1

Ok everyone. Keeping everything below for reference but here’s the breakdown.

Problem: On my Robo 3d R1 printer, my bed was so out of level that the autoleveling feature could not compensate. So in order for it to be touching the buildplate in one area it was smooshing nearly through the plate in another.

Resolution: Manually adjusting height of bed.

What you need: 2x pieces of standard printer paper. M4 washers from local hardware store (optional). scissors.

Step 1: adjust y axis height.

Process: Take standard piece of printer paper and tear 2" thick strip lengthwise. With machine on but with hotend and heated bed off home z axis through software to zero out z location output.

precheck: Robo3d beds sit on magnets protruding from the underside. These magnets sit in recessed pockets on the carriage with companion magnets. ****Make sure that the magnets are seated properly in their pockets***. The pockets are ~10mm ID and the magnets feel like they are 9.9999mm OD. If you aren’t careful placing the bed back down they can be askew and throw off your leveling. the wildly overpowered magnets will make you think its sitting in the pocket but it may not be. If you suddenly start seeing 1 corner raised way above the others, Check this.

(1) Raise z axis (arbitrarily chosen 10mm) Then move hotend to bottom right corner (+x,-y). place one end of the paper between hotend and bed holding the paper at the opposite end. Using software lower z axis until end of hotend barely scrapes paper (my personal rule of thumb is when the wiggling the paper is impeded but not prevented). Record z location output.

(2) Repeat process 1 in top right corner (+x,+y) and record z location output.

At this point one value will be greater than the other. In my case (+x,+y) was about .7 mm greater *yikes*.

(3) lift the low end and add shims. Use the M4 washers for course adjustment (mine were about 0.5mm thick) and with the remainder of your paper cut out circles the same diameter as the washers for fine adjustment (printer paper is about 0.1 mm thick).

(4) lower bed back down to carriage taking care that your shims stay between both magnets and that everything clicks in as noted in the precheck section.

recheck steps (1) and (2) and compare values. They should have gotten closer together. keep repeating until satisfactory leveling has been competed (dealers choice on where that is).

now apply steps (1) ~ (4) on the -y side of your bed and apply shims as needed.

(5) perform step (1) again at the middle left edge (-x, y=0) and record value. then perform (1) again at the middle right edge (+x, y=0) and record.

(6) to bring these values equal (x axis level) bring the hotend to the higher of these two values (the high side) and perform (1) again. With the nozzle at the paper scraping height, through the software move the hotend to the other edge (x axis movement only). then reach in to your printer and twist the threaded rod by hand on the side your hotend is on (the low side) until the hotend is lowered to paper scraping height again.

Now everything should be level. I recommend printing a large 1 layer high sheet to confirm. Thanks everyone in the community for helping me out with this. It should increase my print quality significantly.

Original Post:

Afternoon All,

I’ve spent a lot of time dialing in my z offset and when I get just under the sweet spot I start to see this. Bells went off though because its only in one corner of the part. I have a Robo3dR1 printer with auto bed calibration and I have tried it across several different beds/substrates so I dont think it is a buildplate flatness issue.

From the picture, to me, it looks like the extruder is getting too close to the bed in that area and pushing the plastic out and up as it prints. I should also clarify. The corner closest to the camera has the issue. The other 3 corners are perfectly smooth and ideal for bed adhesion.

My suspicion is:

it has something to do with my z axis. Maybe something to do with going from one side of the bed to the other causes it to be off?

or my auto calibration. Maybe its not calculating the compensation right and is not maintaining the offset properly.

Anyone seen this before and have any suggestions?

20160405_122831.jpeg
20160405_122831.jpeg1632×918 718 KB

2

Hello Hilttech The picture is not so clear to me. It looks if you’re filament isn’t extruded very well in this corner. Maybe the filament is bent to much in this corner so the friction is getting to high. Try to extrude by hand and look if you feel a difference. Bert

3

Even though the picture is a bit blurry, it kinda of looks like the nozzle is too close to the build plate OR your trying to extrude too much materials. Both these situations can cause the plastic to “squish” out around the nozzle instead of laying down a uniform bead

4

I think thats true. My confusion though is that it is only in the +x/-y corner, all other corners come out correctly. This is true across multiple build plates so it cant be any non-conformance in the bed. Which just leaves the z axis threaded rods machanical defect (either in setup or material), z axis control (from slicer or arduino), or the auto calibration as the culprit.

I’m hoping its the z axis setup. Do you (or anyone on this forum) know if a small relative height difference between the +x and -x z axis rods would make a difference after autocalibration. That is to say if I were to reach into my machine and manually turn one threaded rod a full rotation would the auto calibration correct for that by rotating it 1 turn in the other direction or just see it as an unlevel bed and adjust the z height as it traveled back and forth? (hope that question made sense)

5

Hi Hilttech,

It is definitely too close to the building plate in that corner but I can guaranty it is not an overflow of material.

Perform multiple bed calibration, note all the values and compare Them.

If your +x/+y corner average value is always the highest and the -x/-y, the lowest. Bed calibration might not be able to compensated for both.

What worked for me was to remove the bed and adjust the bed support. Solved my issue. Never came back.

Nick

6

Thanks Nick. Ill try adjusting tonight. Do you have a robo3d printer? And if so how do you manually ajdust the support?

7

I have this issue with my Robo3D, as well! In my case I troubleshot it to the plastic bearing support pillows that hold the print bed rods! They were wearing out and they were causing my print bed to move up and down if I pushed down or pulled up on the front side of it. This play in my bed was too great of an adjustment for auto calibration to compensate for! My solution was to do as Nick did and take note of all of the z-height measurements at the beginning of it’s auto calibration cycle and then add little round pieces of paper on top on the y-carriage print bed magnets in the lowest corners until the bed was as close to level as I could get it! I am thinking about removing the top half of the pillows and drilling two holes on either side of each of the bearings and zip tie them down! My prints basically would be great on one side and then thin out on the other and the lines would begin to separate, which would then cause lines not to adhere to each other. When I tried to overcompensate in the z-height the opposite would happen. I would get better results in the offending corner, but plastic buildup and then the extruder would drag and scrape plastic away and deposit it somewhere else, making an awful mess! This seems to me is what is happening in your case. You have a high corner, so you might have to bring the other corners up! Make sure your y-carriage rods are still tight too and if not give your bed a good push down to seat those bearings into their pillows and then take your z-height measurements! Hope this helps you!

8

Follow Up Time:

Heres where I am at:

z offset: 1.07 mm (I know its big but its what works)

first layer thickness: 0.3 mm

z probe autocalibration output

Bed x: 15.00 y: 20.00 z: -0.43

Bed x: 107.00 y: 20.00 z: 0.10

Bed x: 199.00 y: 20.00 z: -0.37

Bed x: 200.00 y: 130.00 z: -0.47

Bed x: 108.00 y: 130.00 z: 0.03

Bed x: 16.00 y: 130.00 z: -0.37

Bed x: 15.00 y: 240.00 z: 0.12

Bed x: 107.00 y: 240.00 z: 0.42

Bed x: 199.00 y: 240.00 z: 0.08

Eqn coefficients: a: -0.00 b: 0.00 d: -0.34

planeNormal x: 0.00 y: -0.00 z: 1.00

echo:endstops hit: Z:0.08

I printed a plate shown in the image below 5"x5"x0.3mm (1 layer) shown attached to this post. Clearly my y axis is off. I’ll try the paper shim technique listed in previous posts.

Also for reference:

Plate thickness

-y edge: 0.41 mm (0.11 mm oversized)

+x edge: 0.46 mm (0.16 mm oversized)

+y edge: 0.43 mm (0.13 mm oversized)

-x edge: 0.41 mm (0.11 mm oversized)

20160412_074305.jpeg
20160412_074305.jpeg904×906 412 KB

9

Yup! That’s exactly what mine was doing only towards the right front of the print bed! Your numbers look very odd, however. Seems like your print bed is warping upward in the middle, and dipping slightly lower on the left side compared to the right. Then somewhat leveling off on the back corners, while still remaining warped upward the the middle. Not sure what is going on here! I would start by adding paper shims to the front corners and see how the numbers change. Hopefully your bed isn’t warped too much that it is not correctable. The object is to bring all your numbers as close to zero as possible, and hope that the auto level does the rest. Did you do your first run fine tuning paper calibration cycle yet? If you have not done this it might help your auto-bed leveling to do this first! Are you using MatterControl?

10

Ok so I am using Repetier. Specifically slic3r in repetier.

I did the paper test to calibrate the x axis before the picture shown above. This is why its relatively symmetrical across the x.

I agree my numbers imply a convex bed plate but it says the same thing for the actual bed and the sitall bed I clip ontop of it. It seems odd that both would have the same warped geometry

Lastly I added paper shims to the front corners this morning and it seems like it is generally helping. Couple of issues though:
(1) First the amount of paper I have to use: I am at like 8 or 9 pieces between the magnets in some areas. so there might be a chance of instability if the thickness of the paper stack goes beyond the depth of the countersunk hole.

(2) The numbers seem to change: my suspicion is this is related to how the magnets seat themselves in their respective holes (aka crappily) but I am looking at ±0.1 at all corners after I lift and then re-seat the bed.

(3) Getting it to zero. Thats fine for any of the negative values. Just add shims. But for the positive values I cant subtract shims that arent there. Are these values all relative to some starting value or to some absolute machine value?

My goal right now is to just get the deviation as close to zero as possible, not the actual values. Anyway I ran another print this morning post shimming and it is much better. However with it more level now my z offset is incorrect. Rather than it being equally smooshed down in all areas like I want it is equally too far away from the bedplate in all areas. C’est la vie.

I’ll post an update tomorrow morning once I dial in the shimming more although any help (especially with the positive, high spots) is much appreciated.

Geoff

11

Your right the shim technique will not work on the positive values, but try running your fine tuning paper cycle again and possibly set your z-offset back to where it ought to be. See if this makes a difference. I think once you clip your sitall glass on top of your already warped print bed, it warps your sitall glass. All Glass is flexible after all, but only to a certain point before it snaps. Not sure how much trouble you want to go through to solve this, but you could try shimming your sitall glass to see if you could get it more level then your bed. You could just as easily buy another bed from Robo3D and hopefully you get a better one this time. Sorry you have to go through all this frustration, but we’ve all been there and it’s not easy. It’s the nature of the beast in this arena.

12

This is a very common problem. It isn’t due to warped glass, it relates to two things when your printer is probing the bed:

1) probing the centre of the bed causes the glass to bend giving a lower-than-it-should-be reading

2) when probing the left-hand side the x-carriage pivots on its rods only tripping the one microswitch also giving inaccurate readings

The upshot is to manually level the bed as people here have already discovered.

More info and some interesting solutions including IR probing here:

13

Thanks for the heads up. I’ll take a look at this asap. I’ve also scheduled a meeting with one of robo’s techs to get this sorted. I figured manual was the way to go when I kept getting varying results from the auto level.

14

Nice post you linked to! Not sure if this is what is going on in his case, as far as the warping issue though, because his glass is warped upward not downward. Unless there is a differential math thing I am overlooking. His numbers suggest his center point is much higher than the outer points not lower.