+ Join to Post a Comment

Print Bed Flatness

A common issue with most 3D printers is maintaining print bed flatness. Aside from adjusting the flatness with the four corner thumb screws, planarity is also critical. Print bed warpage is hard to correct, and that's one reason why some people mod their printers with glass beds (not to mention adhesion issues).

I performed that mod on my Mega since I had some flatness issues, and I wanted to use smaller nozzles, which required critical flatness (thanks BETLOG for the heads up!). What I found was interesting so I thought I would share my experience.
I ended up buying this glass and thermal pad:

However, after installing it and performing the flatness adjustment, it too seemed just as warped as the original aluminum bed. I could get all four corners adjusted perfectly, but the center was almost .2mm higher. What gives? Some people were reporting that their glass was coming in warped from the supplier, so I checked mine with an accurate straight edge and it was dead on flat, so where is the warpage coming from? I thought maybe the paper clips I used to hold down the glass was warping it, but I also checked it installed and it showed perfect flatness, so how is the center rising up .2mm?

As it turns out, there was a bend in the two y-axis shafts, which I found odd since they are made from high precision steel. Well what I discovered was the shafts on their own were pretty straight (after removing them), but what was causing them to bow were the end mounts. The end mounts don't have a very large mounting surface so they can easily tilt if there are any surface variations in the sheet metal cover that they are mounted to, and in fact, they mount very close to the edge where there is a bend radius. You can tell if the alignment of the end mounts are off by loosening the four set screws (one on each side of both ends) and attempting to slide the shaft back and forth. If everything is aligned well, they should easily slide and rotate. Mine didn't. What I did to fix the problem was loosen the end mount attachment fasteners (to get the jammed shafts to slide out) and take a 5/16th drill bit (slightly smaller than the shaft diameter) and insert the smooth end into the end mount hole. Carefully use the drill bit to bend the sheet metal slightly, correcting the tilt. It will take several attempts to bend and check free movement of the shafts. Here is a picture of what I'm talking about (exaggerated in Photoshop to illustrate the tilt):

Once everything was straight, the bed was much flatter and I could successfully print at .05 layer thickness. I think the Anycubic design team should consider widening the end mount footprint so that it would be less likely to tilt with small surface variations, or check for free rotation after they are mounted. I hope this helps anyone with print bed flatness issues!

I haven't had this issue.

But, as with anything that is manufactured there are going to be anomalies in some parts. Since the aluminum plate is metal it really shouldn't be an issue to re-form it by applying a bit of force with a bottle jack (hydraulic). If you know of anyone with a press or have one yourself then just push the center down until you have it flat. Why wait for a replacement part? 3d printers have always required a certain amount of DIY. Lots of cheap bottle jacks out there.

I'm not sure what you mean, the bed aluminum plate or the chassis? The bed aluminum plate is perfectly flat so there's nothing to reform.

I've just heard back from Anycubic - they're sending out a replacement rod and 2 brackets: my left rod seemed fine and moved freely within the brackets, but right rod wouldn't move at all - stuck tight! I'm not expecting it to be a miracle cure and have resigned myself to not ever printing a full 200x200 object. At present I can get about 140x80 reliably printing if I use a raft. so will hopefully improve on that.
On a side-note, I've got an infrared measurement gun now and found a 25mm strip on the left from front to back sits at 30-40ºC when the rest of the bed is set to 68-70º.

FYI: most of those guns work fine on the bed, but when the temperature range is higher, like on the hotend, they are very inaccurate. Of course it's also much more difficult to hit the hotend with the actual measuring beam (as opposed to the laser) as well, but that's not the only problem.

I took a lever and check the level of the mounts .. made them all level and that seemed to fix things just fine for me.

Just got my i3 from GearBest. It is REALLY bad. The center is 0.59mm high. Sent support an email and they pointed me to this page.

I loosened the set screws on the rods and they would not even rotate. Had to loosen the support screws to get the rods out. They are bent but not huge. Tried to put the rods back. The side with two bearings are miss aligned so much the rod will not go back. Will need to remove the bed and adjust bearing mount just to get back together. Sent an email to support with pics and asked for a replacement base that has been CHECKED. We shall see what they have to say. For a machine that should have been unboxed and start printing I'm not happy.

I got my printer from GrarBest also .... There is no way I can level the bed. I have added the Ultrabed and made the adjustments to true the rods in their holders. Still I can never get the bed level. It seems that no matter what I do the left rear corner is always jammed up against the nozzle. Also I have a granite counter top that the printer sits on and there is a >2mm wobble.... it just wont sit flat. I wonder if the chassis is bent. I also have contacted Anycubic with my concerns .....and pics. wish me luck!

One thing you can try is to add a shim (the thickness of the offset at the left rear) under the left front rod mount. Then, try leveling again. What you are describing sounds like the result of the rods not being parallel.

I tried to dial mine in using measurements of bed height rather than free movements of the shafts... and it did kinda work but it made my head hurt at times. The whole y axis seems like a pretty poor design, I might have to make an upgrade at some point.


aaaaannd... did anyone look inside the case at how the mounts are connected? It's possible that the issue could be more efficiently addressed by simply adjusting the mounting bolts or mounting surface itself, or maybe adding a shim.
I'm just guessing though.

The other thing that probably bears a mention: is that this issue may be a non-issue for many people, and is only being interpreted as one by assumption. No bed is ever perfectly flat, and the people focusing on it are often looking for it in their first 3d printer, and subsequently highlighting any deviation as critically important. Obviously perfection is ideal, but in most cases the typical flatness deviation is fairly inconsequential.
Given the obvious inaccuracy of peoples measuring techniques, a surface deviation of half a millimetre is a fair bit, but possibly not serious if it's really anything less than that. ironically this is what 9 point auto-leveling is for. ..even if we can't do that on the i3 mega.

Putting any pressure whatsoever, or even just general vibration to the machine can and usually does cause the unpowered Z axis motors to slouch... ie: the x carriage crossbars drop. Almost always unevenly and imperceptibly.
It would be far more reliable to measure from the plate metal on the case somewhere that isn't a moving component.

Nice video, thanks for sharing :)

Yes, excellent video indeed! I think yours (Funkster's) was worse off than mine, and I agree about making those mounts bigger or perhaps one piece. I too had a slight residual variance even when the mounts were eventually trued up. My shafts had a vary slight, but permanent bow left in them (you can see it slightly by rolling them on the table), and I used that bow to fine tune the residual variance by orienting (rotating) them in the mounts, pointing the bow in the direction I needed to compensate. If you need just a little compensation you would rotate it only a few degrees in that direction. If no adjustment was needed, you would just leave them with the bows outward or inward (not up or down) so they wouldn't affect height at all as the table moved. The amount of bow was not enough to bind the rollers so it seemed ok.

Sneaky! I didn't take the shafts all the way out to check for straightness, but turning them in place and measuring the change sounds like a reasonable experiment. I did notice that there was a side-to-side deviation when unhooked from one end, which could have been due to bend in the shaft - I assumed the mount was misaligned turnwise also. Alternatively it might be that the bearings on the bed are not the correct distance apart!

Given the robust construction of the rest of the machine, the flimsyness of the Y axis is pretty poor. I've got a feeling that I might be turning the handles on my friend's Bridgeport again some time soon...

Can't believe this , had the same issue myself and just worked out myself that the problem was the rails brakets ( all misaligned) and now after reading this I've got the confirmation of it.
Little bonus was that not knowing about the reason behind earlier ive contacted anycubic immediately and requested a replacement bed . Got it and found out about the real problem .
Now ive got a spare bed

Thanks for posting this, I checked and this was exactly the problem. I unscrew the 4 ends to give some relief and it was almost perfectly flat, but after tightening again, the problem appears. So the only solucion is what you have described.

You're very welcome. This was a tricky one!

pour un meilleur contact avec ton lit chauffant avec du PLA utilise de la LAQUE pour cheveux EXTRA FORT ça marche super bien
je nettoie avec une éponge et de l eau

Oui, j'ai lu que cela fonctionne très bien, mais je n'ai pas encore eu l'occasion de l'essayer. Merci pour la suggestion!

I have just stripped down my printer and sort of see what you mean. I have also opened up the bottom case to have a look inside and replace a noisy fan. When removing the two 8mm steel rods (362mm long), they are both bent, and when removing them from the holders at each end, they are immediately forced down as if being held at a wrong angle under pressure. I intend to replace the rods tomorrow but have two options to solve the out-of-alignment holders. I can either unscrew the holders and add some plastic washers to allow enough flex, or I could use the old bent rods to prize up, and hopefully not break, the clamps so that there is no downward or wayward pressure on the new rods. I prefer washers. It would be interesting to know what tolerance the bed is supposed to allow. Obviously the flatter the better. I would have expected 0.1 and below. Mine is all over the place and need to replace both the heated bed and the heated bed supporting plate, which is flimsy enough to be bent out of shape by the spring loaded thumb screws. I have bought a 4mm thick piece of aluminium and it does not bend if I stand on it, but it did not arrive flat! I think the next step is either a garage head skimming machine or a new bed support from a service like at tehnologika.net. I have emailed them to see if they will design one if i supply measurements. I tried glass, but that was being bent into a dome by the clamping to an already bent heated bed.

Yes, the overall design of the machine is very good, but there are a few areas that need attention.I didn't use washer/shims to correct mine, I used the bending trick I mentioned (with a 5/16 drill bit) and it worked out well. You can dial in precisely how much to bend the sheet metal (although it is very subtle) and it felt like the end mount was plenty strong to take the load. The only drawback when using washers, is you are going to need a collection of washers with varying thicknesses in order to pick the right amount of shimming.

Regarding the glass option, I used a 3mm borosilicate glass, which is relatively thick and very strong (also heat resistant), so it really didn't change shape much after being clamped (my bed was also slightly domed up). Additionally, if you use the silicon thermal pad underneath, it is soft enough to compensate for the uneven aluminum.

Also, you should assess flatness with a good straight edge and not rely on the hot end clearance with the bed. As you see from my post, the bed appeared to be .2mm high in the center when checking clearance with the hot end, but measured perfectly flat using a straight edge. The bowed rods was deceiving you into thinking the bed was warped when in fact it was the shift in travel as the bed moved along the rods.

FYU: aside from the fact that one side is matte, and not as sticky as the other shiny side, I bet these would be functionally the same, for cheaper, than the 3dprinter-specific silicone stckysheet.
I'm actually planning to try printing on it to test if it's a viable surface that easier to remove really sticky prints from. I imagine it will fail, but should be interesting, especially with 0.5 or 0.8mm nozzles.

Those look interesting and worth a try as a printing surface. The silicon sheet I referenced is actually a thermal pad that is designed to go between the aluminum bed and the glass for the purpose of conducting heat into the glass.I don't think you can print on it directly since it is not stiff enough. Other people have had success using thin sheets of PEI ultem plastic for a printing surface, which would go on top of the glass like the suggestion you had.

Yeah, I know. But still, it looks like effectively the same thing for a lot cheaper.
...and if filament will stick to it at layer 1.. then it might have some potential as a surface.

Would it be easier to simply drill the sheet metal hole larger?
I'm having trouble understanding your fix description. But I imagine if I started poking around in that connection it would become obvious.
( Given that I have another printer that I can modify, I try to avoid wanting to modify my mega - because I can't modify the firmware if that becomes necessary. )

I added a picture to my post to help visualize the issue:

The drill bit is used only as a means to apply torque to the end mount in order to bend/straighten the sheet metal underneath and correct the tilt. If the misalignment was in the other axis, then yes, opening up the hole a little would allow me to rotate the end mount into position.

Ooooh. For some reason I was visualizing the X axis mounts. Ok, so that makes a lot more sense.
iirc the box has threaded tabs welded into the other side for all of it's M3 surface connectors, so yeah, it should be pretty robust.