Big frisbee print shifting in Y and X halfway in print - firmware bug?

[gadgetfreak 7]

Hi,

I've been using my Ultimaker for almost two years.

I'm mostly using Cura (13.03 now) and I have upgrade to the v2 hotend and the v2 extruder and v3 bolt.

Most of my models print fine, some really great.

But this is the first time I print something really large in X and Y (the whole of the print bed).

I modeled a frisbee (a flying disc of the "disc golf" type) and I'm trying to print it full size (about 20cm in diameter).

I scaled it to the max with external support in Cura 13.03.

The firmware I have is from Cura 13.03 as well ("Ultimaker: 13.03").

Now, halfway into the print (about half the height, total height in about 1,5cm), both the actual model and the support structure has shifted 8.5mm, mostly in Y direction, a little bit in X direction.

 

In Cura, the model looks fine after slicing:

 

In RepetierHost the model looks fine after importing the same gcode I ran (used the Ulticontroller to start it) during the print:

Could this be a firmware bug?

Other models I have print fine (although not as large in X and Y as this one)...

Thanks for any troubleshooting help.

/Daniel

 

[illuminarti 18]

Most likely it's just skipped steps due to additional friction on one of the fast moves, causing the head to lose track of its position at some point.. All further parts of the print then happen slightly offset.

Might well be due to the large size of the print causing the head to stray into the corners that it doesn't often visit - there might be a more friction due to less lubrication or slightly misaligned axes in the corners.

You might try moving the head by hand, with the steppers powered down, and see if you notice anything different at the edges/corners. Be sure and lubricate the axes and cross rods - especially at the corners,. You might also check if there's any chance that the end switches triggered while printing as that might throw things off too, I think.

 

[gr5 2,210]

There's three possibilities I think. Either the steppers skipped as illuminarti suggests which would probably be fixed by tightening the set screws (all 6 for each axis). Or the plexiglass may have moved (I doubt it) or the 3rd possibility which I think happened here is that the frisbee came loose off the bed. You should be able to tell better than us if this 3rd possiblity happened. In your photo it looks like the frisbee is off the blue tape and outside of the normal build area. Also it's very rare for X *and* Y to slip at the same time (same layer).

Large parts have a tendency to warp and the edges lift slowly off the tape. This would be the worst when you are printing the underside of the frisbee (the top from this view). When it is printing those solid layers and they cool they shrink and pull very hard and warp the sides of the frisbee up off the tape.

You can fix this by putting down fresh blue tape and cleaning well with isopropyl alcohol. This makes the blue tape AMAZINGLY sticky.

Also it's helpful to set the skirt distance to zero and add 4 or more loops ("line count") of that. This makes it harder for air to get under the part and continue lifting it off the bed. I have no idea how skirt and support will interact and it might not work properly. So after slicing make sure the skirt is touching all around the frisbee and that it and the support structure don't interfere with each other.

 

[jeremy 0]

Is it possible you are hitting the end-stops?

 

[gadgetfreak 7]

Thanks for your replies guys.

I'll will re-check the set screws the next time before I start the print again (not until earliest Sunday now unfortunately :( )

The plexiglass cannot move in the Y direction, so that can't be it.

The frisbee (or the support part of it) has skewed a little bit off the blue tape, but the blut tape is also a couple of millimeters to short on that end.

I think the axes are well oiled, it was not too long ago I oiled them.

I'm starting to lean towards Illuminartis

"You might also check if there's any chance that the end switches triggered while printing as that might throw things off too, I think."

because I know they were really close to getting hit (the first run actually failed a bit funny because of that I think).

However, isn't the print supposed to stop altogether if any end stop is hit?

I can check this somehow from Cura by executing som M commands towards the printer, can't I? I remember there are some settings on that in the firmware.

And how would the things be "thrown off" so to say? I would actually consider that to be a firmware bug then ;-)

The print surface was pretty bad considering all support needed to be removed from the top of the surface (printed closest to the print bed) so I'm planning on designing some manual support rods and print it diagonally with slicer-generated support and let it be 20cm high instead. Do you think that would make a nicer print perhaps?

Cheers

Daniel

 

[illuminarti 18]

I vaguely remember some testing where software end stops would kick in and not allow the print head to move past certain maximum coordinates, as well.

So I think what happens is that if the head hits an end stop (and/or just tries to move past the maximum coordinate allowed) then the internal counter increments, but the head doesn't move. So then when you try to move back to the next part of the print, you overshoot because you are traveling as if from where you wanted to be, but actually starting nearer than that.

I'd need to look at the firmware code to double-check. I suspect there is room for improvement, yes.