I had an idea for measuring squareness of the gantry which goes something like this:
Use a laser line generator and a 3D printed apparatus to project a vertical line parallel to one of the gantry rails. The 3D printed part must have fine adjustment capability to get the laser exactly parallel to the rail. Then, a mirror with its back side placed against the other gantry rail reflects the laser line back to the laser source and this indicates squareness.
In this diagram the green block holds the laser (red line) parallel to the gantry rail (dark blue). The light blue line represents a mirror and by holding the back face of the mirror against the other rail, it is forced to be parallel with the rail. The reflected laser line (orange) indicates whether the mirror and in turn the other rail is perpendicular to the laser.
This arrangement indicates only whether the gantry rails are square to each other, and does not care whether either one is square to the frame.
To ensure the gantry rails are square to the frame and to each other, you would place the mirror on the frame instead of the gantry rail to indicate whether those are perpendicular:
Now if you know whether the rails are square to each other or to the frame, how do you adjust it?
I created some knobs that clamp onto the end of the pulleys. By turning these knobs, it is easy to get the motors to skip steps one at a time. There is quantization in the possible adjustment, but the resolution afforded by the increment (four whole steps) seems good enough to be on par with the accuracy of the laser.
Here is a demo of this method, although I have not yet gone all the way to drawing a rectangle and measuring diagonals.
Neat idea. If you flipped the laser holder to the other side of the rail, pointing behind the router instead of infront of it, it wouldnât need to stick out so far. Dunno if that would make any difference in accuracy/setupâŚ?
What laser are you using? How confident are you that itâs shooting straight? (Iâve seen them shoot crooked)
They may be leaving at an angle, but that would be an easy adjustment to point it at the mirror. After that, the light wonât bend (unless youâre close to a black hole ).
How does this compare to a tape measure from one corner to one roller?
I appreciate the ability to square it at some location, because we used to advise just pushing the gantry to the start and turn on the machine. That is a very simple way to get started, but you need to start square.
Cool process. Thinking about something like this would keep me up at night obsessing over LIGO type lasers, mirrors and measurement and that philosophical question: just how square should square be?
I am pretty confident on being square. Itâs the tram that is still bugging me. Is it the bed or is it the axis? I used Ryanâs tram gauge and got it pretty good, but I still have some type of slope to my bed and may just have to machine in flat. It really doesnât affect most of my projects but I did a 50mm round slip fit box and the sides are a bit wonky.
The specific lasers I went with were from this thread.
The light travels straight of course but the line thatâs projected is not quite a plane, it has a slight curve to it meaning itâs a very slight cone shape. I donât think the amount of curvature of the cone is enough to affect me because Iâm not using its straightness over a wide angle.
I could make the arm shorter but I decided to make the arm long enough so I have freedom to put it on either side of the router, so I can make the measurement whichever way is most convenient.
There is another aspect which I omitted from the original post because it was getting long, and that is that the laser line must be not only parallel to the rail but vertical. The reflection of the laser is not all on the same Z plane so if it is tilted then it throws the measurement off. I have a string with a weight hanging to help orient the laser vertically.
I havenât measured the roller-to-corner distances to assure squareness because I donât trust my measurements well enough to have them stack up: x gantry parallel to y side rail, y side rail perpendicular to x side rail, x side rail parallel to y gantry. Although now it could be interesting to see if I use the laser to square x the gantry to the frame and the y gantry to the x gantry, how much difference do I get between the rollers-to-corner distance at each end of the y gantry. In other words do both methods agree in getting the gantry parallel to the frame.
I see this as something like a steel square, but one that works on non-intersecting tubes with extra junk at the intersection. I would rather not use it every time, so I am thinking I will use it to get square and parallel with the frame, and then shim my hard stops so that both motors âpopâ at the same time. Then the relative steps between the motors should be repeatable and everything will be square. Then I can draw lines on the spoil board which for my purposes should be good enough to orient the workpieces.
Can we get one of the cross lasers, mount it on the Z axis, shine it down and put a target on all 4 rollers? If that beam is wide angle enough that could be a super fast way to get all four at once!
Update: Iâm finding the lasers to be too fiddly and I have a hard time trusting that the laser will stay parallel with the tube. Even though it should be kinematic, resting on four points on a cylinder, it is not as stable as I would hope and so I donât feel I can trust it.
I had a different idea, using the mirror resting directly against the rails but using two of them facing each other:
I can definitely see the difference and I think optimize to within one cog of one of the motors, which is 0.64 mm over a span of about 34 inches or 864 mm, which is 0.042 degrees.
Since the belt can move by 2mm increments on the pulley, with 0.64 mm cogs of the motor, you can get smaller than 0.64 mm resolution if youâre willing to relocate the pulley on the belt. If you move forward by 3 cogs its 1.92 mm and moving to the previous belt tooth you get a net movement of -0.08 mm. So you can really get super fine resolution if youâre willing to go to that level. This could be helpful for squaring smaller machines where each cog is a bigger angle.
Maybe a very fine vertical wire can help eyeball whether the infinite hallway curves left or right. Itâs a bit hard to judge from the sides of the hallway but itâs easy to see it move. Resolution is there but accuracy is questionable.
Could you use this method to check whether both belts have equal tension?
If the belt tension is not the same, or more precisely, the distance between teeth is different for the two belts, then the angle of the gantry tubes will not be constant over the whole workspace. If the angle stays the same you can be pretty sure that both steppers travel the exact same distance.
If you donât want to get the belt off of the pulleys for the fine adjustment, you could try skew correction in Marlin to get the last few tenths of a mm correct. Youâd need to line up the mirrors the same way every time then, but not exactly straight. Accuracy will also be questionable with this method though.
This could work but it gets more difficult to observe curvature of the âhallwayâ as the mirrors move farther apart. You could measure the gantry rail against each side rail but then you depend on the side rails being parallel to each other so thats another variable.
You can always install a pen and draw a large test pattern. The downside of a test pattern is that it takes time to iterate while making small adjustments.
@jamiek, whatâs the physical explanation of why a stepper always skips 4 steps at a time? I think I read about it in another thread, but I cannot find it anymore. I also googled for half an hour, but I couldnât find anything.
I was surprised to find out (experimentally) that a 1.8 degree stepper moves 1.8 degrees before skipping a step. But maybe thatâs obvious when I understand the âskipping physicsâ.
If you turn the stepper so that it skips steps, you should find it jumps in increments of 7.2 degrees, which is four whole steps. You should be able to deflect the motor by a maximum of 3.6 degrees and have it return to the correct position.
As for deflecting 1.8 degrees before skipping steps, I think this might be that the resistance to torque increases until 1.8 degrees and then decreases between 1.8 and 3.6. In the range between 1.8 and 3.6 it still wants to return to the original position but it gets weaker. Beyond 3.6 degrees the resistance is ânegativeâ meaning it jumps forward to the next step.
So if you apply a constant force and keep increasing it, 1.8 degrees corresponds to the maximum holding torque and thats where it yields.
I just tried turning the energized stepper with the encoder attached to it. I used a long pair of pliers this time so I have a lot of leverage and can easily feel the amount of resistance I get from the stepper. While turning and feeling the resistance, I looked at the encoder angle.
Basically, everything you guys said is true:
To go from A to another A, you need to jump 4 whole steps: confirmed!
It jumps in increments of 7.2 degrees: confirmed!
You should be able to deflect the motor by a maximum of 3.6 degrees and have it return to the correct position: confirmed!
The resistance to torque increases until 1.8 degrees and then decreases between 1.8 and 3.6: confirmed!
In the range between 1.8 and 3.6 it still wants to return to the original position but it gets weaker: confirmed!
Beyond 3.6 degrees the resistance is ânegativeâ meaning it jumps forward to the next step: confirmed!
If you apply a constant force and keep increasing it, 1.8 degrees corresponds to the maximum holding torque and thatâs where it yields: confirmed!
).
