This is the initial article that summarizes necessary steps that get you from your idea of a product to a finished version. It should comply with regulations and be manufactured with reliability. I will discuss here only the issues connected to engineering problems and I will omit business topics (i.e. business plans, distribution types), as they are a separate group of problems.
Creating a device that functions may seem to be an easy task, but the preparation of the product for production is a different story in the engineering world. In order to go smoothly through the whole design process, one should consider following some guidelines. However, keep in mind that they’re not sacred and can be reshaped, bent or even broken. Before you do that, please just ask yourself if you really, really need to do it.
I will divide the product design process into several subtopics:
2) Working Prototype
3) Final product prototype
4) Certified product
Then I will briefly discuss these points, as each step is a broad topic for several articles.
Let’s start with a nice and popular quote from software engineers:
“Weeks of programming can save hours of planning”
At the concept stage we should not only think about the shape and features of our device, but also do some serious planning and quite a bit of research. Ideally after this stage we should have a more-less detailed roadmap to our holy grail, which of course is a certified product freshly straight out of the factory. The results from this stage should be:
1) initial shape of the device,
2) initial casing documentation,
3) final/desired functional description,
4) regulations requirements,
5) REALISTIC production volume forecast,
6) estimated price target.
The information from all these points forms what we call - the Technical Documentation. Why is it so important to prepare all this information at such an early stage of the project? Why not leave the production and certification problems for later stages?
“But I just want a working prototype fast” – you say.
“We need to show something working to out investors ASAP, preferably next week” – I hear.
Well, if you can sacrifice the time and money for a quick prototype, which will be almost entirely redesigned in the later stages, then it’s fine. Usually, you will lack time or money and in some cases both. The earlier you'll start thinking about late design problems the better you can deal with them. Be aware, that you shouldn’t spend too much time on this stage. If you are caught in the endless loop of concept designs and decided to just scratch that 17 th prototype, then it's a clear signal that something is not right. Analysis paralysis is a pitfall which can affect you and will result in delay in the project.
I want to underline the fifth point regarding volume forecast. This may be quite hard to calculate because at the end you don’t know how your product will be received by the market. This parameter has a large impact on the design process, because it will determine the available technologies that can be used in the production. Furthermore, it defines prices in BoM and the amount of testing that must be performed on each piece.
Volume and mass production make low prices, so you should think twice before comparing the price of your 1000 devices to the price of 10 million devices from Singing Sam.
2. Working Prototype
This is the stage at which most of the start-ups think that the design work is done, and it is finally time to start production along with preparing rakes for the money inflow. Unless you’re aiming for <100pcs annual production, you still have a long way to go. The prototype at this stage should prove the functional feasibility of the project and should give clear signs that reaching the final shape of the device is doable.
3. Final product prototype
This stage should end with the product in its final form.
This prototype should not be a subject to any design changes apart from changes resulting from the certification process or from production feedback. Modifications here are costly and require significantly more effort than at earlier stages. Why is it so hard to change things here? Because the designers have to assess the impact on the whole design. Power budgets, thermal budgets, mechanical integrity, PCB layout, etc. must be verified again to ensure proper device operation.
One of the extreme cases that I can recall from my experience was when my colleague had to add a single 0603 resistor onto his PCB. It took him 2 weeks of layout redesign, because the PCB was already so densely packed that he had to move most of the components and traces to make space for the new resistor.
Optimization to BoM, production costs, reliability, testability are main topics for the designers at this stage. These require experience in the field and are main causes why start-up fail or underestimate budget and effort.
As Arthur M. Wellington said:
“An engineer can do for a dollar what any fool can do for two”
Investing work and time at this stage will result in huge savings in the future, simply because the product will be cheaper and you will not get as many service returns.
4. Certified product
When I ask start-ups about where they want to sell their product here's what they say:
“Of course, we want to sell it in Europe and USA… and China would be nice, and Japan too… oh and Brazil…”.
Believe it or not, but there are many different markets that have various requirements (often they exclude one other). When you hope that one can design and certify a device that will cover the whole world - it is just wishful thinking.
Let’s take a look at the two most common markets:
For the EU the product must comply with proper directives. Usually, these will be EMC and LVD, but if you have a radio installed then it will be the RE directive (RED). For the US market FCC and UL are most common certificates.
In general, as a manufacturer, you must show that your device is safe (fire, electric shocks, battery explosion, thermal, mechanical hazards, etc.) and will not disturb nor be influenced by other equipment (radio emission, conductive emission, radio immunity, ESD immunity, etc.).
Depending on the market, you will have to present documentation to respective institutions in order to gain the certificate. Some institutions, like UL, take this a step further and make the tests themselves. Additionally, depending on the product classification, UL will perform cyclic audits not only at the final assembly site, but also at suppliers’ sites.
At the end the costs associated with certification will be covered by the manufacturer (i.e. you). For reference, full FCC and UL certificate for a simple wearable device can cost roughly 20k$-30k$ and can take several months to finish. If you want to cover other markets, then the corresponding institution will also have their own fees. Think over carefully on which markets you wish to operate, because each one requires significant resource investments.
Summary and remarks
There's a certain saying that designers use:
“We design well, fast and cheap. However, you can choose only two options”.
Keep in mind that if you’re pushing one aspect you will have to pay with the other. This short article only indicated and summarized common problems that need to be addressed during the design of any product.
In the following article series I will try to discuss each problem more thoroughly.
Author: Rafał Widórek, PhD