After hearing about troubles aligning lasers I decided to pay more attention to the process and think about it more mathimatically. There is a lot more involved than just bouncing off a bunch of mirrors until you hit the workpiece. This is especailly true when setting up a home built machine, because the mirrors are being located for the first time.
As I see it, hitting the center of any given mirror is not very important. What is important is hitting the the exact same spot on any moving mirror regardless of where it is. If it hits a different place on the mirror at different locations it means that the beam is not aligned to your movement planes. This will result in distortions on your workpiece. Line lengths and angles will be incorrect.
Therefore, before you worry about hitting any mirrors, you need to make sure the mirrors are in the right locations. Imagine two planes along the movement axis, one vertically and one horizontally. The mirrors, ideally the centers of them, have to be on the intersection of the planes. I brought all my mirrors as close together as the movement would allow. I then measured off common vertical and horizontal references to get them onto the same planes.
- Important Planes
Now you are ready to align the beam with these planes. Start with the first mirror that sends the beam down to the first moving mirror. Don't think about trying to hit that mirror, think about squaring the beam to the movement planes. Place a target that can be marked, like thermal paper or a piece of wood in front of the mirror. If the beam is not square it is going to make marks at different locations. In the graphic you can see the beam goes down. You want to move the beam so it follows the green line. Or, at least one parallel to it. The red line is the beam before correction.
I make my first mark as close the the mirror I am working with as possible. As you can see that will always be closer to the desired line than the further away points. Then move as far away as possible and make another mark. They will be off up and down as well as side to side. Now keep adjusting the mirror until you can make new marks right on top of the original mark. Actually you might want to overshoot a little if the two marks are quite a bit apart, because, as shown in the graphic the original dot is off too. In theory you could calculate exactly how much overshoot, but that would only apply if the mirror was rotating exactly about the point the beam is striking, which won't be true on our mirrors. (Google gimbal mirrors)
Repeat this a few times until they are right on. The beam is not focused at this stage, so it is a relatively big dot and the burn mark usually has fuzzy edges, but do your best.
Now check to see that the beam will actually strike the next mirror near the center. If not, you must move the mirror. Do not re-aim any mirrors. Move the mirror. Repeat the above process for the next moving mirror.
On my machine, from a complete tear down, of the axes, it took about 30 minutes to position the mirrors, 30 minutes to setup the Y axis and about 20 minutes for the X axis. The mirrors did not need to be moved from positions I setup in the first step.
To verify the result I ran a 6" square and pulsed the laser at the corners. It appears to be accurate within a few thousands on an inch on each leg and the diagonals. I took more time this time and the results paid off. It is more accurate than ever.