Top Layers not touching - UM2

[visu-al 2]

NEW INTERESTING FINDING

i was just playing around i holded the filament and pushed wich gave a better but not good result then...

..during print unscrewed the long printhead screw at the front left just by 180° and it imiadetely printed the surface nice and filled !!

i know about that problem but i did unscrew and rescrew to test in several strenght but i did not test changing a single screw.. i will print on to see if it's not a coinsident

@Nicolinux and @mevander try to play during print just with that screw (the one that is closest to the nozzle)

 

[visu-al 2]

allready printed over 7 surfaces without gaps or problems!!! (print speed 35mm/s 0.1mm layerheight)

very nice

the only thing to mention is that now it sounds strange a little bit like a loose screw resonating..

 

[Nicolinux 288]

So? What is causing the problem then? Something that pushes on the bowden tube or teflon piece and increases the tension in a weird way?

 

[SandervG 1,521]

Are you sure it is not a screw belonging to a fan?

Can you tighten the rest a bit more, so relatively that one is still the loosest?

(but more tight so it doesn't resonate).

Do you have some pictures to illustrate the difference?

Very happy to hear we may are closer to a solution.

Sometimes it is not as difficult as we may think...

 

[DidierKlein 729]

I think he's talking about the four long screws that hold the head together.

If I remember well i read a while ago someone mentionning that tightning them too much could cause some underextrusion... i have to dig around to find that subject back

 

[visu-al 2]

i just had a little shock seeing massive gaps in the print from far but then i saw it was the first layer after infill HAHA shock away ^^

@SandervG the sound is gone allready again without any changes so i can't search for it any more

the screws are not loose it was more like i used a slight force to screw them in as i would think is good and then on half unscrew the "nozzle"-screw

i will take a video as soon as possible, but to do right now i would abort the print that doesn't have more big surfaces..

 

[DidierKlein 729]

Found it:

http://umforum.ultimaker.com/index.php?/topic/4586-can-your-um2-printer-achieve-10mm3s-test-it-here/?p=39708

 

[nallath 1,120]

I've asked our mechanical engineers; They believe it might be because of the bowden tube (tightening the screws creates a lot more tension between teflon part & tube). Could you also try to tighten the screws and ensure that the bowden tube doesn't have a lot of tension?

 

[Nicolinux 288]

How could one test that? You mean the downward pressure of the bowden tube on the teflon piece right? Or the bowden tube in general? If the four long screws are super tight, I can still remove the bowden clip, press down the tube coupling collet and slide the bowden tube out without any resistance.

 

[visu-al 2]

How do i check the tension of the bowden tube ?

Me too when the underextrusion happened i was able to remove the bowden tube like usual..

 

[opit78 0]

I've informed UM technical support about that issue long time ago but they never reply to me about that.

There is one possible error in the design of UM2.

Y axis's bearing is squeezing the bowden tube.

Please take a look at the picture below so you see what I'm talking about.

Same problem on left and right side.

 

[DidierKlein 729]

Some more information from illuminarti:

 

We saw that on our early production machines a long time ago. I think there can be a couple of things going on:

 

1) Since the bowden is clamped at the top of the head, if you tighten the thumbscrews after seating the tube, you end up compressing the end of the tube along its length. This could cause the the tube's internal diameter to close up a bit.

 

2) If the screws are too tight, the hole head assembly gets pinched. The metal plates are totally open on the front, so you can end up with the top and bottom plates not quite parallel, throwing things out of alignment, and for instance, adding a bit more friction where the Bowden meets the coupler, and the coupler meets the nozzle.

 

@opit78: Could it be the angle at which the picture was taken or do you have evidence of this?

 

[visu-al 2]

..but i remember re-inserting the bowden was not as easy as it is on my older UM2!

 

[opit78 0]

You can look at your own printer, why should I prove anything ?

Remove the bowden tube and just look inside.

Reinserting the bowden tube while some filaments is inside is very often not possible, you won't be able to "pass" this bearing.

 

[DidierKlein 729]

I'm not questionning what you say i will take a look tonight if i can spot something similar. I never have any problems putting the bowden back in place on my printer (with or without filament inside).

I'm guessing your bowden shows some kind of marks because of this?

 

[nallath 1,120]

Well, checking it should be a matter of feeling (by hand) if there is any tension.

Sending design flaws to support is probably not the best way to do it. Most R&D people visit the forum every so often, so this is generally the place to post those things. Even so, we do miss things.

 

[visu-al 2]

..putting the bowden and filament together in the printhead is very hard to do..

im gonna film the effect of unscrew those screws now but need to wait cause the first layer of my big surface doesn't tell the whole story^^

 

[mevander 0]

I tested this "trick" and while I see a temporary effect that improves the print, the threads go back to separating after a few layers.

 

[gr5 2,167]

@mevander how the f*** could you maximize your image??

 

lol! It's not that big of a deal because people can click on your images and see larger versions but to answer the question...

After uploading your print to the gallery, go to view the image and get the largest possible image and right click and do "open image" or something so you can get the complete URL to only the image. The URL *must* end with ".jpg" and not with ".html" or something else.

Then write your post and click on the little "tree" icon below the "smile" icon. Insert the URL there.

 

[opit78 0]

For us only solution was to increase temperature to 245*C (Ultimaker PLA) and flow rate to 105-110%.

Still, that's only temporary solution, high temp is no good for detailed prints.

 

[visu-al 2]

@gr5 thank you very much

@mevander i did too but did you also try seperate tensions on each screw?

Here is my video starting with underextrusion:

 

And here is a picture showing it as well:

 

[gr5 2,167]

I think you might also experiment with different filaments (some softer, some harder) and tension of screw inside extruder. These can also effect how well the extruder grips the filament.

I have to say my extruder "skips back" at the smallest provocation long before it would start to do as much underextrusion as you see without any skip back. So as long as your extruder is not skipping back I think changing your flow to 110% might be the best solution or part of the solution.

 

[IRobertI 520]

I think all it's doing is moving the print head down ever so slightly thus squishing the layers more. I'd expect it to go back to not being completely filled after a while just like mevander reported.

edit: This is regarding the loosening of the head screws.

 

[mevander 0]

@gr5 thank you very much

@mevander i did too but did you also try seperate tensions on each screw?

 

Yes, I did experiment with loosening all of them and trying to just loosen the one closest to the hotend. It still went back to separating lines after 2-3 layers.

 

[visu-al 2]

@robert

cannot be the only factor cause i was able to print a part 10 times faster then before without any critical underextrusion at all. i will do more tests tomorrow to determine some more exact speed limit relation