Looking for some advice on a project. When we got the snapmaker, I asked my kids of they wanted to make anything, and my son’s response was that he wanted a sword. Specifically, the Fierce Deity Sword from The Legend of Zelda Majora’s Mask. Not the Master Sword or anything (there are lots of those), but the one with the weird curves. Of course, there was nothing on thingiverse that looked quite right, so we modeled our own, since my wife is really good with SolidWorks. Model looked great, but given that we want the final print to be around 48" long, it obviously has to go into pieces. Starting from the top down, I have the first section printed, but the second one was giving me grief. It was printing fine yesterday, but this morning I checked and saw this:
Looks like the piece broke off the raft, started shifting, and then got knocked over. Here’s what the curvature on the piece looks like from one angle in Luban:
Obviously, the center of gravity is a problem, which is why I had a 30mm raft on the whole thing, but that was obviously not enough to keep it steady. Supports are going to destroy the surface, so I’m not wild about doing those, but might be a necessary evil in this case. They’d also be required if I printed the bits lying down, and vertical seems to give me the best surface anyway.
At this point I’m thinking of splitting these sections in half and just adding more joints, since that would minimize the overhangs. Given that this is going to be sanded/filed/painted for display anyway, it shouldn’t be as bad as the substantial supports that would be required otherwise. Curious whether there are other ideas for how to handle something like this (STL for the piece I’m trying is attached).
It’s a bit annoying since the sword curves in all possible directions.
For a straight sword, you could slice in half along the length and glue both halves together. Then you have large flat part.
I think you could probably still do it this way. Surface finish might be a bit less due to the print direction, but if you need to glue parts together some sanding (and painting) is usually required anyway.
Do note that a 3D printed sword isn’t necessarily very strong. (at least not the way my 6yo plays with it :p)
Do you have z-hop enabled? Bed adhesion, especially with a raft should be more than enough to compensate for the skewed C.o.G. Your part is likely colliding with the nozzle and as it gets taller the moment is magnified.
I would try again with z-hop enabled and your travel/print speed slowed and if it still gets knocked off then you should cut your parts shorter to compensate for the relatively small surface area contacting the bed.
Or maybe cut them at an angle so the center of gravity is, well, more in the center.
But depending on the print speed, you still do get a lot of jerky movements for a large thing high object. Reducing print speed for the higher parts might help too.
Is there a way to reduce print speed at particular layers, or is that just manually going to the controller and changing the work speed? I was thinking that speed might be a problem, which is why I rotated it 90 degrees so that the Y motion (which seemed to be the jerkiest) wouldn’t be shaking the thin dimension. Speed seemed to be fine for the first 70 or 80% (hadn’t failed yet), so I’d like to start faster and then slow down as the print got taller.
I did not have z-hop enabled, but I may try that next. I can see what you mean-bed movement tilts model enough to have the nozzle contact the surface, and crash! I have another piece printing now (much more stable, but 40 hrs), so I may try again with z-hop and a slower work speed.
As for strength, I’m well aware that this isn’t going to be a “fighting” sword. Plan is to make it a display piece for his room. Should look cool when it’s done, assuming I can get the whole thing printed. Plan is to sand/finish with a filling primer, and then metallic paint on top of that. Joining using 1/4" dowel pins, and a 3/4" dowel in the handle for a little extra strength.
That’s how I do it now Some slicers like Prusaslicer & Simplify3D allow you to slow down automatically when the print time per layer gets below a certain threshold. (useful in particular for very fine details). But that isn’t the issue here.
z-hop should help as well.
In that case I wouldn’t really worry about supports. You’ll be sanding anyway.
If still an option (because you haven’t printed too much parts yet) changing the angle of the cut would help as well to make it stand more straight.
You know…that just might work. I’ll have to go back to the model and check it out. I also tried generating a support spine up the front, and it looks like it might help as well (need to tell Luban to support 20% overhangs though). I think I’ll see how the next few parts print before tackling this one again.
Next test is going to be how well a print like this sands/fills.
Update-I tried printing this part again vertically, but this time with 1mm retraction, retract at layer change, 1mm z-hop, and supports at 20 degrees. I also let it run at 100% speed for the first half or so, and then reduced to 75% overnight. This morning I went to check it out and I saw this: https://www.youtube.com/watch?v=EgqdG8FO2z8
In the video, you can see the part move when the head travels over the model, at which point the z-hop seems to be ignored (although it does raise at other points in the print). This seems to have caused all the supports to break off, so they weren’t doing anything. About 20 minutes after I shot this, I noticed a pretty severe layer shift on the camera, so I went and paused the print to try to clean up and recover. No good (resuming seems to have broken the origin coordinates?).
After having some success sanding down another test piece though, I’m going to try printing this one on its side with supports and see how it cleans up. If I can get it smooth, then that might be the best bet for a shape like this one. I’m thinking that it might look better to minimize the number of joints and deal with more sanding instead of worrying about sanding support bits off.
A couple of other questions though (semi-related):
-Should the z-hop be raising the head when it goes over the infill? Seems like that might be a bug.
-The linear modules sounded kind of strange in that video running at 75%. Does that sound normal to anyone else? It seems OK at 100%.
-Does pause/resume via the touchscreen work? When I resumed, it seemed like the work origin was lost, and that the nozzle was extruding off the printed surface.
Running slower would lower the pitch, I think that’s what’s going on with the sound.
Printing it horizontally sounds like a good idea, sanding is much easier with the right tools. Also eventually the extra cost of retrying prints can add up to pay for said tools.
Not exactly sure and I think it depends on the slicer and settings you use. I faintly remember that it usually only does this over outer walls. But can’t check right now.
I think it’s “just” resonance causing the metal strip to vibrate at that particular speed and movement. I’ve had it at well that when increasing or decreasing the speed you just hit that bad spot where things start to resonate.
But just to be sure, because I can’t see it very well in your video. The y-axis modules do fit nicely in the cutout space in the bed? At tome point I had the impression they were a bit too far out o the back, but it’s just not possible to make out on just the video. (And it’s a rather common mistake people have been making)
For me it always has been working. Do at some point you hear a “bump” when the bed moves all the way forward or backward? (You shouldn’t) and it shouldn’t be a problem anymore with newer units, but some people have been affected by it in the past.
Yeah, I think Cura might give you more options here, but not sure. I’ll have to check it out at some point.
Sounds changing with lower speed makes sense, and so does vibration on the strips. Is there a recommendation to use something like lithium grease on the screws for the modules at any point?
Just curious-which tools are you thinking of for PLA sanding? Recommendations I’ve seen so far (and what I’ve tested) are just stepping up from 120 to 1000+ grit sandpaper, and not using anything high powered to avoid melting the plastic. Dremel is OK if you’re careful for taking off specific bits, and I’ve drilled out holes to make pegs etc fit better, but that’s about it so far.
I don’t use PLA very often, preferring PETG. However I have a belt sander like that in the garage with a variable speed knob. (Can’t find any brand information on the one in the garage though so just found a similar looking one on amazon…sorry if it sucks(Check for local pre-owned ones as well, got mine a thrift store for 70$))
I probably wouldn’t use it for fine sanding on the plastic, but removing the burs and getting it to “hand sanding” levels on something your size would take like…idk 5 seconds maybe 10 given the curve?
OK, I see what you’re thinking. Might work, but meltiness probably depends on speed. Going to try doing this by hand (broke some of the supports off already). See how it goes. I used one of the other failed bits to test sanding/painting, and they seemed to turn out OK.
What percent of infill are you using? Also, Cura is the basis for the Luban, I use Cura to slice my projects. Another thing to try is a light coat of glue (water soluble glue stick) to keep the raft down.
Also, when the first layer prints is it flat or rounded on the print surface?
I’m using the default of 15% infill. I haven’t tried the glue stick yet, since things have been sticking to the build sheet OK-the problem with this one was trying to print a tall, wobbly model. Doing the pieces on their side seems to help, even if the supports are a pain to remove and sand out. I was actually getting some deep scars in the surface from the supports, but wood filler seems to work really well so far, and pairing that with the filler/primer seems to be giving me a decent surface. Still lots of finishing work to do on this.
Some slicers like Prusaslicer & Simplify3D allow you to slow down automatically when the print time per layer gets below a certain threshold. (useful in particular for very fine details). But that isn’t the issue here.