To Mirror Or Not To Mirror

In my previous post I mentioned using crossed hair targets as well as mirrored targets for doing alignments. It occurs to me that some folks doing optical alignments may wonder whether they should use a mirror-backed form of alignment target, and why they would use them, or just stick with the less expensive stretched-filament type of crossed hair targets supported in air. So here’s a discussion of this topic.

We know from basic geometry that any two points in a three-space will determine a straight line through space. So as long as you are trying to generate a single straight line within a piece of equipment (it does not matter whether it is optical or mechanical equipment), you could argue that you need only 2 reference points in that package to define your desired line of sight (LOS). By the way, all optical tooling alignment work is based on setting up some desired line of sight (LOS) by means of optical observation. The equipment and instrumentation vary a bit. When you have to fold a beam, then a new LOS will be required.

A mirrored target is just some sort of precision alignment target with a very high quality flat front surface mirror placed behind it. Sometimes, the target is actually fabricated right on that front surface mirror. You can buy various versions of these mirrored targets from vendors of optical tooling equipment; usually they are the same folks who make alignment telescopes (ATs). You can also have a target custom made to suit your needs for special applications. It is not easy to make your own, so this is not a recommended way to go, unless you have a sophisticated optical fabrication shop as part of your facility.

If all you need to do is set up a LOS between two points in your equipment under test, and the exact direction in space is not too critical (so long as it goes through those two points) then you can do this with a good alignment telescope and two stretched-filament crossed hair targets supported in air. You can make these yourself. A series of LOS observations with your AT will establish the axial LOS within your equipment under test.

But some applications are more complicated. A common situation in optical equipment requires a LOS through a plane which not only goes through a reference point, but also is normal to that plane. In this case a mirrored target may be the best choice. Then you can use auto-reflection or auto-collimation instruments to set up your LOS. This application mandates that the front surface mirror is located accurately so that it establishes the same “plane” you care about. This can be tricky. This is when an honest discussion with your machinist about the problem can save you lots of grief.

In a project I did fabricating a sensor package for a large optical telescope, I needed to get the optical components within it aligned to the theoretical optical axis of the main telescope, which would enter this package at dead center of a round aperture and normal to the plane of that aperture. I built my own mirrored target in which the mirror was removable. Once I got the LOS from my autocollimating alignment telescope to coincide with the reference center (point) and also be normal to that mirror, I removed the mirror and could then look past the crossed hairs to other components within. The mirror had to go back in place with high repeatability when I needed to use it. That is difficult to do when you are talking about LOS angular repeatability of 2 arcseconds or less, but it can be done. It worked. My assembly and alignment project was successful.

“Whether to mirror or not to mirror” is a decision which should always be driven by your application. Do good work!