Z-Offset issue: too close in the middle but too far on the edges of the bed

Hi all, I hope you can help me with an issue with my J1 that has been haunting me for the past few months.

When I use the full width of the print bed, the printer prints too far off the printbed on the left and righthand side on the bed, as if the glass bed is somehow thicker in the middle.
The result is that I can only get a decent first layer if I print on the middle half of the bed.

Some more info:

  • The issue only shows up along the X-axis. It doesn’t matter if the printer is printing at the front or at the back of the build plate.
  • The initial layer height of 0.28 that is Luban’s default is not enough to make it work. If I use the full width of the build plate, I get elephants foot and a skipping extruder in the middle, and loose or almost loose threads at the outer edges.
  • I have two PEI Glass plates. This happens on both of them.
  • It doesn’t matter if I print on the textured side, or on the smooth glass side. It happens on both, although the textured side works a bit better, because the filament can squish in the valleys of the texture more.
  • It doesn’t matter if I print with the left- or with the right extruder. They both show the same behavior.
  • There seems to be some play on the carriage that holds the bed. Despite tightening all the screws, that play is still there.
  • When I print on the center of the bed, prints otherwise look very good. At least as good as my other printer (a somewhat modified Ender 3).
  • I use Luban as the slicer.
  • I converted the printer to a J1S, with the new fan and backplate. Other than that, it’s completely stock.
  • The printer is less than a year old. I believe the problem has gotten worse over time, but when I first got this printer, I didn’t print as many large things or used copy mode as much as I’m trying to do now, so maybe I just didn’t notice it as much.

Tonight I did some extensive maintenance on the printer; cleaned the entire machine, tightened the belts, cleaned and relubricated the linear rails, linear rods and lead screws. I also checked and tightened all the screws. The printer makes slightly less noise when moving, but other than that this made no difference.

Might I have a linear rail that is not straight? Could the play on the bed-carriage be the culprit?
I’d greatly appreciate any and all ideas to troubleshoot this. Thanks!

@3dbjorn Hi! The problem you describe is not new to any alignment of small CNC machine so to 3D printer as well. I can hardly imagine bed should be a problem in a kind-a-curved motion of X-axis over it as this is still piece of rather precise glass. But - you can place a good straight ruler over glossy surface of it and see if there is any light gaps visible. In my professional work we use so called precision straight edges for this but any ruler with visibly straight edge should be of help. Having elephant feet in the middle of build plate you should have convex surface there.
Even you have a play in bed carriage, gravity should force it down into more or less stabile position. Even this position is not well aligned, it still should lead to linear error from one end of table to the other but not curvilinear (of course if we assume glass plate is straight indeed).
I would rather agree to your worries about straightness of linear rail. Error here should not be large to facilitate elephant feet in the middle - 0.1mm is already sufficient. While bolting down 12 or 15 mm linear rail to carriage is not problem at all to bend it sideways in the middle for that extent - one have been doing this will agree.
Depending on what tools you have available you can check if this might be a problem. You need to go down with Z axis to let dry and clean nozzle heaving gap of about 0.5mm till glossy surface of build plate. Moving from right to left or vice versa you shall measure actual gap between tip of nozzle and glossy surface. The best is to use feeler gauge for this but set of papers with different thicknesses should do the job too. With papers being used you can not get absolute value of gap precisely but for sure you will have understanding of current situation.

Hi @Karass3D, Thanks for your reply, that helps a lot. I had thought that the idea of a warped X-axis linear rails was far fetched. But I guess it is not.
I ordered a set of feeler gauges and will do some tests.

If the X-axis is warped, is the fix as simple as undoing all the bolts on one side, and redoing them while applying a little bit of pressure towards the bed?

Hi @3dbjorn !
If we suppose warping to be a problem (you will soon be able to check it, great!), it might come from several reasons - native problems in production of the very rail (rather rare but I have seen on 12 and 15 mm rails), assembling mishap, problems on base surface. To eliminate all suspected cases a full disassembling of particular linear rail unit and inspection should be carried out and this is rather complicated task, avoided as far as possible :slight_smile:
If measurements indicate bow in linear rail, I would do the following:

  1. Release for 1.0-1.5 turns all bolts fastening linear rail except the most left and right (rail remains securely bolted on both ends. Check again distance from tip of nozzle till bed surface in tree places - both ends and in the middle. Hope rail has straightened back. If so, gently secure middle bolt and do measurements again. Gap between nozzle and table shall remain about the same as before tightening the bolt. If it goes back again to initial problem, there might be something wrong with alignment of bottom of rail against base surface - some debris in between or similar. If tightening middle bolt goes well, gently tighten middle bolt of left half and middle bolt of right and so on until all bolts are tied down. Important here is not to overtighten! For M3 bolts tightening torque should be about 1 N*m.
  2. The way described above is not the typical for proper alignment. The better way is to tighten bolts from one end to the other in a row. In particular it takes more time and effort. Normally it should be a datum surface side of the rail is aligned against. I am not by J1 now for several days, can not check axis construction. I would do it like this.
    2.1) If X asis body has a datum surface (simply a shoulder) rail is set against to. Slightly release all bolts holding the rail and from one end to the other bolt by bolt gently push rail sidewards to the datum surface at place of bolt and tighten the appropriate bolt. Check result - rail straightness - at the end.
    2.2) If rail is on the free surface - can flote sidewards. Than instead of datum surface it has to be aligned against print surface. To do alignment just prepare a piece of hard plastic or non-ferrous metal in thickness of 0.5-2.0mm in size of a coin - this will be a distance gauge. You can use also any suitable thickness on your gauge set. Move Y-axis as deep as you can comfortably reach by hands.Move hotend to middle of X-axis span and place distance gauge under nozzle and lower it till it gently touches distance gauge. This is your reference point. Slightly release all bolts on linear rail (but not let it become loose). Start from one end of axis, tightening bolts gently (!) pushing axis against your distance piece/gauge. You should push very gently, yet firm - not to introduce too much force that build table is pushed downwards - this is why better to work at far end of Y-axis as table has cantilever build! So follow to the other direction of X-axis tightening bolt by bolt ensuring still the same distance between nozzle and table surface. After all bolts are tightened, check final result. Now you should have your linear rail as straight as flat your table is.

Sounds a bit complicated but once you get used to these procedures, not that scary indeed.

Hi! No, that actually doesn’t sound all that complicated. Thanks!