Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Print distortion, need some help please.
#1
I've been getting a distortion in some of my prints for the first 5mm of the print. It is fairly consistent, in that, for a particular object/print the distortion is always there regardless of print options, speed, layer height, infill, etc.

I'm using XYZware 1.1.30.16 with FW 1.1.J . I'm using XYZ filament in un-reset cartridges. I have firewalled the XYZware to stop updates. I was using XYZware 1.1.32.x but the problem existed then.

I have made the bed more stable by shimming/hot glue the lower right hand z bearing. Stabilised the glass on the bed so it won't move. Lubricated all the guide rods and the Z screw thread (dry PTFE lubricant). I use plenty of glue.

My testing print is a half-cube is 25 x 25 x 25mm with 5 mm thick bottom/sides. Here's some pictures of the problem.

The object was printed in this orientation. (yes the print lifted slightly on the right, need a bit more glue) There is a slight curve on the right vertical edge. This would be fine but the edges at the back are far worse.





And another print that distorted.



I thought it might be a slicer problem, so I tried to setup XYZOpenMod and Slic3r 117. But I have no experience with slic3r. The print last night ended up as a blob of plastic that only vaguely resembled what it was meant to be.

I've read through every related and unrelated threads, the only other conclusion is the bed/hotend temperatures are a bit low. But I'm not at a stage where I can start messing with them.

Any ideas to what might be causing this?[Image: IMAG0974.jpg][Image: IMAG0970.jpg][Image: IMAG0972.jpg][Image: IMAG0976.jpg]
Reply
#2
My suspicion is the you have a bent z rod or you have a backlash problem. I would check out the threads on the back lash first,specifically look for the threads regarding circles not being around. It seems like maybe the backlash is off on only one axis.

A z rod issue might present itself when moving the platform up and down. In some cases it might be bad enough to watch the platter actually move left / right during up and down travel.



Kieth
Reply
#3
Thanks for the reply Kieth. I was worried that it might of been some defect in the z axis, but the problem doesn't occur in all prints.

I was trying to get XYZOpenMod/slic3r working and started looking at the g-code produced by XYZware and slic3r using the calibration cube.. Looking at the path taken by the extruder on the GUI you can see what's going on. On XYZware there are gaps where the layer perimeters start . The gaps alternate mostly between 3 of the corners and only for the 5mm bottom of the cube. After that the gaps are on the inside corner of where the 2 sides meet and inside the object..



My guess is that these, gaps cause some sagging and warping of the edge and so you get a curved edge



This is the default setting for slic3r where the layer starts are Aligned. I'm guessing that this will produce some warping as well.
There are 2 other settings in slic3r for Seam position Nearest and Random.



This is random, hard to see on this screen shot but the gaps (start/end) are in random places in each layer. I will try printing this as soon as I get the correct slic3r configuration for my da Vinci. I think this will solve the problem.

I gotta admit that I am very much a newbie at slicing for 3D prints. Previously I've always used UP! printers with propriety software. The slicing didn't produce this sort of problem.[Image: XYZ_sliced.jpg][Image: slic3r_aligned.jpg][Image: slic3r_random.jpg]
Reply
#4
I've had this when printing small cubes, with the issue showing up always on the rear-facing corners. I think it's a temperature issue (too much heat). I haven't printed this kind of calibration objects for quite some time now but haven't seen this problem anymore, possible because I installed a extrusion cooling fan. But besides that you can try lowering the extruder temp and setting a slower speed for small perimeters.
Reply
#5
I've had a chance to test my theories and to make a long story short, this is what I've found.
The alignment of start/stop points on the print does contribute to the warping, but it is not completely to blame.
As Oscahie has found, it is also a heat problem. If the bed stays too hot then this sort of warping will happen, even if the start/stop points are not aligned. The warping is less but it still happens. (Oscahie, I also get more warping at the back of the prints.)

So to the solution.

Now I am currently using the XYZ software (1.1.31.7) to generate a g-code (.3w) file using export. I haven't got slic3r working properly, or as good as, yet. I also found that stl files that slic3r finds broken and won't slice, the XYZ software will slice with no fuss.

To stop the XYZ slicer from aligning the layer start/stops, I rotate the object some random amount about the z axis. (39[sup]o[/sup]seems to work) This causes the slicer to place each layer start/stop in different place with no or very few alignments.

Then I open up XYZOpenMod and import the .3w file. XYZOpenMod loads the .3w file as un-encoded g-code. I then edit the g-code and make the bed temperature drop to 85[sup]o[/sup] for the 2nd layer and then to 80[sup]o[/sup] for the 3rd and remaining layers.

Doing these 2 things have fixed the warping problem.[Image: Untitled-1.jpg][Image: Untitled-2.jpg]
Reply
#6
Nice work there Tony !
Some of us have been struggling with this for some time. It never occurred to me it could be a slicing issue.
I also found using a fan helps but didn't solve the problem completely. I'll give your method a go.
Could you upload a pic of your results please.
Davinci 1.0 with repetier 0.92 & E3D hotend
Slicer - Simplify3d 


Reply
#7
First image is the calibration cube rotated 39[sup]o[/sup] and sliced with XYZ software 1.1.31.7. The slicing is not the best in terms of layer start/stop distribution but is much better than not rotated. The slicing is for 0.1mm layer height. This print was done without modifying the g-code and the bed temperature was a constant 90[sup]o[/sup]. XYZ slicing always has the first layer at 0.35mm and then increments at the selected layer height (0.1mm in this case), hence the thick bottom layer and "skirt"/"flange" (what do you call it?)

You can see the warping is very bad.

Next image is the same g-code as before, but edited to make the bed temperature drop from 90[sup]o[/sup] -> 85[sup]o[/sup] -> 80[sup]o[/sup] on 1st, 2nd ,and 3rd layers respectively. The temperature settings don't tell the printer to wait so the temperature drops while the printer continues to print. The final 80[sup]o[/sup] may not be reached until the 4th, 5th, or later layer. I have never checked that aspect. Oh, the attached file - quick calibration.gcode is the g-code for this print. Lines 181 and 326 vary the bed temperature. (lines 180 and 325 aren't needed unless you want to vary the extruder temp.)

As you can see most of the warping has gone. Any warping/unevenness may be due to the slicing or z wobble.

Last image was a slice using XYZ software 1.1.30.16 . Again rotated 39[sup]o[/sup] and 0.1mm layers. There are very few start/stops that are aligned. The only problem that I found is that the extruder pushes out a lot less filament on the first layer, even though the layer is set to 0.35mm. This is a setting in the slicer used by XYZ. This gives the problem of the first layer of not sticking every time. I edited the first layer height to 0.30mm, which made the second layer 0.15mm and helped the adhesion to the bed. This print also has the temperature drops included. See the attached file quick "quick calibration_XYZ_modded.gcode" .

Again a much better print than the first print but I can't see too much improvement, if any, over the 2nd print.

I also checked at what temperature the glue on the bed starts to let go of the print (monitored temp while a print was cooling). The temperature is just below 50[sup]o[/sup]. So it is possible to drop the bed temperature down to 70[sup]o[/sup] or even 60[sup]0[/sup] (I wouldn't go below 60[sup]o[/sup] to be safe).

I think the problem is the filament is not cooling enough before the next layers and is still soft and collapses slightly under the weight of the layers above. Lowering the bed temperature after the first layer allows the plastic cool faster and solidify enough to take the weight.[Image: IMAG1005.jpg][Image: IMAG1006.jpg][Image: IMAG1007.jpg] https://forum.voltivo.com/images/gcodes.zip
Reply
#8
There are many strategies that you could use to give the extruded plastic more time to cool down, like lower the extruder temp, lower the printing speed, put more infill, or simply put more objects in the build plate or even a several layers tall skirt. Ah, also increasing the layer height to 0.2 mm (which I think in terms of visual quality is very similar to 0.1) should help.
Reply
#9
Yes, any of those strategies could/would work. A few of those will slow down printing speed and as we all know the da Vinci isn't that fast to begin with. I was already printing the cubes at slow with 30% infill, and that would take 1:26 hr to print the whole cube. That's why these last ones were stopped about half way, to save time and filament. I was just trying to solve the problem and came up with just one of many possible solutions. If I was using slic3r then there is an option to reduce bed temperature after the first layer. Also I got a clue when I Googled and found this Part Temperature.

It's all good, we are all adding to the knowledgebase Smile
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)