Cfrgtkky

You have a setup like in the image above. But it seems like detector A does some weird things. You should better check it with detector B. What is the minimum number of mirrors you have to move (translate and rotate arbitrarily) to bring the laser beam onto detector B?

• Please leave the mirrors which are used to direct the laser to detector A as they are, because if it turns out detector A is not broken you want to switch back to it. I mean it's quite well shielded from stray light compared to detector B.
  


• The thick black lines are walls. Please don't burn any holes in them.
  


• The mirrors all look the same. They have only one side (blue) with reflective coating.
  


• The grey grid is only for orientation.
  


• The detectors work for any angle of incidence.
  


• Make sure that the beam actually hits the detector, not only stray light.
  


• And don't look into the laser beam.
  

Assuming only 45 degree mirrors,

We are going to need two moves, and there are several ways to do it.

But since the detectors are omnidirectional, we can do better:

First, take a look at the two mirrors nearest detector B. Mark the position B' as the location of the detector B as seen through the nearest mirror. Then, mark position B'', which is B' as seen through the second mirror. From that point, draw a straight line to the center of the third mirror in the rightmost column.

Notice that the line moves 1 square sideways in 6 vertical squares. It also passes through another mirror in between. Pick up this mirror, we'll use it when we figure out where to place it.

Now, reflecting at the mirror, continue the line 6 squares to the left, during which time it moves 1 square up, because of the reflected angle. What joy, we are at a spot through which the original beam passes, so we can move the mirror we picked up here. We must place the mirror at an angle that is exactly half of the desired beam deflection (which is $text{arctan}(1/6) approx 9.5°$), or about 4.7 degrees, in order to hit the detector.

Like so:

I think you can do it with two moves: one translation and one rotation:

Bass' answer is perfectly valid, but I just wanted to share another answer that works in the same way:

One can solve it moving only one mirror.

A lazy laser physicist would balk at the extra calculations they'd have to do if any of the mirrors aren't oriented exactly at 45 degrees (to horizontal). They are not actually allowed to adjust the newly positioned mirrors on a trial and error basis (to lazily escape the calculations) because (a) they'll burn a hole or more in the walls and (b) perhaps someone will accidentally look into an unstable laser beam.

The solutions (including the one by the OP) which use mirrors at angles other than 45 degrees are smart (respect!!!) but not really suitable for use by lazy people.

Since it appears that two is the minimum number of mirrors to move, our lazy scientist would prefer one of the easiest solutions (both on their hands and brains) which should be the ones involving simple translation (with no rotation) assuming the setup allows them to do it easily enough.

One such solution would be:

The arrow marks show the translation to be done; the green line is the new laser beam.

Please excuse the shabby and not-to-scale drawing, if any.

With the assumption of easy reversibility of translations (so that we can go back to the original configuration easily), we can take it a step further and use just one movement:

The translation can be reversed to reach detector A. In effect this movement can be toggled to use detector A or B as required.

With some minor fidgeting, I'm pretty confident a setup like this can work well. Just move any unused mirror into the blue spot as shown: