How to prototype and test – Insights from Paul Bristow, Co-Founder Pangloss Labs

17 - Dec - 2015

Hi, my name is Paul Bristow and I like to make useful things. As well as having decades of experience in product design – mostly around consumer electronics – I am one of the co-founders of Post Tenebras Lab, the Geneva hackerspace and a cofounder of Pangloss Labs, an open innovation centre and entrepreneurial, ecological Fablab in Geneva. In this blog post for the ICRC Enable Makeathon I’m going to share some of the things I’ve learned along the way to help humanitarian makers around the world.


Let’s start with some basic tips for rapid prototyping

Home-made front bicycle light mounting, made from Polycaprolactone (photo from Wikipedia)

According to Wikipedia, “Rapid Prototyping is a group of techniques used to quickly fabricate a scale model of a physical part or assembly using 3D Computer Aided Design data.” Usually using 3D printing. As a maker, I have a different definition: “rapid prototyping is any simulation of an object or process to test a proposed solution to a problem and get you closer to your final design”. Using this definition, rapid prototyping becomes an interactive process that helps you get to your final solution, testing along the way.

In practise, what this means is that you don’t necessarily need to use CAD, 3D printing, Lasers and a CNC from the start of the process. Often simple tools will be enough to get a first rough prototype and start testing. For example, if you need to make a long load-bearing tool, it might be easier to use an aluminium or bamboo rod, and fix the ends to a smaller part, which could be carved, drilled or even hand shaped with my favourite thermoplastic.

Polymorph, (also known as Polycaprolactone, Shapelok, or half a dozen other names) is a thermoplastic which melts to a play-doh like consistency at around 60C, can be shaped by hand, but is hard enough to make tools or cut threads in when it solidifies. It’s also reusable, so it’s great for making and modifying prototypes. Just use hot water to melt it down.


Technical Aspects of Rapid Prototyping

USB Microscope

USB Microscope (photo by Paul Bristow)

For 3D printing with different materials, the help pages from Lulzbot (or your 3D printer manufacturer) are really useful. When it comes to large prints with ABS or Nylon, avoiding drafts becomes really useful to avoid cracking of the print between layers and warping up from the bed, so you might want to consider walls around your printer. Don’t put the electronics inside heated areas or you might only make one print.  For prototyping, you are really looking for mechanical accuracy, but strength may be useful, too. If you are thinking of using a fibre reinforced filament (like the MarkOne from MarkForge does) then consider carefully the orientation of the layers with respect to your part. The part will be stronger along the layers and less so (sometimes a lot less) across them. If necessary test different orientations. Don’t be afraid to use online services to access printer technologies that you don’t have. It’s easier to take a too strong part and reduce it than take a broken too weak part and strengthen it in the same design.

One thing I’ve found really useful is an inexpensive USB microscope. At 600-1000x magnification you can easily see the effects of varying parameters on your 3D prints, and tweak the process (varying speeds, temperatures) to get the strongest parts possible. This is less useful for laser cutting but also good for CNC machining. Don’t be afraid to search online – there is a high chance that someone else has had the same problems as you and posted a solution – or at least a more effective thing to try.

One last lesson. Be aware of the resolution limitations of your toolchain. I remember very well my disappointment when I first CNC’d a round disk in aluminium. It seemed to have facets on its edges. It was only when I went back and magnified my CAD model onscreen that I realised that my CNC had exactly replicated the facets in my model.

User Testing and Product Testing

User Testing should start way before you have a product to test. The Human Centred Design Methodology is a great framework for sitting down with your target users – and the people they interact with on a daily basis – and finding out from them what their problems are and most importantly, which ones are the biggest problems. These can range from show-stoppers (because of X I simply can’t do Y) to annoying niggles that are repeated a hundred times a day (so solving them would have a significant positive impact on the quality of life) . If you aren’t familiar with Human Centred Design then is a great place to start, and has a short and succinct Field Guide you can download to help you through your first time using it. This will help you nail down what problems you should be solving for your target users.

Most importantly, don’t have an ego when you are testing. If you get into an argument with your end-user trying to defend your design that is a sure sign that you are wrong!

Once you have a physical product to test it starts to get interesting in a different way. You must test with real users. But you must be very careful with which questions to ask them. Don’t ask closed questions like “do you like the product?”, because they will probably say yes to try to please you. Which is nice, but doesn’t help improve the product. Ask them open questions like “How would you use this in your daily life?” or “which problems do you think this would solve for you?” and listen carefully to the answers. You will almost certainly discover insights you would not have thought of, because you don’t live with the problem!

When you think you are ready, try using something like the system usability scale to get an independent view of your products usability. You can download it here. This will tell you if you’ve done a good job of design. But don’t forget the most important question at the end: “Would you like to have this product?” In this field, even if your design is not perfect, it may be better and less expensive than anything else available.

When it comes to testing the product itself, you must test the basic physical characteristics of the product in use (“will it bear the weight of the user?”, “can the user fix it into position by him/herself?”) but also in abuse (“what happens if the user jumps on it?”, “what if bullies pull it off?”, “what happens if someone hits someone else over the head with it?”). You must also think about storage or charging if needed – I’ve seen some great products that have fragile charging points that are awkward to connect and break, ruining the product.

What is Human-Centred Design?

To learn more, please watch this video from

How, when and if to jump to mass production?


Modern Assembly Plant (photo from flickr)

First of all, you need to be sure that your product is one that lends itself to mass production. If it’s 3D printed and every part must be customised to each individual user you are probably going to make it massively available without mass production through fablabs, makerspaces and maybe 3D printing networks.

If it’s a product that does need to be mass manufactured, you need to make the best possible estimation of your market and how much of that market you can win. The ICRC have made some market studies available to help.

It is unfortunately easy to design a product that is not mass manufacturable. That definition changes based on the manufacturing technique. 3 axis CNC is more limited than 4 axis CNC but less expensive. Stamping and folding a metal case might be cheaper still. 3D printing a million parts might cost a million times printing 1. The right manufacturing methodology will vary based on the design and the number of units you expect to ship. Getting a handle on the manufacturing techniques that will be used and learning their limitations is key. So find someone for your team who knows this world and listen to them!

You also need to be absolutely sure that your design is correct. Remanufacturing 10,000 pieces with just one hole missing can be really expensive. Talk to your manufacturing partner and get their input into the design. They have probably made 1000’s of different products – you haven’t. Use that expertise. Be aware that tooling charges (tooling is things like the moulds for injection moulded parts) can be quite expensive – certainly $10’s of thousands of dollars and easily into the $100’s of thousands. Given that you will need to spend that money before you ship a single product, it’s very important to be sure you have the funding in place to bridge that gap. Crowd-funding (a method of collecting pre-orders) is one way – but be sure you have already found a manufacturing partner or you will not be sure of your costs, and hence your pricing could be way out. A majority of failed crowdfunding projects are because the mass production costs were underestimated and so the price was set too low. It’s always easier to reduce a price after you have started selling than increase it.

There is a third way – have some parts of the product mass manufactured and some parts locally customised. Quite a few medical braces use the thermoplastic I mentioned earlier to allow nursing staff to perfectly fit a standard brace to the patients shape. Another way might be to have mass produced metal joints in a 3D printed prosthetic to add strength and durability without losing the ability to custom fit.

In short, assuming you have the finance in place, you are ready to manufacture when your design is final, working and manufacturable, not just when it’s working.


Finding the Right Team

If you want to build a business about your solution you will need a team of people. You cannot do not it all yourself. So how do you find a team? Who do you need? First of all, find a book in your native language about this. You will need someone capable of writing a business plan, someone capable of sales, someone capable of buying the components, someone capable of getting the product built. These are not the same skills as designing the product. Basically what you need is a team of people that each love doing their part – that way you will get the best performance. An engineer trying to sell will neither enjoy him/herself nor be an effective salesperson. The best way to find these people is to talk about your project at local entrepreneurial meet ups, or startup pitching nights, not just saying what you are doing but why. Your goal at this point is to find people with the same values to be part of the team, not (yet) to raise money. If you can’t find a local meet up, organise one. Who knows, you could be responsible for creating a startup community in your town.


Key Lessons for Prototyping something intended for production

Iteration, iteration, iteration. I’ve talked about human centred design, design for manufacture and the need for user testing. Apart from building a product that actually solves a problem, can physically be built and is useful to the target end-user, iteration (design – prototype – test – design) solves another purpose. It lets you catch design problems. Every product manager has learned this golden rule over time.


“The earlier you catch a design problem the cheaper it is to solve.”

And not just a little bit cheaper, generally the cost of fixing a design problem goes up by an order of magnitude with each step in the process. So at the concept stage it might mean changing a few words on a page; at the rough design stage redo a sketch, at the prototype stage rebuild something; at the design stage redo the engineering drawings; at the tooling stage add another $10k, after the first trial production run $30k to fix something and after mass manufacture the costs can easily be $300k to $millions.

So catch your errors early. Good Luck!

Paul Bristow

Paul Bristow  is the Co-Founder of Post Tenebras Lab, the Geneva Hackerspace, and Co-Founder of Pangloss Labs, an open innovation centre and entrepreneurial, ecological Fablab in Geneva. He led the webinar on Rapid Prototyping and Testing for the Enable Makeathon and offered to share his insights in the form of a blog post.

Links and resources mentioned: