optical sweep generator

Eh? If a singlemode fiber takes 800 nm, it's not singlemode for 1600nm.

nm.

Far from true. Unlike metal guides, dielectric guides have no low-frequency cutoff.

Cheers

Phil Hobbs

How much power, and what sort of modulation bandwidth?

Cheers

Phil Hobbs

We buy fiber-coupled lasers (850, 1310, 1550) and corresponding photodiodes. They all look alike. We'd like to test the incoming parts and our final products to be sure they are for the correct wavelength.

The lasers are both singlemode and multimode. The photodiodes don't much care.

I tweaked my beautiful GHz o/e converter and later discovered that the wrong photodiode was installed. I'm concerned that we don't verify wavelengths on our products.

Most OSAs are narrowband, high resolution.

For that, a tungsten source and a filter wheel would be one approach. Here are some Thor Labs possibilities.

Tungsten source:

Filter wheels:

Stock bandpass filters:

They also sell choppers that work very nicely:

but a regular old muffin fan from the junk box would probably work fine since you aren't trying to do lock-in detection with it.

Cheers

Phil Hobbs

I've had the same problem, lasers and diodes with wrong wavelength being placed (a very subtle difference in type number). In the company junk box I found a coarse WDM multiplexer, several fibers in, one combined out, all with SC connectors. Added a set of adapters to it. It was easy to check lasers 850 / 1300 / 1550 with it, using 2 detectors and a DMM. The reverse, differentiate detectors with a set of known lasers, is also easy. Hookup a few 'scrap' lasers (rejected for bad moding, etc) to a battery as a set of known sources. A detector quickly shows its type with that.

When you are using SFP's with AC coupling a set of modulated sources is needed, a 50 MHz squarewave is enough at the TX side.

Arie

I was going to say a tungsten lamp into a spectrometer/ monochromator is pretty broad band. Some NIR/IR grating. (in a past life I had access to these McPherson spectrometers with ~2" curved diffraction gratings, the gratings were treated as gold.. better.)

George H.

Yup. But for this sort of use the etendue of a 25-mm filter is hard to beat for the price. Also monochromators can be misleading if you don't guard against order overlap (aka aliasing). A tungsten bulb is quite a bit brighter at 750 nm than at 1500.

Delta filters in Denmark make very nice circular variable-frequency filters--sometime when I've got $3k burning a hole in my pocket I might buy one to play with........nah. Cute though.

Cheers

Phil Hobbs

For my photoreceivers, I could get a tungsten source, a motorized filter wheel, a bunch of filters, and some sort of lens to focus the filtered light into a multimode fiber. Spin the filters and look at the PD output on a scope. That's not totally horrible, but I probably wouldn't get much optical power. I'd like some decent fraction of a milliwatt.

Or I could get a bunch of diode lasers of different wavelengths, scan their drive, and find some way to concentrate them all into one fiber. Crude: I could get three fiber lasers, 850, 1310, 1550, and connect them one at a time to my receiver and note the levels. That would be a pain in production.

In the other direction, checking lasers, this looked great:

until I read the fine print. None of the heads will do wavelength measurement over my range.

I can imagine some number of fiber lasers, at least three, that go into a combiner of some sort and come out on one fiber. I can pulse the lasers sequentially, drive my photoreceiver, and sync it on a scope; a 3-point spectrum analyzer.

Do people make fiber combiners that work from 850 to 1550?

Or an electrically-operable N-to-1 switch, ditto?

Could you just get an Ethernet switch with SFP sockets, and get three (850, 1310, 1550) transceivers? Plug in the right source, send a broadcast, and watch your diodes' output?

SFP modules come with output power specs.

Not cheap not single mode. I know of 1200-1700nm tunable lasers and similar spectrum 'white' lasers. In ten years you'll find them from scrap yard, but at the moment you'll need deep pockets and event that may not be enough.

I've got HP70950B, which does 600-1700nm. Shorter wavelengths are second order from monochromator, but that is now a problem with narrowband sources. It's not very sensitive (compared to cooled InGaAs spectrometer with long exposures), but you won't need that with lasers.

You don't even need to send data - SFPs generate a continuous output which is amplitude modulated when there is data to send.

It is easy to view the output of 850nm SFPs with a smartphone camera.

John

A black-and-white camera with a 0.3 mm thick Si filter sees light from a 1550 nm communication laser. Is this a two-photon absorption or an random short-wave laser light ?

Good point. I guess turning the current down won't help much. :^) I recall these nice bulbs we would use. They had a filament that was a coil in a line. You set the filament to be vertical and focus that onto the spectrometer slits.. It worked OK.

Huh, neat. I never heard of those. What's inside? Some variable interference filters?

George H.

I think they must use a rotating chuck with a shield, and change the rotation speed cyclically so that the layers come out with varying thicknesses.

It would use quite a bit of tool time, which probably explains the high prices. It's not a complete solution for JL, because the tuning ranges don't cover both 850 and 1500 nm.

Cheers

Phil Hobbs

What kind of camera, and what's the optical setup?

Cheers

Phil Hobbs

We have 3 wavelengths of fiber-connectorized laser diodes in stock, and can get more. And we have power meters. The issue is whether we have to connect them one at a time to the DUT, note the output, and do some math, in production. Of course we'd have to trust about the wavelengths of the test set lasers, or send them to some lab to be checked.

The power meters want you to tell them what the wavelength is. They don't measure it themselves. Optical instrumentation is barbaric.

My Fluke DVM autoranges to measure voltage over a 1e6:1 range.

I can't get optical stuff that covers 2:1 !