0.1mm play in the extruder head linear bearings...

[Daid 306]

I'm having all sorts of problems with the amount of detail I can get from my Ultimaker. Infill not connecting to the perimeter at some times, while it worked at other times. Inaccuracies in the perimeter.

And I found out why, there is about 0.1mm play in my linear bearings of the extruder head. Causing sort of a backlash problem. To illustrate the problem, it causes this:

http://daid.eu/~daid/IMG_20111213_101806.jpeg

See the eyes and the nose. You can see it 'stripes' to the right after those, so the lines are more in front compared to the rest of the face. I've checked everything from belt tension so the generated GCode. And the only thing I could find is the 0.1mm play in the linear bearings.

If I take the extruder head in my hand, I can "wobble" it around about 0.1 mm. The axes don't move, and the bearings to. So there is play between the bearings and the axles.

So I'm wondering:

 

• [*:3s4jgmjy]Is this normal? I know atleast 1 other person has this. But I don't think you could print the ultra quality yoda like this.
  [*:3s4jgmjy]What could I do to fix it?

 

Note: My extruder came pre-assembled, so I didn't mess up the assembly.

[ddurant 0]

Not exactly sure what I'm seeing.. Is it the same type of problem seen here?

making waves... by ddurant123, on Flickr

If so, that's usually caused by running at too high of a feed rate. I think it's more "resonance" than "backlash" but could well be wrong..

[Daid 306]

That's fucky, looks like waves. My lines are actually pretty straight. My print was only at 60mm/sec. But it could be the same problem.

I think the problem is also more noticeable at the X axes then on the Y axes, because the bowden tube puts keeps the head pressed to 1 side.

Some custom GCode that prints a 1 perimeter square going clockwise and counterclockwise on each layer should show it I guess...

[Daid 306]

I did some more organic prints, and my conclusion is that this problem doesn't show up on those types of prints. Giving a bit of a false sense of quality.

http://daid.eu/~daid/IMG_20111213_223459.jpeg

(same settings)

http://daid.eu/~daid/IMG_20111213_232929.jpeg

(half layer thickness 0.1mm)

[ddurant 0]

I did some more organic prints, and my conclusion is that this problem doesn't show up on those types of prints. Giving a bit of a false sense of quality.

I think it could still be that waves problem.. On that printer (my old MakerBot Cupcake) you could actually see the x/y belts vibrate like guitar strings when it changed directions. On print areas like a straight line in x with a small y detail in the middle, it'd cause that ripple effect like my picture.

I don't know how to solve the problem but to see if it's the same issue, you could just sketchup something like a 30x30x5mm box with a bump down one side and try printing it at different feed rates. The issue won't be anywhere near as pronounced on an Ultimaker as it was on a Cupcake but you may be able to see a bit of a wave effect.

Or it could be a slicing artifact.. Have you looked closely at what it's doing when it prints that area? Maybe there's a Z change or a change from inner-infill to outer-infill right there?

[Daid 306]

If it is a vibration problem then it might be solved by rotating the model 45deg. Because then the belts are always moving.

But I don't think it's a vibration problem, because I'm not seeing waves.

Maybe this picture helps in showing the effect:

http://daid.eu/~daid/IMG_20111214_164139.jpeg

It's the same red bunny, shot from the top (I wish I had a better camera). You can see the perimeter around the infill. It's printed from bottom to top with a small nudge in the middle. After the nudge it's 0.1mm to the right and no longer connects to the infill.

[mikeq 0]

Daid,

Do you still have excess play in your linear bearings? I've had the same issue with my linear bearings for quite a while now(over a year). I have new bearings and was going to replace them. I was wondering if you replaced yours and still have issues with play.

Also, I have printing issues with non organic objects like you had mentioned in a post. I tried tightening my belts but still have these problems with fill not touching the walls in places. Did you ever resolve these problems? Any help would be appreciated. I don't want to go through the trouble of replacing the bearings if I will end up having play again.

Thanks,

Mike

 

[Daid 306]

I resolved most issues in my machine, except for these bearings. I noticed that printing at 50mm/s makes the bearing issue go away (I think it's the bowden tube keeps some pressure on the head removing the effect)

I've also checked a lot of other machines for this defect. I couldn't find any others anymore.

I do have a new set of bearings, but I haven't installed them yet. But my infill is touching the walls just fine now, so you'll most likely have a different issue. Maybe an extrusion issue. Posting photos of your print will help, as we can see a lot from those with our experienced eyes.

 

[mikeq 0]

How did you resolve your most of your issues? What were the things you found? I will take some photos and post.

Mike

 

[Daid 306]

Most of it was short-belt slack. I was also missing a screw in the slider blocks, causing some play on the X/Y rods. And I added some tensioners on the long belts, as well as securing my hotend better. Simply put, everything was a bit lose :-)

 

[mikeq 0]

I went through and checked everything also. Motor screws, pulley set screws, installed belt tensioners, made sure head screws were tight, short belt tension. Calibrated my steps for e. Installed the extruder tension bearing mod.

The only thing left is the linear bearing play.

Here are some photos.

 

The last two are to illustrate that some prints with the same slicing settings come out perfect.

I seem to have trouble with print that have circles and curves. The curves do not have fill to the edges.

I have to see. I have other prints I will have to dig out that may help.

Mike

 

forgot to mention in the one photo you can see that I'm able to bend back the outer wall. It did not stick to the inner wall. This only happened on certain out side perimeters. Not all the way around the piece.

 

[mikeq 0]

I installed the new linear bearings. There was a bit of disassembly involved. Which is why I hadn't replaced them for two years. There is no play at all! I installed them as suggested by someone on another post here. I put them in with the bearing lines in a cross configuration. 12 and 6 as he said. Interestingly, the bearing that had the worst play was installed in an X configuration. I will have to see if any play develops over time.

I have printed a couple of test objects and they look great. I have to print something I think will really show an issue and see if my problems are gone.

Mike

 

[snowygrouch 1]

I think it might also be well worthwhile measuring the YX 6mm shafts.

For linear bearing performance - the tolerance of the shafts is very important for smooth running and accurate position.

http://www.skf.com/binary/12-55843/SKF_6402_EN_11_10_19_GB_high.pdf

pgs 39..42

The recommended shaft tolerance is Ø6 h6

which means it has to be (for best performance)

Max Ø6.000mm

Min Ø5.991mm

If you have a micrometer (a vernier is totally insufficient for this), and measure your standard shafts

I think you will find they are maybe undersize. Probably about Ø5,970 > Ø5,980 range ( I cannot remember

what mine were when I measured them, but it was in this area).

This will cause the bearings to run very badly, as the little balls will clatter around instead of rolling

up and down their slide-ways.

I would recommend replacing the shafts with properly toleranced prescision ground shafts like these:

http://www.produktinfo.conrad.com/datenblaetter/225000-249999/237086-da-01-de-Silberstahlwelle_D_6.pdf

Of course this is not a magic cure, but will make sure the linear bearings perform to their specification.

C.