3D Printing Tutorial / Pointers / Guidelines needed

I’m going to try to make a brief tutorial or at least a set of guidelines for 3D Printing, but I’ll need some help from youze guyz. Including representative example Blender files with a part or parts doomed to failure or partial failure would be great too.

Starting with the Print3D add-on, what things can cause a model to fail? Non Manifold Edges surely. Bad Contiguous Edges? Intersect Faces? Zero Faces? Zero Edges? Non-Flat Faces? Thin Faces? Sharp Edges? Overhang faces?

I assume a simple closed manifold is sufficient to make a 3D solid. Will intersecting manifolds produce another closed manifold that is the sum of both?

1 Like

Make a manifold object, that has no self intersections, the rest is really up to your slicer and support building outside Blender.

Sometimes Cura has managed to slice non-manifold objects properly, but wouldn’t really want to rely on that.

1 Like

Thanks for that. I assume self-intersections falls under ‘Intersect Faces’, so ‘Intersect Faces: 0’. Check.

This is not a Blender thing but before printing it’s always good to check all the layers from the preview and scroll through them to check that everything is as it should be.

1 Like

I use STL export format from Blender, because it appears to work well between AutoDesk’s wonderful Fusion 360 mechanical CAD product and Blender. But other formats and their benefits are invited.

I assume you’re familiar with the Print3D add-on. Do all of the entries in the Print3D ‘Output:’ section need to be 0s?

I began part of my work with a rigged human model from TurboSquid I think. It’s got 24 non manifold edges, 23 intersect faces, 13 zero faces, 29 non-flat faces, and 386 overhang faces. I may just have to accept some partial failures from that and hope for the best. I assume if they’re small and the print holds together and aren’t unsightly or lead to mechanical failure of some type that it’s not really a problem. But always how the slicer actually parses up the model into slices. Check.

If you haven’t downloaded a separate slicer software and are using 3D printing services, you should probably download Cura, use the slicer and then use the preview to debug various issues. Nothing to do with Blender at all.

Step 1. Do a manifold, non-intersecting model in Blender
Step 2. Import the .STL to Cura, slice it and preview it. If there are errors, go to step 1.

1 Like

Thanks. Will do so immediately.

Installed. Thanks ambi. Best advice so far. See a typical platter full of parts from the comfort of home before sending off to the 3D printer blind, with evidently a number of different machines to choose from. Just ask your printer what machine they use. After studying it, I’ll include my impression of Cura.

Cura probably don’t have any of the professional printers big printing houses use (Stratasys etc). I don’t know what service you’re using, but Shapeways has a web interface to analyse the uploaded part and tell you if it’s actually printable. I’ve previously had no problem with it really. I just made the part with Blender and uploaded it to Shapeways, and they did the analysis for me if it can be printed.

1 Like

Yes, choosing a printer will be a challenge. I’m trying to print a miniature human figure for which I want smooth skin. I also have small gears which must mesh and roll against each other smoothly. So I need precision and high resolution. I’ll need to research and ask lots of questions. I plan to try ICO Mold, because my ultimate intention is to cast multiple copies of the parts myself, though they should be able to produce the raw parts instead, which might be wiser initially, since I still need to check that everything, mechanism, etc all work properly:

I’ll check out Stratasys and Shapeways too. My first experience with a printer was quite good. The cream colored parts were sharp, high res, smooth and just right with no warp that I could find. But that was years ago and I haven’t the slightest idea who that printer was:

My second experience with a discount printer (MAKEXYZ) was abysmal:

I’m going to ‘like’ your posts just as much as I can, because I value your advice so much.

I’ve set up supports in Blender for a resin printed model. I haven’t used the 3d printing add on yet. I have two FDM printers and two resin printers. First question is what type of printer are you using? Resin or FDM? Resin slicing will be more forgiving since it’s just outputting a cross section of the model. FDM models need to be put together better. i.e No intersection polygons, welded edges, watertight etc.
I’m using Autodesk mesh mixer and Zbrush for some clean up work on models and for FDM printing I’m using Simplify 3d for Slicing. For resin I’m using Chitu box but I’m looking for a better slicer.
Biggest problem I’ve run across is an area not being thick enough for the resolution. For example the ear lobes on a human head model. I always have to go in and inflate that area a bit to get it to print. The biggest failure point is supports, not having enough or in the wrong place. Also the first player not having enough contact to support the model. For example printing a human from the feet up. You need a base layer, or raft as it’s often called and supports going to the hands. Feet alone are not not enough to give the model stability while it’s printing.

1 Like

Thanks Mash3d. I don’t have a 3d printer myself. Wish I did. I have to find someone else to do the printing. I’ll keep your advice in mind. Can either of you current responders (or anyone else) tell me based on appearance alone what type of printing and/or material the cream colored parts pictured above were most likely made from? Those were of sufficient quality for my own needs.

The manufacturer of the miserable looking grey parts said they were made from a material called ABS with a process called FDM and claimed that for their FDM process, the expected tolerances are +/- .080” for the first inch and +/- .005” for each additional inch. But clearly, the result did not match their claims at all.

I can understand that a figure’s hand posed so as to hover over the ground has no support from below whatsoever, is the first encountered before the rest of the hand then arm then body by the slicer and must require a support. Is it also necessary to support, for example, arms raised above the head at an angle? In other words, is newly laid material insufficient to hold up extensions from it - i.e. does it remain soft and semi-molten for a while? I assume the raft is to keep the model from toppling over. Was wondering what that mote surrounding the figure were in Cura.

The cream colored piece looks to be from a resin printer.

The grey one is from an FDM or FFF printer (Fused Filament Fabrication). It looks like it’s poorly calibrated, and was probably printed too fast.

I’m curious how many pieces you plan on having printed? Shapeways and other 3D printing services are pretty expensive, and so their comes a point quite quickly where it would actually be cheaper to buy a 3D printer - which thanks to China - are now incredibly cheap at entry level.

Tolerance levels are relative to the printer and how you’ve set it up. A poorly calibrated printer (like the one that printed that grey piece) will have a much worse tolerance than one that’s been set up well. My printer for example is a Wanhao i3 Plus. I spent time calibrating it and I know roughly that I have a tolerance level of about 0.5mm. Which means, if I have an 8mm hole (for a rod say) I’ll make the hole 8.5mm in diameter.

A high end resin printer however (and even most of the cheap ones) will be much more precise, due to being able to print at much higher resolutions. There are two main types of resin printer (that I’m aware). One uses a laser to cure the resin, and another uses a high resolution screen to cure it.

I’ll also say that some of the stuff mentioned above about intersecting meshes being a no-no when 3D printing is not true…not anymore at-least. I’ve printed plenty of things from Blender with intersecting meshes that print fine. I’ve also printed files straight from sculpting, with intersecting geometry, and millions of verts, and they print fine also - just as an FYI

1 Like

Thanks Peter. Would you (or anyone else) have any best possible quality and resolution parts that you can photograph and link here (hopefully along with info about what printer you used and key parameters guiding its best performance)?

Decent FDM 3D printers cost a little over 200 bucks. Just the tooling for injection molding costs something like 5000 per part, which I assumed you’re planning to do. I’m little confused how you’re dividing your budget.

2 Likes

Ambl, looks like FDM probably isn’t going to suit me, as I have meshing 5mm gears with course teeth, smooth skin. High tolerance seems to be what I need. I’m a little confused too. I have what I hope is a good, innovative idea for a mechanical doll (about half the height of a Barbie), something like a music box ballerina. I’m reluctant to discuss it in detail yet, as I plan to seek patent protection for a few of its aspects. Ultimately I’d like to build and sell them myself. For that I would need steel or aluminum molds and would probably need to figure out how to injection mold at home and produce something akin to plastic scale models for hobbyists that I would assemble.

Complicating matters is that I first have to verify that the mechanicals will work as expected. So, I probably won’t head right for an expensive mold. Initially, I plan on having a precision set of parts made and I think I’ll need tolerances on the order of 0.05mm, so I’ll probably need the professional printers. If all goes well with that, I’ll probably try to make copies of the precious pieces with the blue stuff or green stuff - making cheap molds at home as described herein:

http://www.plasticmodelsworld.com/taxonomy/term/313

and see how that goes. If the mechanicals work and I still find her charming, I’ll probably try to see if there might be a market for her. If that goes well, then I’ll probably go for injection molding. Or I might try to see if a Walmart, a Target, or a leading toy store might want to license the potential patent. I have about $5000 right now, which I’ll be splitting between living and investing in that and a couple of other inventions I have in mind, before I have to go back to cab driving or looking for another job. So… we’ll see!

It’s looking like buying a printer might be a good idea, as Peter suggested. Or departing from Blender except for those pieces that have skin or fabric shells and making molds directly from actual gears, steel rods and brass in either silicon molds or materials like ‘blue stuff’, ‘green stuff’ or silicon caulk as recommended in the linked page (i.e. hacking my prototype that was cobbled together from those items). And sanding the 3d printed skin and molding the smoothed result. Gotta start somewhere.

If you’re making small parts that require high precision, SLA resin printers are probably what you want. The prints they make cost a lot more per weight than FDM and the max size is tiny, but if you’re doing small parts that require high precision, that’s what you need if you want to make things at home.

1 Like

Thanks for the input, folks. I encourage everyone to mention the printers they’re using and what settings they’ve had success with hopefully send a picture or two to show us.

The Cura slicer reported 2 days and about 20 hours (presumably to run on the printer - I chose Extra Fine - 0.06mm) with current settings and the Cura slicer preview is taking so long that I stopped it and will restart it before going to bed.

Would grouping the parts close together on the platter provide any cost reduction or otherwise take less time to print, being sure of course to leave space between them when viewed along the vertical direction if and when vertical support stilts prove necessary? Ambl, I may well send you a picture of the Cura pre slicer preview and ask if you see any potential problems and are willing to keep it confidential.

My parts aren’t particularly large, but a few of them are quite high poly. In fact, I turned up the subdivision surface modifier until skin appeared shiny and quite smooth, which was probably a mistake, and then Applied (probably another mistake rather than applying modifiers only during the STL file generation and leaving myself room to back out of the high poly to a lower poly). They’re essentially just the same parts as low poly, though a bit smoother. Do high poly parts cost more to make than low poly in addition to more time in Blender and the Cura software?

Would there be any advantage to connecting a multi-object design so as to make it essentially a single part much like plastic scale model kit parts are formed in the injection molding process? Thanks all

This time I started the Cura preview, thinking it were the same thing as the slicer. Evidently not. Yikes, a bit of red is showing up in the X Ray view. I assume red=bad. Better see what the cryptic Print3D add-on has to say.

CuraRed

The first tan color part was probably printed with a resin Straysys or Objet printer. Not sure what material.
The Second grey color part was printed with a potato. Here are some prints I did with a Crealty CR-10 and a Kudo 3d bean resin printer. https://imgur.com/user/mash3d/posts

1 Like

Thanks Mash, very helpful. In looking at the close-up of your standing figure closeup, any idea if the sort of roughness at 0.05mm is about what the pros can achieve? Not that you’re not a pro. If it is or their prices are too high, I’ll probably need to sand where appropriate and cross fingers where gears and fine clearances are involved. ICO Mold does offer an ‘insert molding’ process, where actual metal parts can be added inside the molds - probably expensive. I wonder how scale model kit makers achieve that excellent smoothness on their parts - probably also expensive. I’m always surprised when you folks take time from your busy day to help out a help sponge.

Ambl, I tried the Print3D and found lots of non-manifolds. I’m not sure those are a problem if they are tiny and the bulk of the braid mass is intact and it looks ok and doesn’t fall apart. Make Manifold decimated the braid and left a few persistent non-manifolds. The other deeper red region might be more of a problem though.