UM2 Bed Levelling - Mission Impossible?

[uit 14]

 

Spirit, thanks for your input on this thread,

 

looks like you were able to deal with your issue by applying this workaround.

regarding the feet I already printed spacers and levelled the chassis to my table, so that it is at least standing an all of its feet and is not wobbling and moving around, but it had next to no effect to my bed-levelling-issues.

 

your second part of the fix, which involves the spacer is actually the same approach as the washer/nut-combination under my left-rear-side, which made this trouble corner less unusable. but i cannot get my rear-right corner to the right level. the tolerances are too high, so that the levelling-screws are to short to compensate. furthermore the bed seems to be under tension, and I don't think this tension mixes up very nice with the hot/cold cycles and the mechanical movement of the bed.

 

 

it is what it is - a messy workaround.

 

 

Nearly a year ago, i ordered an UM2 because I don't want to mess around with basic calibration and tweaking on an Ultimaker Original. I wanted a well calibrated, "plug and play"-type of printer as promoted.

 

The first Support Ticket I openend on this specific Issue was December 16th 2013.

between this date and today there was a lot of Support-Ticketing, documenting, narrowing down the issues to the bent chassis, assuring I get a new Printer, some hotfixes from my side and even a repair by ultimaker which made this specific issue worse.

 

All I wanted was to print architectural models, which have a base size of 20x16 cm to 20x20cm which has driven my decision to an UM2. Since 9 months I'm trying and failing, including 2 superclogged nozzles related to the bedlevelling, days of wasted time, effort and money.

 

[goligo 0]

I had also had some issues with leveling. In the beginning I just had small items located in the center which all printed very well, but with larger parts I noticed that the left side seemed to have more distance and the first layers didn't stick properly, although the bed was properly leveled using the paper method recommended by Ultimaker.

After some trial and error I discovered that there is a noticably diffence in leveling dependent on the heat of the print bed. So while I was perfectly leveled with a cold bed, after it was heated the left side did get more distance and large prints didn't work well.

So my advice is, first heat up the bed using Maintenance -> Advanced -> Heatup buildplate to the temperature used for your print material, then do the guided leveling using Maintenance -> Buildplate.

 

[gr5 2,270]

@uit and @spirit

What you have done in effect warps the glass so that instead of the glass being flat it is now U shaped looking at it down the diagonal.

This seems like a reasonable fix although it does put some tension on the glass.

An alternate fix is to fix the actual problem which is that the top 4 corners of the UM2 are not flat - they are not in the same plane. These top 4 corners set the movement plane of the head as it moves around. If these 4 corners are not in a plane then you can't match the glass to the gantry without warping the glass into a U shape like you did.

More specifically it's not the top 4 corners but the locations of the 8 bearings embedded into the walls of the machine - the 8 bearings that hold the 4 thicker rods in the gantry.

 

[randyinla 14]

I too have the same leveling issue, only my low corner is the back left. I first posted that I had this issue a few days ago. Sandor had seen my post and contacted Chris McAdam @ Ultimaker. (Thank you, Sandor!) I received a nice email from Chris with instructions on how to re-level the z-stage and a link to the Assembly Manual. I was just about to start the process of placing three calibration objects under the build plate (which seems to be flawed, making an assumption that the base of the U2 box is indeed level?) when I started reading this thread.

The comments here prompted me to check if my U2 box was square and it is not! The front right corner is up off the counter surface approximately 1mm (the 1mm shim I use for bed leveling juuuuuust fits under that corner and prevents rocking). I took it into a kitchen with a lot of flat stone counter tops, in case my cubicle desk wasn't flat enough, and the front right corner is up off the surface there as well. I then rotated the box 90 degrees. Same issue with that front right corner. Rotated another 90 and another, and in every position, the front right corner footing was not making contact with the table top surface. If the counter was not level, then it seems that the raised corner would always be whichever corner happened to be in the front right position.

I have a printout of the Assembly Manual. Should I loosen the appropriate screws around the entire frame, put weights on the four corners as suggested previously in this thread to ensure all four are touching the counter surface and then re-tighten? I would then probably go through the z-stage alignment, using the three calibration objects under the build plate.

 

[randyinla 14]

Just re-read through the thread and I don't think shimming just the underside of the U2 box with printed adjustable feet or paper would do anything for the internal alignment. Reason being that the cube is out of alignment with itself, not that the cube is perfect and is just not flat on the bottom. I'm assuming the (laser cut?) walls of the cube are the exact right size and shape. The fact that my cube wobbles tells me it's not a true cube. By shimming under the high leg or shimming to lift under the low leg to stop rocking just rotates the out of alignment cube to stop wobbling on the table. Like Gr5 suggested, it would seem the front and back X rods would still not be parallel to each other.

I'm going to loosen the frame screws and see if it just naturally shifts and re-aligns itself to my counter top. Since so many people seem to be having this issue, I wonder if some U2's were all assembled on the same surface at Ultimaker that happens to not be perfectly flat?

 

[Blizz 53]

Randy: Keep us appraised if you want. I have the same problem with mine (not with the buildplate but with the cube'ishness).

 

[randyinla 14]

I loosened all the screws around the edging of the frame about a 1/4 turn and gently bumped my fist into the top 4 corners while pushing down all around. Within a few seconds of this, that front right corner stopped wobbling and all four seemed rock solid. I tightened the top corners first, starting with the front right, back left, front left, back right... testing wobble after each screw. If a slight wobble came back, I would loosen that screw again and bump my fist in each corner (tapping sideways and downwards) until it went away again. It was rock solid on the kitchen stone counters, rotating in all directions. I am back at my desk now and the wobble is gone. At first, I could still slide a single piece of printer paper under the front right corner with some resistance, but sliding the printer back towards the center of my desk stopped that from happening. It is MUCH more stable than before.

I then ran through the menu-based paper leveling routine and printed a large, thin item that I've never been able to print before due to it warping itself off the build plate and it made it all the way through without coming up. For the first layer, done at .3mm, the lines were flattened out towards the front and still more like laying on top of the plate towards the back left, but a little more 'pushed in' to the tape this time. Also, I didn't have the blue tape back when this piece was failing to print... so it might be more the tape than any adjustments I made to the outer box.

I plan on unscrewing the 8 screws to the linear bearings, re-leveling the z-stage tomorrow and printing it again. Might bring in my air bubble level to check how the rods are aligned too.

My belts don't seem to be tight enough either. I saw someone's youtube video on belt tension for the U2 and they had a musical note when plucked. Mine were MUCH lower in pitch. I know this has nothing to do with my back left lowness of the build plate, but might explain some patterns I see in the side walls of my objects sometimes.

 

[dholi2k 0]

I've actually had the same issues described in this thread when I first started using the UM2, but I've implemented a different fix using two parts.

The printer that I have has the issue where one corner (front right in my case) was ~4mm higher than the other three corners and so the printer wobbled when printing. I didn't want to completely take apart the printer to fix the issue, so I came up with an alternative fix instead. The first part consists of length adjustable legs that attached to each corner to stabilize the printer:

 

 

 

The second part was the actual fix for the heated bed and involved a printed spacer wrapped in kapton tape to prevent the spacer from melting/warping. By wedging the spacer into the back left corner between the PCB and aluminum plate, I've turned the three point leveling system into a four point one with one corner being set at a fixed height. By tightening the rear screw, the height of the back right corner can be adjusted without affecting the back left corner. This adjustment is also independent from the front corners. It takes a bit of time to get it levelled this way, but once the bed is calibrated it stays pretty consistent from my experience (I print at 0.05 - 0.1mm layer heights).

 

 

Wish there was a more permanent fix, but it doesn't seem like there is one without taking apart and reassembling the printer.

 

Dear Spirit,

I have the same issue you are having with the back left corner not being level with the rest of the machine. Would you be so kind as to give me more information on the spacer part you printed and if possible, send me the file so I can print one myself? I'm a newbie to this so any help would be appreciated. Many thanks.

 

[spirit 0]

Sorry, I don't think I have the original file of the spacer that I'm currently using. It should be fairly trivial to make one though as it's just a rectangular box ~12mm in height.

 

[gr5 2,270]

I loosened all the screws around the edging of the frame about a 1/4 turn and gently bumped...

 

Nice work! This is basically what I would have done. Don't know why people didn't try it sooner. You make it sound easy.

If this fixed it only part way, then do it some "more". Put a shim (thin) under a leg to warp it even further than you already have and repeat the previous steps.

These printers aren't being assembled this way of course - UM manufacturing isn't that dumb. The problem instead is with the shipping. If you tilt the box up onto a bottom edge and then sit on the opposite edge... and if you weight 400 pounds, then you can easily cause this problem where one leg is high.

 

[randyinla 14]

I am very happy to report that my build plate seems MUCH more level than previously possible!

After adjusting the box's squareness, I then followed Chris's suggestion; took off the back plate around the z-stage linear bearings, loosened the 8 screws on the linear bearing plates (again, don't take any screws out) and realigned the build plate to the now (more) square inside bottom. I didn't have any calibration thingies or dice or anything else... then it dawned on me that the three thumb screws provided by Ultimaker for fine tuning the leveling are RIGHT THERE! I unscrewed them and used them as the three calibration devices. Worked great. Put it all back together and ran through a typical leveling routine and printed that hard to print piece a second time (it's a propeller protector for a quad-copter that spans the far back left to the very front right of the build plate). This time the back left first layer looked identical to the front layers! Flat, slightly pushed into the tape, not laying on top. I have printed three perfectly so far.

If your box is wobbly, with one corner obviously rocking a millimeter or more, do not simply print out adjustable feet or put a piece of folded paper under it. It needs to be adjusted to be made square and then the z-stage re-aligned to this new squareness. It is NOT difficult to do B) The only thing I would suggest you watch out for is to NOT loosen the screws that hold the stepper motors in place. They are close by the frame screws and are the same type of screw, but there will be four of them in a perfect square, outlining the stepper motors. I'm not sure if loosening them by mistake would cause the motors to move or not, but I do know they don't need to be loosened to adjust the squareness of the box.

I also replaced the noisy tiny fan just behind the print head so now I can't tell it's even on except for the lighting!! It's like a brand new machine! Looking forward to tightening up the belts and making a front door for the box!

Thanks for everyone's suggestions in this thread (and other threads).

 

[Nicolinux 288]

Hi,

I noticed that my UM2 has the same problem in this regard. I loosened all screws around the base and tried to push/bump it into form but nothing changed. The printer still wobbles but only when it is turned sideways. At first I thought that my surface was not levels, but after trying almost all available surfaces in my apartment, I think it is the printer.

Do you have any ideas as to what could be wrong? I loosened _all_ screws on the outher frame except for the feeder housing and the motors. Should be enough right?

thx.

 

[twistx 0]

I was just dealing with this bed calibration issue over the weekend.

My home position on the bed is lower than the back right position. After pulling the bed apart I determined that my heated aluminum plate is not completely flat for one, maybe from colliding with the print head, and the corners have different degrees of height. The middle seems to be lower than the rest.

The biggest culprit though was the rear spring on my UM2. It is badly curved, effectively angling the rear of the bed in whatever direction its sitting in since there is only the one point creating a fulcrum. I rotated my rear spring so it curves on the x axis rather than the y and was able to get the height between the back corners to within 5-10/100ths of a mm. The front is no issue since there are two points to calibrate with.

I understand the 3 point system is supposed to be easier but it seems very susceptible to defects especially when the Z platform and aluminum heating bed are so flimsy. The only rigid part is the glass.

I didn't have this issue on my ultimaker original which had the 4 points to calibrate the corners with and a thick, rigid piece of acrylic. Even though the wood of my Z platform was extremely warped (off by 5mm on the front right corner) I could still calibrate just fine.

 

[randyinla 14]

Nicolinux,

When I loosened all of the screws around the entire box, not just the base, I could feel the box's squareness changing depending on which corner I put pressure on. I didn't loosen them to the point of almost falling out, but enough where the alignment of the sides/top/bottom could be manipulated slightly. All I can suggest is to try again and maybe get someone to help hold it down against a flat surface while tightening? I would think if you tighten the screws while being held tightly against a flat surface, you'd have a good chance of retaining the alignment. If you can reduce or eliminate the wobble of the cube on a flat surface, then you must loosen the 8 screws around the two linear bearings for the z-stage (behind the back white panel with the logo on it) and go through the initial z-stage alignment. Then do a normal leveling using the paper.

To be clear, I think the printed adjustable feet are a great idea if the surface the U2 lives on isn't flat. But not as a fix for a wobbly U2 on a flat surface.

 

[MoneyMike 0]

I just got my machine and I also have this problem. One thing different about mine is if the top panel of the case is off, then its square. As soon as I put it back on, it becomes wobbly. I thought about taking it back off and using a dremel to notch out the alignment slots a little for some play. Thoughts?

[Kaminoan 0]

I just recently purchased a UM3. Initially bed leveling was working fairly well and holding day over day. My printer is about a month old now and bed leveling is consuming more and more of my time so I have been reading the forums and Googling. I have a few questions due to my lack of knowledge.

My UM3 seems to be square and all 4 feet rest on the counter top, so I believe it's frame is level. The print bed, should be it level with a bubble level or is the head out of level because of the two holding rods and the bed is level to the head? I assume it is the latter as when I level via manual leveling it is out of level via bubble level. I have ordered some metric feeler gauges, but is there some other critical tool for leveling? Also has anyone set up a video of UM3 leveling that is recommended? From the broad strokes to the fine tuning?

Thanks

Jared

[tenjin 0]

This is an old thread, but just wanted to post to say thanks to the original poster and the people who replied.

 

I had a similar problem to the OP, and when I tried the paper calibration method the bad area just moved to a different corner. 😀

 

I printed the grid from the OP post, and then (as one of the replies suggested) just manually adjusted the screws until I got a good print.

 

That worked a treat, and my UM2 is producing the most uniform prints I have ever seen.

 

Thanks

 

Darren.

 

[Smithy 1,146]

Thanks Darren for your feedback!