I am building a new MPCNC and have most of it together but came to a point I am not quite sure about. In the final section, it says to screw down two feet and then square up the others to prevent a parallelogram. If I understand the assembly, there should not be any flexibility to do so given the rail mounts are rigid (that is designed to be held at 90deg). I have the rail lengths measured pretty accurately on all four sides at 23” between the belt attachment points, but my diagonals have an 1/8” difference. With no flexibility to adjust anything. Is that too far off? If so how can it be adjusted without bending something? If it matters I am planning to use dual endstops. My first thought is I need to loosen all the top rail mounts and try and simultaneously adjust square without effecting the lengths, but due to mount design it seems like it would just be forced to a position regardless. Any suggestions?
You should definitely be able to adjust out a 1/8" difference in square. Just do as the instructions say: fasten down 2 feet on the same side, then bump the free side till you get to where the measurement is the same. Remember 1/8" out of square across a 23" square is not much at all so you think you’re going to be forcing it but really you’re only moving things a fraction of a degree.
I have reflected a bit on this question and it seems that the squareness of the gantry makes all the difference, and squareness of the side rails is much less important. Now, if the gantry is adjusted to be exactly parallel with the side rails then it will inherit the non-squareness and that would be bad. But it seems to me that depending how the endstops are positioned and whether they support fine adjustments, it could be possible to have an accurately square gantry even with a fairly severe parallelogram in the side rails. And such a configuration ought to still make accurately square cuts.
The instructions imply that a parallelogram base cannot make square cuts, but I can’t see why an accurately square gantry would also need a square base. Can someone help clarify for me whether this is truly a problem?
Thanks for the replies. I’ll try to further adjust it to get rid of the parallelogram that I have, but I guess I am questioning if it is really even possible given the design of the top mounts. They are rigid and by design seem to force the corner angles when tightening them down so pushing the feet around does not seem to make much sense to me unless there is enough flex in the plastic to allow it to work that way.
Jamie, I can under stand your statement about the gantry and I am pretty sure that is the next battle I have to face. My initial build and check show it is not square and my first quick atttempts to square it did not seem to help much. Perhaps I am being a bit too timid in not over tightening the adjustment bolts. After initial assembly one of the cross rails was not contacting the a bearing and jiggled. I was able to fix that some with the B screws but maybe need to go further. I put it aside till I got the base set up. One step at a time I guess.
There is enough play in the corners to get out of square enough to screw up your cuts. Try to get it as square as possible before tightening the corners. The squarer the frame, the easier it is to square the gantry. Parallelograms will still cut parallelograms no matter how you set your endstops. The stops just make sure you’re the same distance from the corners, they don’t account for the diagonal slant of the tubes.
Just to add a bit to this, it should not be off by much when you build it. Build it carefully as square as you can and try it out before you spend days trying to get it “just right”. Some things do not require any accuracy (signs) others you will want to add things like dual endstops to get repeatability.
Think about it some more. The motion of the gantry will move parallel to the respective X and Y axis, even if the gantry tubes are exactly perpendicular. This is because the motors are attached to the frame, not the gantry. Moving on the X axis, if the X tube is not perpendicular to the Y tube, will include some amount of movement in the Y direction.
Okay, I’m still not seeing it the same way. Let me elaborate and you can tell me which part you disagree with.
My main claim is that the tool head moves parallel to the gantry rails and not parallel to the side rails. Consider Figure 1 where nothing is square:
[attachment file=92740]
If the tool moves purely in the positive X for example, (left to right) both ends of the X gantry rail are fixed, and both ends of the Y gantry rail will move left-to-right. The tool head will move left to right along the X rail, but since it is not square it will also move slightly downward along the Y gantry rail. The Y gantry rail cannot exert an axial force on the tool because it is on rollers. So the tool head move parallel to the X gantry rail when a +X move is made. The angle of the side rails does not play a part in this motion (except maybe the length of movement could be off by the cosine of the angle error).
And likewise a +Y movement will have both ends of the Y gantry rail fixed, so the tool must move parallel to the Y gantry rail. This is irrespective of whether the gantry rail is parallel to the side rails or not.
To reiterate, the direction of tool head movement is determined by the angle of the gantry rails, not the side rails, and the squareness of the gantry rails determine 100% of the squareness of the cuts, while the squareness of the side rails make no difference (within reason) to the squareness of the cuts. (The side rails had better be parallel, or there will be axial loads on the gantry rails, but they need not be square.)
As another example where the gantry is square, and I am claiming square cuts will be made, consider Figure 2:
[attachment file=92741]
In this case, X movements happen to be parallel to the side rails, but consider Y movement. The tool must move along the Y gantry rail. The endpoints of the X rail will move diagonally, and the whole rail will translate a bit to the right for a +Y movement. It’s true the motors and the rail are attached to the sides and will follow their direction, but the tool head can easily roll along the X rail, so the tool will still move purely vertically along the Y gantry rail.
As an opposite example, Figure 3 where the sides are square but the gantry is not. Still, movement will be parallel to the gantry and the fact that the sides are square does not help.
[attachment file=92742]
Finally, I am also claiming that with a parallelogram base, careful positioning of dual-endstops can still produce a square gantry. Figure 5 shows how this would work, with the red segments representing endstop positions. As long as the home position of the two gantry rails are mutually perpendicular, it should be possible to achieve a square gantry (and therefore square cuts) even if the sides are not square.
[attachment file=92743]
I hope this does not sound argumentative. I am hoping if I can clearly articulate my conception then you can point out why this is not correct.
Not argumentative at all. You took the time to illustrate and make your assertion. I appreciate it, actually.
For whatever reason, I can’t get my head back into the space it was in when I wrote my previous comment. Whatever made sense to me at the time, no longer does.
I think you have made a good point, and I can no longer think of a reason why as long as the gantry is square, it shouldn’t matter whether the frame is square. IOW, I think you’re right and my previous post was wrong.
I apologize for being curt in my previous comment. In retrospect, it doesn’t sound as friendly as I would like to be.
Frame still needs to be as square as you can get it. In your exaggerated example above moving the steppers say 50mm positive Y would not get you a 50mm cut. It would be short a bit. Usually this wouldn’t really be an issue, we’re talking a couple millimeters at the most, on a really big table, or really skewed frame, but if you’re cutting parts that need to fit together, say a wood inlay or something like that, it can cause issues. Besides, it’s really not that hard to get the frame to within a millimeter square.
No with a square gantry on a parallelogram base you will still cut parallelograms. The base has to be square to cut a square. In your last example X will cut a straight line Y will cut a diagonal.
I am not sure what the actual issue is. Dual endstops easily square a gantry even if your build is wonky. So dual endstops and a square base and you should be easily golden.
This is what I am having trouble getting my head around. In Figure 4 above, you are saying that Y movement will not move the tool parallel to the gantry rail, and the tool head will instead move parallel to the slanted side rails? I don’t see how this is possible since the Y gantry rail is fixed.
I generally agree that it is not that hard to square the side rails and there is no reason to have it out of square, but I am trying to focus appropriately on the part that is truly essential in getting square cuts.
I also agree that a square gantry on a skewed frame will cut slightly short. I believe the length error would be determined by the cosine of the skew angle, so for one degree of skew (which is huge) the actual cut length will be 0.99985 times the intended length, or 0.15 mm short on a one meter cut.
[attachment file=92837]
Edit: adding this diagram to help convey even though the X gantry will move diagonally (blue), the toolhead should move along the Y gantry rail (red).
Well, dang I think you are right. This is going to be stuck in my head until I am sure. I finally have my machine at the new place so I might have to go stare at it for a while.
Went out to the shop and skewed my mpcnc. I found it’s really hard to square the gantry by hand, and hold it while trying to hit a button on the laptop to energize the steppers! It still came out slightly. :roll: Arrows pointing to corners labeled 1 was my first 200mm square. Came out to within a millimeter corner to corner. Unscrewed the feet on the back side of the mpcnc and shoved that side over almost 3/4". I heard plastic start cracking, so I stopped. Second set of arrows point to the second square drawn on top of the first. It’s almost 4mm out.
Sharpie lines are where the foot was before I unscrewed it. I finally have a reason to install the new feet I printed months ago… Yea, I’m a procrastinator. :oops:
Thanks for all the discussion! I wanted to post back and say I figured out my issues with the build to get the frame square. It turns out my printed parts for the top corners had just enough roughness on the cradle surfaces to make the fit to the rails a bit too snug. This made it feel that I would be cracking the plastic to push the frame around even a little bit. I sanded the cradles slightly so the rails had a slip fit, which gave it that fraction of a degree needed to make adjustments. Once there it was very easy to just square it up as intended.
Thanks again!