, across the C and L bands.

As noted, Accumux claims to be the first to have an etalon-based dispersion compensator, but at least one other company -- JDS Uniphase Corp. (Nasdaq: JDSU; Toronto: JDU) -- is also working in the field, as visitors at the recent Optical Fiber Communication Conference and Exhibit (OFC) may have noticed.

At the OFC, Accumux was a little secretive about its Light Fixer, covering it with a cloth during most of the show. JDSU, on the other hand, made no bones about having a working prototype of a tunable etalon-based compensator on its booth. David Moss, from JDSU in Ottawa, also presented a research paper on the company’s developments in this field.

JDSU's presence in this market could be taken either as a validation of Accumux's technology, or as a huge threat. After all, how can a 20-person startup hope to compete with the 800-pound moose of the component industry?

What's more, JDS Uniphase's technology may be technologically more advanced. Accumux's device has limited tunability -- temperature is used to tune the dispersion of the device by up to 10 percent of its central value, according to Kocen. But the dispersion slope, which describes the varation in dispersion with wavelength, remains fixed. In contrast, the device described by JDS Uniphase in its research paper has tunability of over a much wider range of dispersion, and can offer different dispersion slopes (although the amount of slope is fixed during manufacturing).

JDSU is still in its development phase and hasn't said when its etalon-based products will hit the market. Mind you, Accumux isn't actually ready to ship either. It expects to field test the product in the summer, with volume shipments starting in 2003, according to Kocen.

Accumux was founded in January 2001. It has raised $7.3 million so far in its series A round, from Fortune Ventures, Smart Technology Ventures, Trimax, and Lite On IT Corp., which is enough to last until 2003. Nevertheless, it's seeking another $5 million before it closes the round, Kocen says.

— Pauline Rigby, Senior Editor, Light Reading
http://www.lightreading.com

(8)  | 
Comment  | 
Print  | 
Related Stories
Newest First  |  Oldest First  |  Threaded View        ADD A COMMENT
Pauline Rigby
Pauline Rigby
12/4/2012 | 10:30:06 PM
re: Dispersion: New Version
I've read the paper again. It was the last paragraph that confused me:

"By intentionally mismatching the FSR [free spectral range] of the two etalons, one can introduce a linear variation of the relative wavelength offset between the two etalons across the band, which will in turn introduce a linear varation in channel dispersion across the band. The resulting dispersion slope (fixed during fabrication) in combination with normal tuning, enables the dispersion to dispersion slope ratio to be adjusted to matched to any fiber type."

My original take on this was that it's analogous to the way they use dispersion compensating fiber to compensate any fiber type. To do this requires two different types of dispersion compensating fiber (DCF) with opposite dispersion slopes. By adjusting the length of the two pieces of fiber, it's possible to achieve any value of dispersion and dispersion slope. (Ax + By = whatever line on a graph that you want.) I originally thought it would be possible to do the same thing in etalons.

However, for tuning etalons, you only have temperature. Tuning temperature gives different amounts of dispersion, but I now see that it doesn't change the dispersion slope.

What you need to change is the FSR, which I think of as the length of time that a light pulse bounces around inside a cavity. If light bounces around for longer, its like sending light down a longer piece of DCF. FSR depends on the length of the cavity, which is fixed in manufacture, as you point out.

My interpretation now is that JDSU's approach makes it possible to design different dispersion slopes, although once manufactured, the slope is fixed. In contrast, using a single etalon means you have to settle for whatever dispersion slope the physics of the cavity dictates.

Can you explain what's going on in the paper any better? I'll update the story when I'm sure I understand what's going on. I did contact JDSU to ask them about it, but the product line manager was out of town.

[email protected]
Biosphere
Biosphere
12/4/2012 | 10:30:06 PM
re: Dispersion: New Version

From my reading of the JDSU paper the slope is fixed during manufacture and not tuneable as implied in the above article.

Any comments?

Also the link in the article is incorrect - the one below should work

http://jdsunph1.jdsunph.com/si...
bigdaddy
bigdaddy
12/4/2012 | 10:30:06 PM
re: Dispersion: New Version
Golly, All these new Chromatic dispersion solutions are really going to obsolete the old fiber based "management" solutions. I wonder how the new devices are doing with Telcordia qualification??? That can be a real tough hurdle if you don't know what you are doing, I hear.

smaller, better, cheaper, will allow you to hit the window of opportunity. Better hurry before the window closes.

What is 5% of a 30million dollar market??? Almost not enough to make a business case out of it.
Perceptor
Perceptor
12/4/2012 | 10:30:00 PM
re: Dispersion: New Version
From a cursory glance at the article, I believe it is tunable. By introducing the mismatch between the two mirrors, a linear dispersion vs. wavelength is formed in the etalon. I suspect that the slope of this disperson is tunable by altering the etalon spacing (i.e. a piezo or some other kind of moving part), which I believe is fairly standard practice when working with Fabry-Perot etalons.

So the etalon is built with an inherent dispersion, the slope of which is tuned via the etalon spacing in order to compensate for the dispersion of the system.
bigdaddy
bigdaddy
12/4/2012 | 10:29:45 PM
re: Dispersion: New Version
Pauline,

As I read it your analysis is correct.

From a detailed look at the paper it is clear that they are tuning the dispersion:

"The value of the dispersion is determined by the relative spectral shift between the two etalons, which we control thermally."

The slope is optimized for the fiber type (distance?) spec. and fixed during production:

"Finally, by intentionally mis-matching the FSRs of the two etalons, one can introduce a linear variation of the relative wavelength offset between the two etalons across the band, which will in turn introduce a linear variation in channel dispersion across the band. The resulting dispersion slope (fixed during fabrication) in combination with normal tuning, enables the dispersion to dispersion slope ratio to be adjusted to match any fiber type.

What is not entirely clear is that if they are fixing the slope and tuning only the dispersion, they are still restricted to a predetermined span distance (e.g. 80km, 100km, Etc... for a given fiber type) because of the devices fixed slope and the increase in slope as a function of span distance.

They mention the "dispersion to dispersion slope ratio" which is most likely the Lucent RDS factor. So they are matching a spec value with the fixed slope and playing with the dispersion to match RDS. Again this is better than what is out there today, but, still not ideal.

They mentioned this device as a "nonlinearity free alternative", but they do not address any high power testing or resistance to NLEs.

They also mention Insertion Loss ripple of 0.1 to 0.4 dB, but, they do not show any of the results in spectral graph form.

Often times the most important things are what you don't see or hear. I look forward to more answers/details from JDS in the near future. This product does have some potential if the cost is right and it can be productionized.

Cheers,

Bigdaddy


dwdm2
dwdm2
12/4/2012 | 10:29:28 PM
re: Dispersion: New Version
However, for tuning etalons, you only have temperature. Tuning temperature gives different amounts of dispersion, but I now see that it doesn't change the dispersion slope.
----------------

Only thing to add is that it is not the temperature rather the gap of the etalon is used to tune the resonance frequency. Some details can be found at

http://www.ieee.org/organizati...

Cheers
Dark Fiber
Dark Fiber
12/4/2012 | 10:29:15 PM
re: Dispersion: New Version
If you believe ElectoniCast numbers for DCM, they claim by 2004 DCM market at $579.5 million with 50% DCM and 50% DCF. So 5% of DCM is $15 million. Not bad for a small companyG«™.
dwdm2
dwdm2
12/4/2012 | 10:29:12 PM
re: Dispersion: New Version
Dark Fiber writes: "If you believe ElectoniCast numbers for DCM, they claim by 2004 DCM market at $579.5 million with 50% DCM and 50% DCF. So 5% of DCM is $15 million. Not bad for a small companyG«™."
---------------

Actually the dispersion market may be bigger than it appears. In a recent report CIBC points out that one-third of all of the optical fiber market is occupied by long-haul fibers. This older generation fibers account for approximately 33% of installed fiber base. This huge investment is not going to be replaced soon with more expensive dispersion compensating fibers; rather, they are going to be in use for some time.

Since a significant portion of the long-haul transport requires viable dispersion management, one can surmise that dispersion management gears will get a sizable market share.

Cheers,
AR
Featured Video
Upcoming Live Events
March 16-18, 2020, Embassy Suites, Denver, Colorado
May 18-20, 2020, Irving Convention Center, Dallas, TX
All Upcoming Live Events