Top Layers not touching - UM2

[illuminarti 18]

Won't it fit with the spring body on top of the belt, and the arms coming off the top of the spring and just the ends hooking under the belt ??

[illuminarti 18]

As an unrelated thought, have you tried lowering acceleration to see if that improves backlash, and hence surface quality?

[Nicolinux 288]

That's the first picture then. Or I could bend the pin the other way and then see if it fits the way you describe it.

And regarding acceleration - I didn't try it. But I did something else.

So to answer my former question - no, it is not safe to downgrade the firmware without resetting to factory settings

I started a print after the firmware downgrade and the heated bed crashed into the head. Nothing broke thankfully. I'll let it cool down, re-level the bed and try again. And thanks for the tip with the acceleration settings, I will try it too.

By the way, we should really start a thread with the UM2 motion settings. So far I reverted everything to the defaults a long time ago. I have only adopted your retraction settings (retraction 5.5 and speed 35).

 

[illuminarti 18]

No, the first pic has the body on top, but the arms coming off the underside. Rotate it 180° about an axis parallel to the belt, so the arms leave the spring coil at the top. Or maybe I'm imagining things...

[illuminarti 18]

Like the second picture, but with the belt going over-under-over, rather than under-over-under, as looked at from left to right.

[Nicolinux 288]

Ok, understood.

But I don't like this solution anyway. When the head moves into position to print, the clothespin crashes into the pulley at the other end because it is too wide. Wouldn't work anyway. So I'd need to bend a metal piece in such a way that it bends the belt without taking too much space.

But since I seem to be the only UM2 owner with this problem, I hope to find another solution "soft" solution.

Oh and by the way, you changed your avatar. Nice, you look more seasoned and authoritative

 

[SandervG 1,521]

If the clothespin is not wider then the sliderblock, and I think it isn't it shouldn't hit into the pulleys.

If you install it exactly in the middle when the sliderblock is in the middle as well you should be safe.

In regard of the actual problem (which was not fitting on a belt tensioner).

Doesn't this remind anyone of an inaccuracy we had with Cura a while back in the UMO days?

The way it calculated wall thickness in combination with nozzle size sometimes created similar gabs.

I believe Simon was also involved in the solution and a detailed write up.

 

[Nicolinux 288]

Ok, thanks.

As I said, I need to create a custom belt clip that's not too wide. Although I don't like the idea to put on a tensioner on the belt that hangs on the sliding block with a built-in tensioner..

Btw. are there any drawings of the slinding block's innards? I'd love to see how this belt tensioner works.

I can at least report that it is not related to the latest firmware. I reverted to 14.01.02 and the problem remains. Also tried another Cura version - still the same (Cura 14.01 and 14.03).

And I played with the x/y jerk settings. Doubled and halved the value - no difference.

This is weird anyway. If you look at this image and the cube in the middle:

 

You can see the layers beneath and the gap between those too. I noticed that the gaps get bigger with increasing z height. The first 2-3 layers are fine. Then for every pass, the strands are further and further apart. Maybe this issue is somehow connected to the bed leveling problems I am experiencing (http://umforum.ultimaker.com/index.php?/topic/5218-bed-leveling-trouble-um2/).

 

[illuminarti 18]

Does your test piece come out the right height? I don't think it's due to leveling per se (as in left-right, front-back), but it could happen if the z-movement is fractionally off, so that the bed moves slightly too far each time. That would cause the effective layer height to gradually increase. Although your top infill layers look pretty well squashed, so I don't think that it is that.

Page 12 of the assembly instructions shows the inside of the slider blocks - it's just a captive clothes-pin spring, basically:

https://github.com/Ultimaker/Ultimaker2/blob/master/um2%20assembly%20manual%20V1.1%20_english.pdf

I seem to remember that we saw before that too-much tension in the x-y system can also cause backlash; perhaps that's the cause for you: that would explain why trying to tighten things up made it worse. Have you tried moving everything around by hand, adding lubricant, and making sure that the cross rods are properly seated in the sliding blocks, and not catching on the walls of the printer (left side panel, especially). You might even try loosening the the 4 long screws in the head a bit, to see if removing tension on the two bearings helps at all.

 

[cor3ys 28]

Stefan

Just stumbled across this belt tensioner for a UM2 not sure if it will fit above the blocks, looks like there using them on the belts at the back of the machine

http://www.thingiverse.com/make:74071

 

[Nicolinux 288]

Hey thanks Corey,

I tried it but as soon as I put it on one belt, the other became extremely loose. I didn't put it on the other one because I have only one slack belt.

 

[Nicolinux 288]

Does your test piece come out the right height? I don't think it's due to leveling per se (as in left-right, front-back), but it could happen if the z-movement is fractionally off, so that the bed moves slightly too far each time. That would cause the effective layer height to gradually increase. Although your top infill layers look pretty well squashed, so I don't think that it is that.

 

Yes the 20mm test cube comes out right. It should have a height of 10mm and I measured something between 9.93 and 10.08.

 

Page 12 of the assembly instructions shows the inside of the slider blocks - it's just a captive clothes-pin spring, basically:

https://github.com/Ultimaker/Ultimaker2/blob/master/um2%20assembly%20manual%20V1.1%20_english.pdf

 

Thanks, I'll take a look.

 

I seem to remember that we saw before that too-much tension in the x-y system can also cause backlash; perhaps that's the cause for you: that would explain why trying to tighten things up made it worse. Have you tried moving everything around by hand, adding lubricant, and making sure that the cross rods are properly seated in the sliding blocks, and not catching on the walls of the printer (left side panel, especially). You might even try loosening the the 4 long screws in the head a bit, to see if removing tension on the two bearings helps at all.

 

I don't have sewingmachine oil right now, but moving the head by hand is ok...ish. Not the smoothest movement ever, but not problematic either. I can move the head without moving the entire printer around (something that I wasn't able to do back when I had the UM1).

The cross rods are fine, they are seated correctly and dont scrape the inner walls (whould also make a hell of a noise if they did and leave marks).

I was just about to loosen the short belts a bit when I noticed that they were already loose again. They are equally slack now but I wonder - did I not tighten the x/y motor screws enough? I don't want to overtighten stuff (because I usually tend to...).

And just as I wanted to loosen the 4 long screws on the head, I noticed that I can slightly move them on the x/y plane. Even if they are properly tightened they do move around sideways. I think this is due to their length and the new added black spacers on the new units. I did loosen them a bit, but it did not make a difference regarding the layer tightness :/

 

[Nicolinux 288]

Oh wow, do they plan to sell UM2 kits? The assembly instructions are pure gold - now I can finally see behind the z-stage cover without taking it off

 

[illuminarti 18]

You can tighten the x/y motors pretty hard. They shouldn't come loose.

 

[illuminarti 18]

And no, I don't believe UM2 kits are in their future - I think it's just that when you need to get a bunch of people up to speed on assembling printers pretty quick, it helps to have good instructions to train from.

 

[Nicolinux 288]

Ok, I tightened the x/y down, the belts have more or less the same tension. But still no difference in print quality :(

It this wern't belt related, do you have other ideas what could be wrong?

 

[Dim3nsioneer 557]

It might be something you already thought of days ago, but is it actually the same with any filament?

 

[Nicolinux 288]

Tried two different filament colors so far (UM orange and UM metallic grey) and it is the same.

 

[Dim3nsioneer 557]

In this case I guess it doesn't get better if you increase the temperature for some degree, right?

[Nicolinux 288]

Well I am already printing at 230° which is pretty hot. But I can try though.

 

[illuminarti 18]

Try printing slower, to see if it's some sort of speed-related under-extrusion. Also, have you marked the sleeve and shaft on the extruder motor, to check it's not slipping?

Finally, double check the diameter on the filament.

 

[mnis 11]

Hm, one of the toothed belt of my UM2 is also a little bit looser than the rest, but I think it does not cause significant problems.

Top-Layers were often or perhaps even generally not closed properly, which I have now so well recognized. At first I thought and think it a bit still, it only affects objects with layer thicknesses >0.1mm.

Actually, I have the thing only really paid attention, after Nico opened this thread.

I had undergone a change from Ultimaker material to Innofill material.

I noticed the following afterwards:

- Filament strands were generally not connected correctly.

- A test cube 20x20mm then showed the problem clearly.

- The measured diameter of the material was usually a bit greater than 2.85mm at various points.

- Older objects (Ultimaker filament) were with the material setting 2.85mm fine.

- A test with the material setting 2.75mm for the new material again showed clean test cubes.

Now it seems to be good so far.

Only thing missing is a bug-fixed version Cura.

Markus

[braddock 1]

I can see some evidence of this on my UM2 also, though it improves in additional layers.

3rd layer or so

and later on.

I always print 1.2mm top and bottom layers.

Does this look normal?

 

 

[illuminarti 18]

Braddock, yes, I think it looks fine.

 

[Nicolinux 288]

Ok, I applied sweing machine oil to the x/y axis and checked the short belts again. Made sure they were al dente with matching tension and tightened the motor screws down like there's no tomorrow.

And all I got was this lousy cube: