Mass Customization with 3D printed systems

Appearing at the 2024 Bay Area Maker Faire

The intro

cover photo

It may be tempting to 3D print the same thing again and again, but it is cheaper and faster to make the same thing using molding techniques. In order to really use the unique advantages of 3D printing, printed objects should be unique. They should be customized to an person or small group of people’s needs. At the same time, getting good at CAD isn’t easy.

However, there are physical systems, generated by code or parametric modeling, that work with 3D printing’s core advantages and are easier to use. I’d like to show off some of the community-generated systems for organization that take advantage of this mass customization such as the Honeycomb Storage Wall and the Gridfinity system, along with metal extrusions, as an example of how to take advantage of 3D printing’s unique advantages.

A textual version of what I was talking about at the booth

booth photo

I got my 3D printer, not to make something impressive or fancy or cool, but because I wanted to print organizers and buying things off of Amazon was just not doing it for me. As I started to look through Thingiverse (Printables wasn’t a thing yet) I noticed that people were not just making organizers, they were making systems of organization.

And 3D printers are really neat things. The are magical. Upload file, get plastic object. But they have their limits. You aren’t going to want to print 200 of the same thing. Thus, in order for all of us to serve the community of 3D printer owners well, we need to keep thinking of better ways for people to work towards the sweet spot of 3D printing - customized things and prototypes and one-offs - and do so in such a way that doesn’t require everybody to be a CAD expert.

The industry term is “Mass Customization”. Mass meaning it’s everywhere and customization meaning that it’s unique and different for everyone. How do we get there? Well, let’s talk about Gridfinity.

Gridfinity was introduced in a video by Zack Freedman a few years ago. Before Gridfinity, there were some very similar systems. I even printed a bunch of pieces of one of them. But, one thing that was happening is that a lot of people were trying to sell you a set of models that you could print, whereas Gridfinity was mostly a set of starter models and enough information for everybody else to themselves go wild.

Some of the things that people cooked up for Gridfinity were parametric models in OpenSCAD. One of the neat things about this is that, for the audience of people who don’t know CAD, they can print customized models using the parametric tooling instead of everybody printing the exact same models.

Most of the customized models I displayed use OpenSCAD libraries, which is even more powerful because now you can think about a library that can be very advanced and friendly instead of a mere parametric thing.

But, even there, if you for example have a crimper, you only need to think about a rough holder that holds your crimper, not the entire organizational system. There’s a clear power to offloading processing onto other people’s brains. And, even better, if you use a Gridfinity holder, if you break that crimper and get a different one, you are replacing that segment of Gridfinity, not a giant plate with everything on it. One subtle feature of Gridfinity that works really well is that the base unit size is close to the optimal size; I’d used a prior system where the base box was about twice as large and it was too hard to make everything fit well.

I mentioned that Gridfinity was created because Zack let people go wild with a base design, whereas some of the earlier versions were such that a person asked you to pay them for the full set but provided you with some starter bits. That’s what drove Gridfinity’s popularity over literally anything else. It wasn’t specific to any particular CAD platform so people could approach it with the tools they had in front of them and it wasn’t limited to the vendor’s worldview. I don’t really sew; if you asked me to make sewing organizers they’d suck… but someone else can make perfect bobbin holders and whatnot.

Likewise, a friend-of-a-friend asked me to do some stuff for her circus aparatus. She was able to use a system for robotics that this company had cooked up to, as a dancer who did not study engineering, get pretty darn far. Except that the company that cooked it up merged with another one and they decided they did not need two incompatible systems for robotics… so they discontinued the one she used which meant that I did some printing of replacement parts.

Unfortunately, this is kind of a bug in capitalism - the best systems for mass customization are the ones where nobody owns them so you end up needing to choose between making the world a better place and making money.

I also had on display the Honeycomb Storage Wall, which works mostly the same way, but is more for the walls, as well as 2020 extrusions and DIN rails, which you can see as the metal version thereof. I brought one of my workshop bits that I designed in FreeCAD out of aluminum extrusions and then ordered a list of parts from Misumi USA. They cut and drilled and tapped the extrusions and I was able to build it up like an Erector set.

I’m mostly talking about organizers so far and I totally think you should think about using Gridfinity and Honeycomb Storage Wall to make your life easier, but it’s important to note that none of this is confined to just organizers, the world is full of systems! My soldering light broke one day and I’d realized that I had an OpenSCAD library that did trusses, some Hirth joints, and a modular electronics board for driving high-powered LEDs and so I was able to make a pretty neat little soldering lamp out of the bits I had already there.

Furthermore, I don’t think we’ve spent enough time really thinking about this problem. A lot of the early 3D printer owners were very very comfortable with CAD but my software-engineering perspective is that neither professional engineering users nor recreational users are well served by the CAD/CAE/CAM software that exists. It’s not easy for users of CAD software to be able to apply a parametric snap-fit or be able to apply one of these systems as a library except if you are using OpenSCAD and I definately understand that OpenSCAD feels very very friendly to me but is almost certainly not for new users.

Some of the prints I was showing

Posted:

Updated: