Swans on Tea

I align lasers for a living. Or so it seems.

That’s not actually my job description, but if you observed what I do in the lab over a long enough period, it wouldn’t be an unreasonable description of the plurality of my effort. That job doesn’t occupy as much of my time as it once did; because of some engineering advancements we’ve made, once lasers are aligned they stay aligned for much longer, but there’s still a bit of time spent doing this.

And this is the bane of my existence

That’s a multi-axis fiber port. And they are a complete pain in the butt to align.

The reason for using them is the “engineering advancement” I mentioned — if you couple your laser into an optical fiber, it’s very easy to send that light somewhere else. If you sent it through open space, you leave yourself open to all sorts of problems — dust and dirt on your optics, things blocking the beam, and misalignment issues creeping in — if you send the light just 1 meter, each milliradian of error moves your beam a millimeter, and thermal changes can “walk” a mirror mount to steer in a subtly different direction. If you need precision, long-term alignment, fiber is the way to go.

Free-space alignment is relatively easy. This is a mirror pair we call a dog-leg (golfers might observe it’s two dog legs, but this isn’t golf)

I’ve drawn in a laser beam (you wouldn’t see an actual beam unless there was dust or other some particulates to scatter the light, and we try to minimize that) bouncing off of the two mirrors. Each mirror mount has two knobs: one pushes on a corner for horizontal tilting, while the other pushes on the opposite corner for vertical tilting; there is a small ball bearing at the third corner, and a couple of springs to hold it all together. Adjusting a knob will tilt the mirror and deflect the beam, and change the angle of reflection off of the mirror (and also off of the next mirror). You can un-do the angle change with the second mirror, so that the exiting beam is headed in the same direction as you started, but having been translated to one side. The adjustments are orthogonal, so you can “walk” a beam anywhere you want by adjusting pairs of knobs, as long as the beam continues to hit the mirrors.

That’s pretty easy, in principle; in practice it’s a little tougher, because you often do this “blind” — you are looking at the target or a display, telling you how well you are hitting the target, and the lights might be off. But it’s a skill that’s picked up pretty quickly.

The fiber port, though, is tougher. The chuck in the middle is for an optical fiber, and in my case it’s a single-mode optical fiber. The core of such a fiber is a few microns across and is very sensitive to the spatial mode (the shape) of the light it accepts. I’ve previously shown what a poor mode looks like, but assuming you have a nice zeroeth order mode, you need to send the light in with the fiber positioned just right — not only hitting the core of the fiber, but at the proper angle and with the fiber tip just the right distance from the fiber, to match the mode characteristics.

There’s a little divot on the left side of the fiber port that houses the screw for moving the fiber chuck in the x direction, and one at the top for y, and they are both sensitive and subject to a bit of hysteresis (turning through some angle and then getting back to the original position doesn’t reposition the holder exactly), and also have the annoying habit of walking off when you remove the allen key/driver. The small black dots on the face of the holder are the tip/tilt controls, which also suffer a bit from hysteresis. In addition, to translate in the z-direction (moving the fiber closer or further from the lens) you have to change all three tilt controls. In practice this means your coupling efficiency goes down as you change the first two settings, and you have to hope you can recover your signal as you work on the third. If you can’t, it may be because you are moving the wrong way along z, or because you changed one of the tilt settings too much, so you have to try many iterations to find out what’s going on.

Just so you know it’s not all pew, pew, pew when you work with lasers.