We live in revolutionary times that some call a “new industrial revolution” – driven by the mixing of computers with manufacturing technology. By far, the all-star of this movement is the home 3D printer – as typified by the RepRap and its hundreds of derivatives. There’s a lot of great energy in the field – which is awesome. However, there’s also a lot of fluff and hype floating around the movement that is distracting from a lot of the real work that needs to be done.
Digital Manufacturing is the Real Movement
3D printing is great, but (as I say below) it’s only a small part of the solution. The real thing is much larger than any one process. It’s the digitization of manufacturing in general – from CNC milling, to factory floor automation and all the way to system-wide integration beyond today’s primitive quoting sites. By applying the ideas of software to this field: automation, abstraction, standards, etc. – you give access to a far wider group of people and create many more possibilities for solving the world’s problems.
3D Printing is a Gateway Drug
Accessible 3D printing is being heralded as the being a new technology that’s going to quickly change manufacturing, and it has been doing great things for prototyping. However, there are a lot of engineering reasons why the technology will continue to not be used for most production products. Refer to “Nothing is ever Democratized” below for why.
3D printing is exciting to me, not because of its future for mass production (which is overhyped), but because it’s a gateway drug for the field of digital manufacturing in general. We are seeing more and more other processes going digital, and this wasn’t nearly as widespread before the explosion of 3D printers – along with their open source control hardware / software.
Distributed manufacturing doesn’t work
In the near term, the idea of distributing manufacturing doesn’t hold water. Not for the environment, price, or perhaps even time. The problem is that we haven’t brought down the minimum viable economies of scale – i.e. the smallest size that efficient production can exist at. Efficiency tends to reflect in all aspects: sustainability, price, etc. Look at the “nothing is ever democratized” section below for more on why.
Distribution of prototyping on the other hand is already happening and integral to the future. Look no further than hackerspaces, TechShop, etc. and you’ll see the means of prototyping to be getting pretty flat. The reason this works is that your main focus is time. The more product iterations you can get, the better your results will be – which requires on time efficiency, not low unit costs like production does.
Nothing is ever Democratized
Did inkjet printers democratize printing? Does Amazon have a bunch of HPs printing off their books? No way – they have very specialized machines doing this, and the same is true for everything in engineering. Remember that old saying: “Good, Fast, Cheap – pick two” – this applies to all engineering problems. You optimize for “good” and “fast”, and this comes at the expense of “cheap” – which means it’s not democratized (few people can own it).
There is no end-state for this either, tech marches on, specialization and all. Specialized techniques and processes are invented to solve specific engineering problems, and this makes them sticky and other processes at a disadvantage. So yes, you may democratize prototype-grade 3D printers, but then others will be make huge, fast printers that are able to beat your per-unit cost by an order of magnitude – but at high capital cost.
This also means that non-3D printing processes will continue to improve too. The trend today is for the ever higher specialization of processes. Design today, for better or worse, is trying to out-do each other and looking at Apple’s crazy processes as an ideal, or in other ways looking at non-automated, handmade aesthetics. These both mean that we’re not seeing designs getting simpler, but more specialized – meaning that you need more machines and specialized craftsmen – leading to less democratization.
Machine Cost Doesn’t Matter
People would ask us at CloudFab why we didn’t have MakerBots on the system, and when I told them it was too expensive they were shocked. The most expensive part of a MakerBot is that it requires a lot of babysitting and produces one part at a time (with a pretty high failure rate), where an industrial SLS machine can produce 1,000 parts over the course of a build and takes a fraction of a labor time for setup / unloading because it’s spread it over many parts.
For prototyping, cost matters a ton so that everyone can have easy access to it, which is why MakerBot makes complete sense there, but not for production tools where it’s more about the machine’s cost efficiency (units / $) than its capital cost.
Humans are the expensive part of production. It’s why we switched to China. If you want to make a dent in this movement, you need to focus on systemic/process stuff: machine utilization, reducing human labor, material costs, etc. It’s why Shapeways bought machines, and they weren’t RepRaps. There is a lot of room for production improvements, and few of them require new machines.
CAD isn’t the Problem
Access to CAD tools isn’t the problem for making hardware. You can pirate SolidWorks on The Pirate Bay with a modicum of motivation, and the software is actually really nice. Open-source CAD hasn’t really moved the needle; it’s a nice-to-have without being able to out-do SolidWorks and similar CAD. Training also really isn’t the issue. There are already tons of educational videos, etc. – I taught myself in a few days and in no time I was able to create parts and assemblies. I’m convinced that most people are actually just being lazy – or are hobbyists who fall into a different market than people who want to make serious projects.
The hobbyist tools (like 123D) seem more show than go at this point – more like scripts in Rhino than tools you can use everyday. Stuff like TinkerCAD is actually useful for applications like quickly getting to a 3D printable model, but the closer you get to production, the more you need the traditional solid CAD packages.
Tons of problems lie around CAD tools though: collaboration, design rule checks, costing, simulation, etc. These are hard problems, but they’re definitely pain-in-the-ass issues that are slowing development. Companies like Upverter are working to integrate everything from CAD tools, simulation, perhaps even to the manufacturing – which is the direction I see being the most interesting. Mechanical CAD is harder to do in the browser, but I see a similar future of very vertical companies with design to production under one roof.
Vertical is Back
The business school ideas of the 80’s and 90’s were all about outsourcing anything that wasn’t a core competency. This makes sense for things like janitorial work, compliance, etc., but outsourcing manufacturing is a trickier business. Yes, building your own factories is often not possible, but often it’s the only way to acheive results. One of the big reasons for this is the iterative design loop that builds organizational knowledge – this is lost in nearly all outsourced projects.
Some of the most interesting companies today are either very vertical or essentially vertical. A good example of the former is Hyundai who built their own steel mill in order to make specialty steels just for their cars. Of course both Tesla Motors and SpaceX similarly make over 90% of their own parts in-house.
Even though everyone makes a big deal of Apple’s outsourcing to Foxconn, etc. – they’re actually very integrated and on the ground there. There’s no knowledge being lost on them and captured by Foxconn. Apple knows that manufacturing is key to their future design, but Foxconn knows how hard it would be to get even a company like Samsung to the level of Apple using only Foxconn’s internal knowledge. This makes Apple essentially vertical without the need to actually own factories – which puts the onus of labor, gov’t relations, etc. squarely on a Foxconn who know the local knowledge more than Apple.
Stop Making Non-Interesting 3D Printers!
Whenever I see someone make a “new” 3D printer that’s just a derivative of the RepRap or MakerBot – I could care less. Only new processes, great interfaces or super-low price points get my attention anymore. FormLabs being a great example of all three – which is why they were a massive hit. If you’re looking for problems: make a cheap laser cutter, CNC mill, or pick-n-place machine – not something that everyone already has with a slight tweek.
This post first appeared on my website.