Telephotonics Starts a Price War

Telephotonics Inc. says it intends to set a new price point for optical channel monitors, making them so cheap that every DWDM (dense wavelength-division multiplexing) equipment provider can afford to put one inside its box.

Right now, network monitoring generally happens in the electronic domain. That's fine inside switches and routers where the signals have to be turned into electricity anyway in order to be processed by an electrical fabric. But there is still a need to monitor the health of the individual wavelengths in the optical fiber, in order to pinpoint faults at the optical level. That's where optical channel monitors (OCMs) fit in.

If OCMs are to be included inside network gear, then they need to be small. Several companies are developing integrated devices that fit the bill in terms of size, including Bookham Technology PLC (Nasdaq: BKHM; London: BHM) and Axsun Technologies (see Axsun Launches Tiny Optical Analyzer). But these cost about $12,000 apiece. In cost-sensitive applications like the metro market, that's not affordable, says Fadi Daou, Telephotonic's VP of product development.

"Equipment makers are telling us that if we can make them for just three or four thousand dollars each, then they will buy them in quantities of 100,000," he told Light Reading. "It would open up entirely new applications."

Daou and other key executives from Telephotonics showed Light Reading a working prototype of the OCM in a private suite in a hotel adjacent to the Optical Fiber Communications Conference in Anaheim, Calif., recently. The demonstration also included a variable optical attenuator (VOA) array. Both products were announced at the show (see Telephotonics Unveils Metro Package), but only a select few were invited to see the demos.

The heart of the OCM is a solid-state tunable filter. It scans across the wavelengths in the system, recording optical power at thousands of wavelength positions. Electronic circuitry then converts this data into information about channel power and optical signal-to-noise-ratio (OSNR). The OCM has a wavelength resolution of 0.1 nanometer, says Daou.

There are all sorts of ways to build tunable filters, from micro-optics to mechanically tuned fiber Bragg gratings (see Tunable Filters Go Solid State). Daou wouldn't reveal exactly how this one works, except to hint that it's something new. It's based on an optical polymer with a thermo-optic coefficient that's 40 times greater than glass, so it consumes very little electrical power. Yet the design is athermal, so it doesn't require temperature stabilization. Furthermore, it's small and it tunes quickly.

But the most salient feature is the material -- polymer -- which is very cheap and quick to process, says Daou (see Startup Creates Component Cocktail). The manufacturing is based on standard semiconductor processing techniques, so the startup was able to give itself a leg up by buying an abandoned silicon chip manufacturing plant.

If it can deliver on its price promise then Telephotonics may spur other manufacturers of OCMs to cut prices significantly. Jim Lewis, one of the founders of Axsun, figures that prices will tumble as the technology matures. Axsun sells OCMs for $12,000 to $13,000 each in small volumes, he notes. "Small quantity prices are not meaningful," he adds. "Volume manufacturing should make things orders of magnitude cheaper."

Other startups, like Cidra Corp. (proposed Nasdaq: CIDC) are working in the same space. It's too early to say whether they think they will be price competitive.

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

Petabit 12/4/2012 | 8:36:08 PM
re: Telephotonics Starts a Price War Ooooh. So Telephotonics showed you a private demo. Cool.

If you are really paying $12k in small quantities for an optical channel monitor, I suggest that you find a different source. They were on sale two years ago for less than $6k in single quantities. That was from one of the four OCM vendors that JDSU now own. Want to bet against them...

Let's address some of the other points in this article.

You don't need to know if you channels are within 0.1nm of where they are supposed to be. If they aren't on grid when they leave the transmitter site, there ain't anything you can do about it further down the line. At the Tx site it's a useful feature, and nowhere else. You just need to know which channel you are looking at.

Athermal? Polymer? With a high thermo-optic coeff? OK. I believe you. Just. I am willing to bet money that none of the big systems vendors will believe you. They will insist on temperature testing. For every single unit. Which will add $500 to the manufacturing cost of every item. Big chunk of the profit.

There is no mention of the speed of the device. Some of the devices (like Bookham's) are fast enough to allow protection switching. They can only achieve this by not sweeping. The control and data processing required by a sweeping device will always make it slower than the integrated ones.

Bit of a me-too device in what is already a very crowded space. Telephotonics will have to be really cheap if they are to carve out a niche in the optical components space.

exphoton 12/4/2012 | 8:36:04 PM
re: Telephotonics Starts a Price War thank you petabit for a sensible post.

i wish lr would do SOME (despite my aversion to upper case, i am forced to make a point here) level of cross-checking before publishing such piffle. just because someone gave you a private demo, you should NOT write a glowing review.

so, back to basics (and reiterating some of petabit's points)

1. first ask for telcordia test results. does the device pass the temperature-humidity test ? (wet bake - ~ 0 to 75 deg.c, etc.)
2. polymers are known for their long-term stability problems. how are these problems solved ?
3. how do you reference the wavelengths ? allow for drifts ? (especially if it is not temperature stabilized)
4. price point is not impressive. ocm's at $5k - no big deal.
5. so at the heart there is a solid-state tunable filter. "solid state" with a polymer ?
6. there are ocm's that work with fiber gratings that are not mechanically tuned.
7. was the voa part of the ocm ? or was this another device ?
8. did you think about the fact that in "cost-sensitive" applications, there may not be a need for these devices ?

etc. etc.

sorry pauline, i usually see folks criticize lr for sensationalism. i am merely picking on your inability to address some key issues surrounding a functionality (channel monitoring) and a technology (polymer) when you file your report.

Pauline Rigby 12/4/2012 | 8:35:57 PM
re: Telephotonics Starts a Price War As I understand it, Bookham's device contains a demux followed by an array of detectors, one for each channel. That's why it can check all channnels simultaneously. It tells you two things: whether the channel is there, and what power it has.

OCMs containing sweeping filters like those from Telephotonics and Axsun can't give all that information simultaneously. But they give an additional piece of info: the OSNR. That's because they can measure optical power between the channels. Bookham's device can't do this.

It's difficult to compare like with like with these monitoring devices. Bookham calls its device an OCM. Telephotonics calls a monitor that can test OSNR an "optical performance monitor" (OPM) rather than plain OCM. Axsun also calls its basic sweeping filter device an OPM. So far so good. But Axsun calls a sweeping filter device with an on-board reference wavelength an OCM. Confusing, huh?

OCMs and OPMs do not necessarily hit the same applications. If you wanted to do protection switching, for example, you wouldn't need to check OSNR, would you? So is this protection switching thing relevant to sweeping filter devices?

Incidentally, Axsun says that having the on-board reference improves the wavelength accuracy from 100 pm to 25 pm. But Petabit said that no-one needs to know channel position with even 100pm accuracy -- if it's off grid, then it shouldn't be in the system at all. I would have thought that you need a degree of wavelength accuracy to be able to calculate OSNR, but I'm not sure why anyone needs accuracy of 100pm.

It's seems to me that the whole topic of optical monitoring is a bit new, and people aren't clear on what's needed. Is this right? I'd welcome any help in understanding what the requirements for optical monitoring are.

[email protected]
Petabit 12/4/2012 | 8:35:57 PM
re: Telephotonics Starts a Price War There have been a couple of really good presentations at some of the conferences recently. OFC didn't have any, I remember seeing one in London, about October last year.

It all comes down to what to want to know, and what you can measure. Ideally you would like to know which channels are going where, how healthy they are and what they are carrying. Very few of these can be discovered in the optical domain - most require some line rate decoding.

You can measure the power in a channel. Useful to find out which ones are present, but no help in where they came from.

You can measure the wavelength, but as we said before, an off grid channel doesn't deserve to be there in the first place.

You can measure the OSNR. Which is useless for two reasons. Firstly, the channels are getting closer together - Ciena were talking about 10 Gbit/s on a 12.5 GHz spacing at OFC. There is no gap between the channels to measure the noise - so you can't measure OSNR is some of these high capacity systems. Secondly, OSNR doesn't actually tell you how healthy a channel is. The noise mechanisms that you are really interested in are the ones where the noise can beat with the signal. These are quite often at the same wavelegnth as the signal (four wave mixing, multi-path intereference are two), and so an OCM cannot tell the signal from the noise.

The only true test of the health of the signal is the receive it and to measure the bit errors. OSNR gives you a hint, but is not always accurate.

The main uses for OCMs today, are for channel inventory (what have I got on this fiber) and for feedback to control systems (EDFA tilt control, Gain Equalisers, etc). Forthe feedback case, the essential requirement is for power accuracy. If you can measure to better than 1 dB you can't control your system to any better accuracy.

What the polarization sensitivity of the Telephotonics device? I suspect that it will dominate the power accuracy.

Sign In