Adapting optical fiber input to a lens or free space

[JohnG]

I'm looking to see if there is a way to use an O/E receiver that has a fiber optic input to be able to sense free space light, i.e. some kind of adapter that has a lens, or could be fitted with a lens.

Also, if anyone can recommend a book on practical optical measurements, I would highly appreciate it. I have some RF background, and a lot of optics seems at least somewhat analogous, but I have a lot to learn here.

Many thanks,
John

[awallin]

you are going to need a high-brightness source like a laser or possibly a LED.
Multimode-fiber with several 100 um core is probably the way to start, with a short focal-length lens (5-20 mm maybe) or even a ball/bead of glass.

these might help:
https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=10417
https://www.newport.com/n/fiber-optic-coupling

with a laser-source it is reasonably straightforward to couple into a 5-10um single-mode core if you have good lenses and mechanics to move the lens or fiber by tiny adjustments - but for a beginner it's a steep learning curve..

"building electro-optical systems" by Philip Hobbs is quite practical, with both optics, detection, and electronics (maybe a bit like horowitz&hill ?).
For basic optics try "optics" by Hecht.

cheerio!

[ejeffrey]

+1 on building electro optic systems. 

Why do you need the fiber?  Is your detector fiber coupled naturally, do you need fiber optic components like switches, or do you just need your detector to be remote from the light?

What is your light source?  Multimode fibers are good for LEDs that can be focused reasonable well.  Single mode fibers are basically limited to lasers.  For even more diffuse sources you might need a fiber bundle or large light pipe.

[JohnG]

What I am wonder is whether or not I can repurpose an optical to electrical receiver with a fiber input connecter to do one or both of the following:

1. Detect some or all of a diode laser output (905 nm for now) pointed at the detector
2. Detect the reflected beam from a diffuse target (look at the spot).

I am particularly interested in looking at the shape (amplitude) of the optical pulse, which is expected to be under 2 ns wide, and eventually under 1 ns wide. Max distances in the setup will be < 50 cm. Right now, I have a single detector (ET-3500), and it will do #1 for me. The issue is that I need more than one setup, and I have access to some amplified detectors with 2-3 GHz BW, but they have fiber optic inputs. I would like to know if there is a way to use them.

Thanks,
John

[ejeffrey]

You can certainly do #1 with a fiber coupled detector.  Coupling an IR laser into an optical fiber can be tricky, but it is certainly doable.

Looking at a spot... well you mostly just have to accept how little light you are likely to get.  You are going to want a lens with a large aperture to collect as much light as possible.  You definitely want to use a multimode fiber.  You would have to look at the exact specs but a 100 mm EFL f/2 DSLR prime lens might work well for this application.  You may just not be able to collect enough light.  Depends on your laser power, detector noise, and desired geometry.

[Marco]

No personal experience, but this presentation has some pictures of some ROSAs, it seems they generally have some free space before the photodiode which should allow you to dremel it out and use it raw. Bit annoying that you can buy SFF/SFP modules cheap, ROSAs relatively cheap ... but for individual photodiodes with the same specs you suddenly have to pay 10x as much.

PS. there are some interesting photodiodes on aliexpress currently, especially the avalanche ones.

PPS. well that didn't last, the most interesting avalanche photodiodes are now now longer available.

[JohnG]

Thanks for the suggestions.

I ordered the Hobb's book to get started. I may have to learn to build a receiver, which would be interesting. I just might not have the time to do it before I need it...

Laser peak optical output power is in the order of several W minimum, but could be much higher.

Thanks,
John

[ejeffrey]

If you have several watts peak power, and are able to collect a 25 mm diameter aperture at 50 cm, then that is about a milliwatt of power.  You won't have anything like unity collection efficiency, but even a few hundred microwatt  should be easily detectable.  A much smaller collection aperture such as a 5 mm aspherical lens will collect proportionally less. The bigger your collection aperture, the more accurate your alignment will need to be.

IR lasers with this output power are extremely dangerous, be careful.  Even the scattered light can be dangerous and since your eye can't see it you won't have a blink reflex. 2 nanosecond pulses won't be a problem, but you want to be safe even if your driver circuit misbehaves.

A narrow band interference filter matched to the diode laser may be quite helpful.  Your SNR may actually be quite high even without it, but the filter makes it a lot easier to tune the alignment in as you can see the much weaker signal when you are poorly aligned,

[mjs]

The book from Hobbs is great. It covers in some depth all of it, but if you start with etendue, fiber optics and start building your photon budget, you'll get an idea if this is feasible with the power levels you've got.

Free space to fiber usually has horrible efficiency and getting things aligned is really painful. Free space to single mode fiber works only in rare cases, multimode is a bit easier, but try to get into using bulk optics if anyway possible.

[nfmax]

Another vote for Phil Hobbs's book, it's chock full of good stuff and wise advice. He doesn't really like fibre optics though, because polarisation. I sympathise...

A few years ago I was part of a team working on an interrogation system for a fibre-coupled Fabry-Perot interferometer, which had to be insensitive to polarisation state because the fibre downlead could move around a lot, and needed high signal to noise ratio despite downlead losses. We worked out a quadrature optical heterodyne bi-polarisation system using pairs of pulses transmitted with two different polarisation states, and some completely non-obvious signal processing to recover the interferometer phase from the I & Q IF channels of the two orthogonal polarisation receivers, for a pair of pulses (8 measurements in all). It was a long and tortuous thought process convincing myself it would actually work, and a truly memorable moment when we got it going on the bench giving completely stable results while we madly twiddled a polarisation controller in the downlead.

One of the best bits of work I've ever done - https://patents.justia.com/patent/8401401

The project was canned shortly afterwards, naturally