Primo gantry won't square with core attached

Mostly Printed CNC - MPCNC Troubleshooting
1

Hi all, first time MPCNC builder, first time poster here. I cannot get my gantry squared during step “7-Squaring”. I would appreciate any suggestions and help.

Background: So far everything in the build has gone very smoothly. Corners on the feet measured less than a millimeter on the diagonal, and almost indiscernible difference on the x and y lengths (which you must also measure otherwise you could get a trapezoid). The trucks went on well, they move slowly when tilted 45 degrees, and they squared very nicely following the instructions (I had to correct about 3mm). The core when together well with minor(?) exceptions. When I first put in the rails, one of them had a small wiggle between one set of bearings (no nuts on the outside bolt). I was very cautious not to over-tighten and the wiggle went away after slightly tightening the nuts between the clamp and core body. The only other (noticeable) exception was the rails slide more easily than the trucks, but they still have reasonable(?) resistance.

The Problem: While trying to square the gantry with the core attached to the rails, I get no movement when tightening or loosening the adjustment bolts.

Root Cause Analysis: I re-read the instructions and previous steps many times, and I don’t find any discrepancies in my methods or build. I went to the forums and found a couple similar posts. The closest problems were resolved by disassembling back to previous steps and redoing the process. I removed the core, verified my corners, loosened and re-squared the rails, everything up to this point matches. I put the core back on and see an immediate deviation of 5mm. The trucks closest to the origin (0,0) are too close to the origin, while the trucks on rails opposite the origin deviate away from the origin. In both dimensions this means I need to tighten the inner bolts. I started with all four bolts loose. VERY loose. Not even seated to the plastic with a visible gap. I tightened only the inner bolts starting at as 16th turn each, then moving the core in figure eights, and measure. No difference. Eventually the bolts become quite tight, and the tension on the gantry rails is so tight the core only moves haltingly. At this point I stop because it feels like any more and the clamps will be compromised. Still no deviation in the square measurements. I loosen the tension bolts and start over with the same result. Then I think maybe the tightness of bolts between the core and clamps may be wrong, so I try the above procedure again with them both quite a bit tighter and quite a bit lo0ser with the same results, actually the measurements are now permanently worse. I also tried loosening and tightening the gantry pipe clamps with no difference. Then I try resetting everything and adjusting the truck bolts instead of the gantry bolts. It seems to make a very small difference, but the trucks ride worse on the rails as a result, almost as if they are fighting against the core if that makes any sense.

Current State: I reset everything back to the original instructions, attached a pencil and drew a rectangle (parrallelagram). The sides are very nearly equal (around 0.5 millimeter, but the diagonals are off by a quarter inch. The difference between rails is about 5mm for both x and y.

Potentially useful information:

  • I measure with calipers from the edge of the top corner to the edge of the truck.
  • My core is printed in PETG (Accidentally bought it based on reviews for structural integrity, thinking I was getting PLA, so decided to just try it, and it’s stronger than my PLA parts.)
  • Clamps are printed in PLA
  • My PLA and PETG calibration prints are around 0.05mm per 50mm average, w/ variance around the same on the measurements.
  • In the photo, the yellow box exaggerate the deformation of my workspace. The green box shows the origin. The pink circles show the bolts I am tightening to try and square the gantry. The non-circled bolts on the core clamps I have tried loosening also. Please ignore all the extra scribbling. The extra lines in the area kitty-corner/diagonal/oblique to the origin are mostly due to variation caused by drawing the box without realizing the belts were causing issues.
  • If possible I would like to avoid reprinting the core. I’m more apt to waste time modifying firmware and add calibration fudge factors.
  • If possible I’d like to avoid speculation on what’s wrong.

Help is greatly appreciated.

gantry_image
gantry_image1152×1350 174 KB

2

You are not going to get a definitive answer. I also have problems with one axis always being off. Did you check if your whole build is square and not a parallelogram? Because the measurements don’t change much if it was.

3

@Tokoloshe, I verified the feet and corners are square, and the trucks are square without the core. Once the core goes on it goes out of square by 5mm on each rail. I realize every build is different and no one is going to have a definitive answer. I’m hoping there might be people with experience to suggest some specific tweaks that might help my target my troubleshooting.

Thank you for your response.

4

You are welcome.

Did you print the new or the old version of the core and clamps? If you have the old version you need modified clamps, otherwise it will rattle. There was a minor mistake in the files.

5

I’m using the C version files posted a couple weeks ago. I don’t know if these have the fix included.

6

C should be okay, they were never broken in the first place. :slight_smile: I am officially out of ideas. I am sorry. :frowning:

7

If you take the core by itself, and put conduit through both X and Y clamps, are they square?

1 Like
8

Also, what happens if you remove the z?

I wonder if the flexibility in the petg plays a role.

9

@kd2018 I removed the z, seems to adjust a tad, maybe?

@Strider_Matic I will try that when I have time tonight. You gave me some new thoughts on measuring the core by itself, reverse mounted, etc. Tried rotating the rails and didn’t see a difference.

10

I measured the square of the rails in the core by itself, and the 5mm deviation is present. I tried many variations of tightening, loosening, adjusting and measuring. Ultimately I forced the rails into square with a speed square and tightened the bolts and bearings to hold in that position.

I also scrutinized the trucks before putting the core back on. Even though the measurements were square, I found a slight amount of twist where the lone bearing on the min x and min y could come slightly off the rail. This tiny, tiny wobble was also biasing the deviation towards the origin. I did the same thing here; I loosened and retightened all the bolts on the trucks, holding them in the desired position to eliminate the wobble.

Finally I put the core back on the rails with everything together. My x axis was almost spot-on, and the y-axis was about 1mm off. I again tried adjusting only the tension bolts on the core, but still didn’t see any movement. I tweaked the clamp bolts and bearings one more time and it came in to less than 0.5 mm deviation. Everything rides smoothly.

Thanks everyone for the suggestions.

2 Likes
11

I am also having a hard time squaring the core. Could you run me through how you used a speed square?

12

Hi Mike. I took the rails off and put them in the core by itself. The corner opposite of the origin, on the plus X and plus Y side, is unobstructed. I put one edge of the square against the x rail, and the other edge against the Y rail. Since the rails are offset, one of the edges will only be in partial contact, but it’s enough to see a gap when the rails are not square in the core. In my case the angle of the rails was less than 90 degrees, so I had to force the speed square into the corner to separate the rails until they were square.

1 Like
13

Thanks! I’ll give it a try!