This is Pt 2 of last week's article. If you're just catching up now, go ahead and read that one first.
From Idea to Prototype
By now you already have a fairly clear understanding of what ideas you want to build and test. Now to build them. Again, if this is outside your wheelhouse it may be worth working with an Industrial Designer to produce a set of designs for prototyping. Some prototyping houses can also produce designs directly if you have a good contact, though your milage may vary.
Designing for Prototyping
At this stage it's entirely acceptable to design with complete disregard to production processes. While that will be a path that you will eventually go down, it is only necessary to design for the prototyping processes you intend to use in this moment.
What you wish to test with this prototype iteration will drive what prototyping processes you use. And what prototyping processes you use will drive the design. For instance, if you need flexible silicone components, you will need to cast them, so create your design to suit the casting process.
Take a look at your objectives for this prototype as this will define your prototyping needs:
- Does it need to be flexible?
- Does it need to be food safe?
- Does it need to be strong?
- Does it need to have a specific finish or coating?
- Does it need to be air or water tight?
The answers to these questions will begin to limit your options for prototyping. There are of course a huge amount of ways to prototype devices, but here's a list of a few main ones and what they're good for:
| Process | Good for: | Bad for: |
|---|---|---|
| FDM (filament) Printing | Fantastic for early stage prototyping. Very cheap. Particularly great for mock-ups. Fast to print and a huge range of filament types and colours are available. Generally pretty tough, especially carbon-reinforced filaments. Can also be used for silicone casting and lost-PLA investment casting. Good dimensional stability and large sizes are possible. | Very high strength parts or parts needing very fine detail. Not good for water or air tight parts. Surface finish is usually pretty rough, but can be sanded and painted. Some materials like PLA will creep if under constant load. |
| SLA (resin) Printing | Great for very fine detail and very small parts. Cheap to produce. Good for mock-ups that represent injection moulded parts, but don't have the same toughness. Excellent surface finish. Flexible resins are available, but isn't very elastic. | Not great dimensional stability. Can warp and the shape will creep if under constant load. Very limited material selection and they're usually brittle when fully cured. Small print sizes. Uncured resin is often toxic. |
| SLS (nylon powder) Printing | Moderately cheap. Extremely tough. Can also be steam sterilised and used in sterile environments. | Only a single material choice. White or dark grey, but both can be dyed. Rough surface finish. Not great at at fine tolerances, but can be manually worked. |
| Metal Printing (SLS or EBM) | Can produce very strong and light parts. Can be (and usually is) machined post printing. For complex parts may work out cheaper than CNC. | Expensive. Limited materials - usually just stainless steels and titanium. Poor surface finish, though SLS is a little better than EBM. |
| CNC Machining | Usually the second choice behind FDM and SLA. Can produce very strong and highly functional parts. Fairly easy to source from local vendors. Wide range of materials and finishes available. Usually it's possible to use the same materials you'd want to use in production. Can sometimes produce parts that look injection moulded. | Can get expensive. For small volumes the set up cost can be high. Capabilities will vary a lot from vendor to vendor. Local availability for some some materials might be a limiting factor. Won't work for most flexible parts, though you can make moulds for castings. |
| Vacuum Casting | Quite affordable. Great for high-fidelity prototypes you want to look like finished products. Usually uses a 3D printed part to create the pattern, and can usually produce 10-20 parts per mould. Tough, food-grade and flexible urethanes are available. Can be pigmented too for any colour you like. | Large amount of labour, which can increase time or cost. Local vendors are pretty thin on the ground. Dimensional stability isn't that great. Limited to PU resins only - silicone parts are usually not possible. |
Prototyping
Producing prototypes is something I take for granted. I have both FDM and SLA capabilities in house, and being in the industry I have a lot of contacts locally and overseas that I work with for more specialised processes. For someone not in the industry, however, it's likely a case of not knowing who to turn to.
There are hundreds, if not thousands, of prototyping houses locally and overseas that could make the parts you need. The best thing to do is ask around. Speak to other medtech founders, mentors, or program managers. Getting a recommendation or two from contacts (or contacts of contacts) will help you sort the wheat from the chaff. The more specialised the process however, the more legwork will be needed to find the right vendor.
The actual process is pretty familiar. Though it is always worth discussing with potential vendors your design. If they have any input to improve or simplify the design to make life easier, this can result in faster and cheaper prototyping. This is more true with local vendors. From my experience however, overseas vendors (at least Shenzhen-based ones) usually don't provide much feedback on the design anyway. They just tackle it head on and figure it out as they go. The language barrier can be an issue too with overseas vendors, however communicating with lots of images and annotations can minimise the miscommunication. I would encourage getting quotes locally and overseas, the price difference may not be as much as you might think.
Testing
Once you have your prototype in hand, this is where you need to go about answering the questions you had in the beginning. Ruthlessly test your device against your design inputs. If you're producing a prototype that tried to answer many questions at once, try and simulate as many use cases as you can. If you've isolated out a few key questions to prototype and answer, then test those. Get it in lots of hands and gather lots of opinions - even the ones you don't like. The important thing is not to be precious with your idea, your design, or your prototype. You want to, and have others, pick apart your prototype. The better your ideas and your prototype stand up to criticism, the more robust the product will be.
As you test and collect feedback, write it all down. Start identifying new things that need testing. At this stage of product development you want to "move fast and break things". This will reduce the risks early, and tease out the true answers to your questions. When you go back around the design process loop for the next iteration, you'll want to refer back to what you've learnt to redefine your design inputs.
Iterate, Iterate, Iterate
As I mentioned, expect to do this all again. Several times. Each iteration will continue to ask and answer more detailed questions, and will begin to include more stakeholders like quality, regulatory, and manufacturing as you go.
Remember, your users can make or break your device (figuratively and literally). Include them early to define your needs and test your solutions. It'll pay off.
If you've got any questions about your device, I'd love to chat about it with you. Book an intro call here.
