This topic contains 9 replies, has 6 voices, and was last updated by
Anonymous 2 months ago.
-
AuthorPosts
-
23/10/2014 at 7:10 am #8312
After a few dozen toys, gadgets and gizmos (including a two-piece 12″ ruler with a name on it for my lad to sell to his mates at school!), I have finally got round to try to print something for work, which is what I bought it for.
I have a mission to try to show all the companies I work with how a 3D printer might have some application for them. I have been sent the STL file for a laptop storage cabinet which one of my clients manufactures (for schools) in the Midlands. It’s a steel fabrication and stands about 1m high. I have sliced off the bottom and top of the model to print them separately as they are a bit fiddly (wheels, handles etc). What is left is the body of the steel cabinet.
It is made of bent sheet steel, about 2mm thick. The problem is, to get is small enough to print (and to be useful) I have resized it to 17% (1:6 scale). This means the walls are about 0.3mm - which I thought would print okay. However, it doesn’t. I’ve spent hours with meshmixer, netfabb, sketchup and 3D builder tying to find ways of solidifying the model more, shrink wrapping it, or thickening up the walls. All to no avail.
Attached is the photo of the final result (as well as a screen shot of the thing I am trying to print). As you can see, all the walls have turned in to whispy hairs of ABS. The only thing of substance to have printed is the support structure. This is ridiculously over the top as well - if a hinge sticks out slightly from the body, there is a support column all the way up the side which is hard to remove as it is too close to the actual model.
As a second best I’d be happy to print it as a solid cuboid, but haven’t found out how to convert the shell to a solid - meshmixer produces a weird blobby solid after a long time, and it doesn’t pick up the walls at all. It’s almost like I will have to get them to design me a version just for printing.
Has anyone else tried to print something like this?
Attachments:
You must be logged in to view attached files.Trying to spread the word about 3D printing to small and medium sized businesses in the West Midlands.23/10/2014 at 7:21 am #8315I think you should really adopt ‘prepare to print, slice, post-process, send to printer, *PAUSE*, inspect resulted gcode with gcode.ws, unpause or abort’ process. It saves filament and time with bad slicing. I am sure you would see something being wrong going through gcode viewer - it this case missing sides of your cabinet. Another thing I did was to load same STL to Slic3r directly (you can download the latest version with gcode preview) and see if there’s any difference in that.
I have been through similar troubles myself - going back to my model and using Solid Inspector with Sketchup to ensure object is watertight and then let STL run through netfabb too. And generally it made no difference to finally sliced gcode. Also, with skids, made my own support and/or made all edges that went at 90º sloping more gently. You can try that as well. Support generated by Slic3r got stuck to my model and I couldn’t remove it properly (same as you).
BTW, I am a bit puzzled why your prints almost always have filament strands sticking out of prints. I didn’t really notice them when I was printing anything. Is it possible that something else is missing not working well?
23/10/2014 at 8:21 am #8317
AnonymousI’d be surprised if you ever got to print walls that thin, that tall, for 0.3mm. Isn’t there a minimum wall thickness setting somewhere? (at work, so can’t check). I seem to remember someone saying that walls had to be at least 1mm to print successfully.
23/10/2014 at 8:47 am #83190.3 mm should be single pass wall - how well it will hold after certain height I don’t know… But when I tried 1mm walls Slic3r did it in two passes and left gap in between - practically making two walls of 0.3mm each with 0.4mm gap between them.
23/10/2014 at 9:28 am #83210.3 is pretty small, but what might work is if you create a custom profile, 2 perimeters, 0.3 nozzle for the perimeters, set the width to 0.1mm for the perimeter and try again (check out the gcode first!). I had good results printing 0.6 mm walls with 2x a 0.25mm perimeter for cookie cutters. If this doesn’t work you might need to get some 0.6 mm walls at least. Professional printing services like shapeways demand a minimum wall thickness of 0.7mm for some materials but 1.0mm for most, they probably have good reasons for that
23/10/2014 at 4:35 pm #8334Yeah, you definitely need to increase the thickness of the walls, also there may be issues with the mesh. NURBS shapes don’t always translate well to polygons.
23/10/2014 at 8:13 pm #8345Thanks everyone for your responses. Just back on to the printer now to have another go.
@clicky - you’re right about the strands/blobs problem. I have tried to play with the needle ejection and open volume values and I can’t seem to fix it. What I have noticed is that at the start of a run, there is often a delay when nothing comes out. It’s okay on a long run (perimeter, infill etc), but lettering and small features are really messy. I also have a support ticket open (relating to head temperature control at start up) which Chris White thinks indicates a ground leakage problem in the head pcb so I may be getting a replacement head soon. Also how do you inspect the gcode? Where is it?
@biscuitlad, @dave and @zach - I’d agree that a 1mm wall would be more practical. The thermal contraction on the thin whisps as it cools is quite destructive. I have been trying to find some way of doing this as I don’t have the CAD model, and that’s my problem. I’ve tried using netfabb, meshmixer etc and can’t find a minimum wall thickness function.
In this instance I don’t mind printing it as a solid with infill, so I am almost thinking of redrawing it in sketchup, but on an ongoing basis this will not be justifiable. I want to be able to take customers’ CAD files and print a model for them to introduce them to 3D printing and challenge them on how they can use such models in their sales process. If I have to do a lot of manual work on the file first, it won’t be scalable.
Trying to spread the word about 3D printing to small and medium sized businesses in the West Midlands.23/10/2014 at 8:21 pm #8348for the gcode, go to the cel folder that was set up when you installed am : Documents\CEL Robox\PrintJobs
sort the folders by date and you will see the latest one created for the last print job and in there are 2 gcode files, the one from slicer and the am generated one . you can drag them in to gcode viewer http://gcode.ws/ and see what is happening but chris white has said that itw won’t show up correctly here due to what am does for its own needs but you can see what it is doing at least
23/10/2014 at 10:30 pm #8355Interesting problem but I don’t think it’s solvable in a decent way without getting some manual work done. A cad file instead of a STL will help a lot on making it easier but you always have the problem that the scale isn’t correct any more so you will either take away some space from the inside of the locker or the outside making the margin between locker and handle a bit smaller. One thing you can try is loading the STL in a 3D modelling tool, selecting the faces of the side you want to move and move them until the walls get thicker. I’ve done this for much simpler models successfully (adjusting height for parts I need bigger when I only have the STL file) using both blender and meshmixer, don’t know if it will work for this though.
24/10/2014 at 12:38 pm #8364
Anonymous@wellmeadow - you might find it useful to have a look through the following suggestions on how to thicken walls, but I don’t how far you can go with just an STL file.
-
AuthorPosts
You must be logged in to reply to this topic.