3D printing or mass production — when does it make sense?

Traditionally to produce your product you had to go via the mass production route. With 3D printing — and digital manufacturing at large — you have a second affordable option available. Of course it depends on the product itself, the components it consists of and the expected volumes which production method makes sense.

3D printing makes sense for one-off and small series production. That is the obvious answer. The more elaborate answer is that it depends on the actual volume of the product and its parts. Redeye (division of Stratasys) made some calculations and put them on the web. They came to series of between 100 and 250 parts. It depends on the material and complexity of the part but in essence when volumes go up mass production methods like injection molding start to make sense. I put in chart to make it more tangible.

The setup costs of mass production methods make it prohibitive for small series production. If you take injection molding the tooling cost start $10.000 and up. The setup costs for 3D printing are virtually zero. That makes the difference.

Other considerations are material, part complexity and source location. 3D printing has a limited set of materials and may or may not suit the requirements for your parts. The complexity of the part is also important. More complex parts require more tooling costs while with 3D printing the cost impact of part complexity is low. 3D printing even allows to make parts which you cannot make using mass production methods. But that is something for another blog post. The last consideration is source location. 3D printing is available around the corner in most Western countries and thus be sources locally effectively saving on shipping costs and lead time. Mass production typically takes place in East Asia because of lower costs.

In small part series 3D printing can be a viable option. I expect prices of 3D printing to go down significantly in the next 5–10 years effectively raising the bar for traditional production methods. Maybe 3D printing will even replace traditional manufacturing methods all together.

Iterative design, 3D printing, co-creation and marshmallows

3D printing makes it possible to produce unique one-off designs for reasonable costs. This enables designers to implement design improvements based on customer-feedback. It is one of the aspects of co-creation where consumers and designers work together on a design.
Using iterative design a design can stay relevant for longer times — theoretically even forever. Technological capabilities are constantly improving and products can take advantage of that by using the new capabilities of the design. Also the expectations on what is fashionable and what is not changes over time. And last user experience can be used as input to improve function. Parts can break due to wear and tear.

Traditionally the concept of iterative design is employed within product design teams. You see it a lot in software development — especially in agile development environments. Developers or designers quickly deliver a prototype which is shared with the whole team and sometimes customers. They provide feedback which is used an input for the next iteration cycle. Also the open source software development mantra “release early, release often” is trying to achieve the same thing.

New technologies like the internet but also 3D printing make it possible to bring the product development cycle into the open and allows for faster feedback cycles. Especially on-demand production and zero stock policies make it possible to adapt a product design immediately.

There is a game often used to in team training session which is called the Marshmallow Challenge. Each team gets 1m / 3ft tape, 1m / 3ft string, 20 sticks of spaghetti and a marshmallow. The objective is to build the highest stable freestanding structure with the marshmallow on top. It is interesting that children are much better at this than adults. Children start building immediately and start over when they fail. Effectively they are using iterative design to come quickly to the best solution.

Recipes for your 3D printer

For several years now I am following Mark Ganter from the University of Washington. He works in the Department of Mechanical Engineering. Together with Duane Storti he runs the Solheim Additive Manufacturing Laboratory.

In their laboratory they work on new material recipes for 3D printing. They publish their findings for everyone to see and use on a blog called Open3DP. A lot of 3D printing companies keep their material a trade secret — with exception when they use really common materials like ABS or Nylon. To me Open3DP is a fresh of breath air.

They have come up with some interesting new materials like:

 Wood Wood  Glazed plaster Glazed plaster

I love what Mark and his team are doing and I am always keen on hearing from Mark what he has been up to.

DMCA does not cut it for 3D printing

The Digital Millennium Copyright Act (DMCA) was signed into law in 1998. It is designed to protect copyright and implements two WIPO treaties. There are many controversial paragraphs in this law but there is one very useful provision. This is the Safe Harbor provision called Online Copyright Infringement Liability Limitation Act or OCILLA. It limits the liability for online service providers on copyright violations by their users. By implementing a strict Takedown Notice procedure these online service providers can get themselves out of the dispute.

With the rise of 3D printing and their accompanying marketplaces and 3D model sharing services a new type of infringement is becoming more prevalent. Copyright infringements of 3D models and 3D printed products are rare. A form or function cannot be copyrighted and only patterns on objects can fall under copyright. Infringements do happen but they are in the area of trademarks, trade dress, patent and design patents.

Unfortunately there is no protection for online service providers for infringements in these areas. In practice this means that online service providers are responsible for the activities of their users and are liable contributory trademark infringement. The same reasons for creating the Safe Harbor provision under the DMCA apply to the other intellectual property rights.

Like we have seen with digital media entering the internet we can expect users start to copy, remix, recreate and improve existing products using 3D printing. Sites that accept and make available user generated content are seeing this already. It is only a matter of time that these sites become interesting enough to rights holders to sue them for infringement. Whatever the outcome these companies will be seriously affected. If left alone this issue will create a lose-lose outcome for all parties involved. A genuine disruptive innovation cannot be stopped and should be dealt with. Rights holders should be protected while this innovation can flourish.

The issue with regard to non-copyright related infringements already exists for companies like Etsy and Ebay. Ebay has been sued several times already and has recently won a case in the supreme court Ebay suffers mainly from users who sell counterfeited goods. To limit their liability and under pressure from the industry they have implemented very strict rules and programs to deal with these issues. This helps them in these cases.

Safe harbor has spurred innovation on the internet and there is a need to extend this to the other IP-related rights as well. We are still early in this cycle of innovation but I hope we can create protection for online service providers in these areas as well. The worst thing which can happen is that we enter a new period of lawsuits slowing down the innovation of personal manufacturing on the internet.

3D on the web — WebGL to the rescue

There is at last progress on bringing 3D to the web. For years, different companies have tried to solve this problem, but the solutions — often with proprietary browser plugins — were cumbersome and limited. With the creation of WebGL, there is finally a way forward.

There are a host of solutions to bring 3D to the web. There are straightforward solutions like animated GIFs of pre-rendered 3D images or a little more advanced like javascript animated JPG or PNG images with limited interaction with the 3D model. These solutions are extremely well supported with most browsers — even mobile ones. A notable example is the 3DNP (3D No Plugin).

Adobe Flash supports 3D since last year and the same applies for Microsoft Silverlight. Flash has a sizeable installed base and is truly platform-independent except for the Apple IOS platforms. The plugin suffers from performance issues on a certain platform like Linux. Silverlight is very much geared towards Windows and Mac OS X. The installed base is quite low, and Microsoft seems to have lost interest in moving Silverlight forward. They are currently repositioning it for Windows 8 mobile. The 3D engine is still immature and needs improvements. I expect that Microsoft will improve it for the next Windows Mobile version to enable gaming.
Performance wise both plugins are limited by their underlying technologies. Adobe cannot go beyond 50,000 polygons to achieve smooth animation and Silverlight is even more limited than Flash.

Another option is Java. Java has a mature openGL implementation called Jogl. It provides hardware accelerated 3D engine which is extremely powerful. The major issue for real 3D performance is the limited memory available to an applet. An applet can only use 64MB, and this effectively limits the polygon count to around 250,000. Another problem is that OpenGL support is buggy. On Windows OpenGL support is provided by the video card vendors, and the level of support varies per vendor. On other platforms — like Linux — it is the same story.

This year WebGL came to life with support from Google, Apple and Mozilla. It is build on top of the HTML5 canvas-element. It is managed through Khronos standardization group who are also responsible for managing the Collada 3D file format standard. Microsoft is sitting on the side lines without any support for WebGL in Internet Explorer.
In short, WebGL exposes OpenGL calls to the javascript engine and uses the HTML renderer to show the result. This enables far better integration with web applications, and this is already shown by online 3D editors like 3DTin and Tinkercad.
WebGL suffers from the same problem as Java applets using Jogl and that are the buggy OpenGL implementations.
For compatibility reasons, the number of vertices are limited to 65k (16-bits) per mesh which is quite low. Fortunately multiple meshes are supported and scenes or models with up to 500k triangles stay interactive in the browser.

To me WebGL is a significant move forward in making 3D possible on the internet via a browser using an open standard. It has all the conditions to be a success. The biggest challenge is to keep WebGL secure. It exposes direct hardware calls to javascript, and it depends on the filtering techniques of the browser javascript engine to make it secure.
In the end, I am confident that WebGL will succeed. We finally have a 3D solution for the web. Now it is time to make use of it.

I would like to close with this extremely impressive demo from Google about the capabilities of WebGL:

Pirated 3D Models

Some people ask me what keeps me up at night. One of the answers is Intellectual Property and personal fabrication. At the moment we get about one DMCA Content Notice Take Down request a week at Shapeways.

Michael Weinberg wrote an excellent report titled It Will Be Awesome If They Don’t Screw It Up on this. He writes:

[…] it is critical for today’s 3D printing community, tucked away in garages, hackerspaces, and labs, to keep a vigilant eye on these policy debates as they grow. There will be a time when impacted legacy industries demand some sort of DMCA for 3D printing. If the 3D printing community waits until that day to organize, it will be too late. Instead, the community must work to educate policy makers and the public about the benefits of widespread access. That way, when legacy industries portray 3D printing as a hobby for pirates and scofflaws, their claims will fall on ears too wise to destroy the new new thing.

A casual glance over at Pirate Bay reveals that the number of infringing 3D designs is still limited. I did found a collection of Transformers models which obviously infringe on trademarks owned by Hasbro Inc.
I find it telling that there is a whole page about trademark on a Transformer fan community site.

For a change I do like how Apple explains very clearly on how 3rd parties can use their trademarks and copyright need to be used. I think it is an example on how to cooperate with your community — either commercial or non-commercial.

There is both opportunity and risk involved for the existing manufacturing industry but trying to stop it is not an option. We have seen what happened to the music industry. They are still recovering after decades of litigation and falling revenues. But they never have been able to stop it.

Design meta language

One of the problems in 3D design is that 3D models do not — or to a limited extend — capture any design intent. The technical requirements for manufacturing a part or product are hard to extract from a design let alone the functional requirements.

This is already a very basic problem in the 3D printing industry with regard to material specifications. There are multiple material printers on the market but they are nearly impossible to use because 3D software does not capture how things are made.

But the requirements go beyond materials alone. Parts need to have particular properties to function as intended. The production process is in itself less important and should be determined by the available resources. These properties need to be captured in the design.

There is also something like design intent which is also not captured. If a designer gets a design from another designer the design intent is not captured in the 3D model. This makes it hard for another designer to make adaptations to that design. He needs to reverse-engineer the design intent to be able to do that. Imagine an adaption of design based on a particular functional requirement. For instance you have an USB stick and you want to change the design to micro-USB. Fundamentally the impact of that decision is low but without knowledge on the actual design it hard to make that adaption.

The current file formats are very poor at capturing design decisions. There is a need for a Design Meta Language on top of the existing file formats which allows designers to store intent, function and properties of parts and components.

So why is that important you ask yourself? Well for one to make it possible to let non-designers customize designs without the need to have a designer available. There are situations where that is not feasible like a war zone or in space or when it just too expensive like in most consumer applications. Consumers can improve designs and share them with others. They can improve or adapt it further and so on. It is called iterative design.

3D printing in space

Today an article in Space.com appeared about tests a company Made In Space did with two 3D printers during a few zero-gravity flights. It is unclear from the article if the tests were successful or not. Regardless there are a couple of reasons why 3D printing makes so much sense.

The first requirement is for manned space flight into our solar system is that we can actual do manufacturing in space. At the moment all stuff is hauled from Earth and brought to space. Sometimes some assembly is required but we send mostly finished products in space. This limits us to wander very far from earth. The production, testing and shooting things into space is extremely costly and time-consuming.

The second requirement is that we need to be able to fix what is broken — even far from home. When we send people to Mars you cannot just order and replace a broken part. Because of this reason spaceships and space equipment are build according to the highest quality standards possible to avoid that they break. But of course things break nonetheless. Just imagine a design flaw which causes a part to break. If we would lower quality standards and can accept things will break in space the cost of design and manufacturing for space equipment can go down significantly.

The third requirement is when we venture further in space we cannot prepare 100% for what we will find or encounter. We need to be able to adapt existing equipment or make new ones.

The fourth requirement is that we simply cannot take everything with us.

Here comes 3D printing to the rescue. It offers a few solutions to these problems. 3D printers allow to manufacture on the spot using basic materials. On a space mission only these basic materials should be on board. I can imagine that we would mine local resources like moon dust to build parts.
When parts fail during a mission because of design flaws astronauts can modify the design — or even receive it from earth — and build an improved part on the spot. New parts can be created as well and produced for opportunities or problems we could not envision when the mission started.

The current state of 3D printing is still not up to the level it is good enough to actually solve the aforementioned problems but I am confident that we can get there. I can only imagine what will happen when NASA would put her weight behind this technology and actually starts to move this industry forward like she did in other industries as well.

3D printing: so what’s it called?

In general 3D printing is called 3D printing. But there are many other names or acronyms used for this technology.

So far I have seen:

  • Additive Manufacturing (AM) — as opposite to subtractive manufacturing (CNC and others)
  • Additive Fabrication (AF)
  • Rapid Prototyping (RP) — old term when the technology was mostly used to make.. well.. prototypes
  • Solid Freeform Fabrication (SFF)
  • Additive Layer Manufacturing (ALM)

I am not sure who thought of the term 3D printing. Personally I favor 3D printing simply because most people immediately have some sense on what it means.