Practical Enclosure Design, Optimized For 3D Printing

[3D Hubs] have shared a handy guide on designing practical and 3D printing-friendly enclosures. The guide walks through the design of a two shell, two button remote control enclosure. It allows for a PCB mounted inside, exposes a USB port, and is optimized for 3D printing without painting itself into a corner in the process. [3D Hubs] uses Fusion 360 (free to hobbyists and startups) in their examples, but the design principles are easily implemented with any tool.

One of the tips is to design parts with wall thicknesses that are a multiple of the printer’s nozzle diameter. For example, a 2.4 mm wall thickness may sound a bit arbitrary at first, but it divides easily by the typical FDM nozzle diameter of 0.4 mm which makes slicing results more consistent and reliable. Most of us have at some point encountered a model where the slicer can’t quite decide how to handle a thin feature, delivering either a void between perimeters or an awkward attempt at infill, and this practice helps reduce that. Another tip is to minimize the number of sharp edges in the design, because rounded corners print more efficiently and with smoother motions from the print head.

The road to enclosures has many paths, including enclosures made from FR4 (aka PCB material) all the way down to scrap wood with toner transfer labeling, and certainly desktop 3D printing has been a boon to anyone who’s had to joylessly drill and saw away at a featureless plastic box.

36 thoughts on “Practical Enclosure Design, Optimized For 3D Printing

  1. Ok I don’t know 3D printing, so maybe that’s why, but this statement doesn’t make any sense to me:

    “2.4 mm wall thickness may sound a bit arbitrary at first, but it divides easily by the typical FDM nozzle diameter of 0.4 mm”

    2.4 doesn’t go into 0.4….

      1. You also have to make sure your X,Y,and Z steppers are also doing full steps that match. if they dont then you need to find your magic numbers for the Z axis as to what a single step REALLY is then calculate from there.

        I can take any 3d printer and after finding what the real Z step is make it print better than the owner has ever seen.

          1. You find the z stepping in your 3d printer configuration, steps per mm. For marlin and repetier firmwares, executing m503 will display the configuration. Find out what microstepping the driver is doing. You’ll need to know what stepper motor drivers your 3d printer uses, then look at the jumper settings on the board. If the driver is controlled by firmware you can find it in the configuration. Dividing the steps per mm by the microstepping gets you the number of full steps per mm. Inverse that using 1/x and you’ve got the distance per full step. This value can be used in the slicer layer height, or a multiple of this value.

  2. I feel like the idiot, I’ve often run into issues with prints doing wierd things with wall thickness and funny voids- I never stopped for a moment to consider the nozel diameter v wall thickness.

    I’ve got a few objects to redesign…

    1. You don’t need walls that are multiples of the nozel diameter. No idea what the article is talking about. The slicer will create an outline that matches your cad design, then do an infill. The key is just keeping walls at last thick enough to fit some infill.

      1. Perhaps not, But it is optimal… making thicker walls just so there is airspace in the wall makes a needlessly bigger case.

        Also 3p printers and slicers often suck at short print distances like would occur in the wall design you suggested. You printer would have to have well tuned retraction for it to work.

        1. I have better luck if the walls have some infill because the infill patterns typically rotate creating an effect similar to plywood. A smooth path with no infill will shrink along the path rather than uniformly. Maybe it varies by printer, but I find infill is usually a good thing.

        2. I am not really sure why they say this is optimal. Printers also apply a certain amount of overlap, so dividing your thickness exactly by your nozzle width might land you in trouble.

      2. Exactly! And not all nozzles are created equal, it might say .3mm nozzle, but it might be .28 or .32 or anywhere inbetween. Aside from that there are a lot of issues with this design. Turn the case over and the buttons fall out. They should be bigger on the inside of the case not just the outside and probably best to join them together at the bottom. The lip wont keep the case together, you’ll need a screw or something along these lines
        https://www.3dhubs.com/knowledge-base/how-design-snap-fit-joints-3d-printing

      3. I think they’re talking about only having perimeters to avoid infill. Less retractions, faster prints and nicer prints.

        However, if a printer has a 0.4mm nozzle and is printing 0.2mm thick layers, the round output will get compressed down to a 0.2mm rounded rectangle shape and I’m pretty sure the width won’t be 0.4mm anymore.

        As far as I know, the only way to make 0.4mm wide lines is to print 0.4mm layers and anyone with a bit of 3D printing experience knows it just won’t work.

        1. The printer can control the extrusion width by pushing out a different amount of plastic. Don’t believe me? Park the nozzle 0.4 mm above the bed and extrude a bunch of plastic. You’ll get a big wide blob of plastic. Clean your nozzle off and repeat with a lot less extrusion. You’ll get a less-wide blob of plastic.

          Also, you can totally print 0.4 mm layers with a 0.4 mm nozzle and get a damned strong print, you just need to make wide extrusions. 0.6 mm width is plenty, though wider still wouldn’t hurt. Forget all the nonsense people spread about “never print layers thicker than of your nozzle diameter”, it was never true for a second.

          1. “never print layers thicker than [your favorite percentage between 50% and 100%] of your nozzle diameter”. Silly comments thinking my angle brackets were supposed to be HTML.

          2. If you want to print thicker than your nozzle hole, the shape of the nozzle starts to have a big influence on the layer quality tho. Some have a rounded tip shape, some have a bit of a flat rim around the hole before they flange to the outer diameter (like most of the E3D or the MK8 or MK10 shapes)… As soon as you go close to or wider than that flat rim around the nozzle hole, it starts to look really messy.

      4. I assume it depends upon how the infill is being done and exactly how wide it is.
        I can see that for small wall thicknesses (up to 3x nozzle diameter) and solid walls, having integer multiples make it run ‘smoother’ because it can lay down single runs of extruded material.
        For hollow/sparse infills, I would have expected that as long as the total width is more than 3x the nozzle diameter, then it can zig-zag (or whatever the preference) to fill in what is present in the gap (anything less than that is effectively treated as solid).

      5. The point is, once you are above a certain thickness where infill kicks in, there is no point in restricting your design to multiples of extrusion width.

        Looking at the picture of the print, his slicer creates two smooth perimeter extrusions at the layer boundaries and then infills the rest. So assuming extrusion width of .4 mm (not nozzle width as article suggests) you can stop worrying about multiples extrusion width after you hit 1.6 mm wall thickness. So his wall of 2.4mm will work exactly the same if it is 2.5mm, 2.3mm, etc.

        1. Well it depends how many perimeters you use. Normally I design my cases with 1.6mm wall thickness because it is stable enough in most cases and gets print only with perimeters (which sometimes better than using infill pattern, because perimeters are more flexible).
          But of cause I would print a 2.4mm wall with only perimeters too, since it it faster because of constant movements.
          But above 2.4mm I would always use infill pattern, and over that the printer will oft cause also hit 2.5, 2.6, 3.5mm almost perfectly. The slicer always tries to hit the outline of a body as good as possible, no matter what the thickness is. The divisibility by 0.4mm respective 0.8mm is only because of better prints on thinner walls.

      6. WHen I first got my #d printer back in maybe 2012 slic3r left me quite often with two unconnected shells. I just started making thicker walls and letting the infill work, that and always setting to 100%. I had some success turning up the ABS temp to around 260, thinner layers, and going over 1 on the filament extrusion multiplier setting.

  3. No no no, not nozzle diameter, but extrusion width! You really shouldn’t be extruding pieces of plastic the same width as your nozzle because your prints won’t be as strong as if they were wider. Fortunately, most (all?) slicers know this and do the right thing. It’s typically adjustable in the slicer too.

    If you’re making parametric models, it’s nice to make a parameter for extrusion width and another for layer thickness, then base the sizes of thin features off of those parameters. Not nozzle diameter as the linked tutorial says, and *definitely* not anything relating to nozzle diameter for vertical thicknesses because that just doesn’t make sense.

    1. I have my .4 nozzle extruding at a .47 width so I design everything at multiples of .47; or .6 on my .5 nozzle. But most the time the slicer does whatever it wants to do.

  4. Before reading this I was assume that it’S a basic rule to design with 3d printer using what printer can do. I use inventor must of the time I design and I have a basic template ant majority of dimension are based on extruder witdh and layer height.

Leave a Reply to JJ DasherCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.