Optical components

Qusion Touts Wideband Modulator

Components startup Qusion Technologies Inc. announced its first product yesterday, a wideband optical modulator based on indium phosphide (see Qusion Goes Wide).

Modulators are key components that control lasers by turning the optical signals on and off to transmit digital data. "Wideband" indicates that the device operates over a large 40 nanometer wavelength range, enough to cover the entire C band. This makes the modulator "a perfect fit" with widely tunable lasers, says Qusion's CEO Philip Wallace.

According to Wallace, the product represents a significant improvement over existing electro-absorption modulators (EAMs), which only work over a limited wavelength range, typically 5 to 15nm. He claims that a range of 40nm is an industry first.

But competitor Agility disputes the claim.

"The thing about that is, we've got one too," said James Regan, Agility Communications Inc.'s managing director for Europe, upon hearing Qusion's claim. Agility developed a widely tunable EAM some time ago, which it has integrated on-chip with its tunable laser (see Agility Packs Three Into One).

Regan concedes that Qusion may indeed be first with a standalone EAM with wideband properties. But, to put it rather bluntly, he can't see the point of a standalone product. In his view, most of the advantages of an EAM result from integrating it with other components.

A second factor is polarization sensitivity. If both laser and modulator are on chip, it simplifies things -- the polarization state is known, so it can be accounted for in the design.

If, on the other hand, the signal has to pass through an ordinary fiber pigtail between laser and modulator, the polarization becomes scrambled because of stress-induced birefringence (caused by bends, vibrations, and so on). The result is high polarization-dependent loss (PDL) at the modulator output.

Some of these drawbacks can be mitigated by adopting a slightly different approach -- co-packaging. This means the laser and modulator are still separate items but are placed inside the same package, along with microlenses to couple light from one to the other. Since there is no fiber pigtail, the PDL problem goes away. Although it is expensive to align the elements inside the package, it's still cheaper than packaging separately.

Qusion declined to provide a technical expert to comment on the details of its approach or whether this might get around the potential drawbacks pointed out by Agility.

Marconi PLC (Nasdaq/London: MONI) and Nortel Networks Corp. (NYSE/Toronto: NT) both make co-packaged laser products. The fact that Qusion plans to sell its modulator in die form suggests that it's planning to partner with other vendors and do a spot of co-packaging. The startup says samples will be available from February 2002.

But in this application, too, Qusion will still have to convince other vendors that indium phosphide EAMs are a better prospect than lithium niobate, and that's not going to be straighforward. While EAMs do have certain advantages in size and low-voltage operation, they still have some catching up to in terms of performance (see CyOptics Claims Modulator Milestone).

Qusion says it's not planning to integrate its modulator with a laser anytime soon. "This isn't our entrée into the tunable lasers; it's not even our next step," says Wallace. "We expect that in a few years, we will have products with lasers in."

— Pauline Rigby, Senior Editor, Light Reading
Photonboat 12/4/2012 | 7:26:07 PM
re: Qusion Touts Wideband Modulator I checked Cyoptics' website, they tout coverage of entire C-band. The Agere specs were not specific as to the amount of bandwidth covered. So, Qusion may have made a nice EA-ML, but not any better than their competition.

Yes, EA-ML does indeed reduce costs. However, it doesn't meet all the performance criteria necessary in some applications, and so LiNbO3 modulators continue to sell and are an emphasis at some firms, such as JDSU. So, we'll probably continue to see EA-ML developed in parallel with LiNbO3.

Anybody care to comment on the performance tradeoffs EA-ML vs. LiNbO3?
pasquinade 12/4/2012 | 7:25:31 PM
re: Qusion Touts Wideband Modulator Photonboat:

"I checked Cyoptics' website, they tout coverage of entire C-band."

*Coverage* is not the same tunability. You get coverage by filling wavelength bins, each of which cover a certain subset of wavelengths. You don't use a single part to cover the whole C-band, for example. It may only cover 5nm, depending on how picky you are about output power and reach. If you want something that works at a wavelength outside that range, you need to order another part (which, apparently, CyOptics has available).

Also, *tunability* is not the same as colour blindness. You can put any C-band wavelength through a single LiNbO3 modulator and not much will change. Try the same with an EAM and the performance characteristics will alter dramatically. There is no comment on how Qusion achieves tunability. One method would be active thermal tuning; however, while this is easy to do for a stand-alone EAM, it will make putting it in the same box with a laser either expensive or difficult (or both).

Re: LiNbO3 vs. EML
* At 10G and lower, LiNbO3 has more favourable chirp characteristics, which translates into longer reach. Furthermore, advanced coding schemes such as duobinary are more straightforward with LiNbO3 and can increase reach even further. With an EML, you don't get the "cadillac" reach of the LiNbO3, but you do get lower cost, smaller footprint, and lower drive voltage (=lower cost and lower power dissapation for the driver).
* At 40G and beyond, links >~8km are dispersion managed/compensated to first order. This means the chirp characteristics of the EML don't matter as much, as long as the resulting optical spectrum is still transform limited and the non-linear chirp is low. In theory, you can get all the economics of the EML 10G, but with the cadillac reach of LiNbO3. Not much has been done either engineering-wise or in the scientific literature on EMLs in dispersion compensated 40G links, so the jury is still out on the practicality. Anyone in industry seriously trying to do metro or long haul 40G links manufacturably is using LiNbO3 to try to squeeze every ounce of margin from their link budgets. This may change as the technology matures and prices are driven down.
Sign In