Prototyping In Engineering: A Step-By-Step Guide

Stop following the norms. Think outside the box to accelerate R&D. Story time: I needed to test a new printed circuit board (PCBA) that was going to be a 'hub' board for a few USB camera modules. But I hadn't designed it yet. How could I test it? In my previous life, I would sit down in CAD and model the entire system. I would get files to EE after a few days, and they would be sitting around waiting on my files. I didn't want to send them until I was happy with them. This would take a few weeks. Then, a few years into my career, I started finding shortcuts. Shortcuts exist when we know what QUESTION we're trying to answer. Without a question, we're just designing everything, and all the parts take priority. I broke it up into three main questions that needed answers: 1. Is the circuit board big enough for all the EE components? 2. Can I fold up a flex like I'm imagining? 3. How can I get moving on software in the meantime? Here's how I shortcut the PCBA prototype, step by step: 1. I created a scrappy DXF (2D) model (10 min) in Inkscape that I thought would roughly approximate the flex board that I was going to design. 2. I gave the DXF to the EE resource, and asked them to start the layout. It told them it was only representative, but it gave them something to start with. (I'll get them a real outline later) Now that part is moving. 3. Then I cut out paper/cardboard and taped up a rigid/flex board to test the folding inside the enclosure. Bingo. Confirmed enclosure size with ID. 4. Then I went on Amazon and ordered a series of flex ribbon USB-C cables and hubs to connect to the camera modules. This would be a surrogate prototype of the functionality of the custom hub PCBA. This gets the software team working on the image recognition bits. The most complicated tool I used in this workflow was the board outline software (Inkscape) and never touched CAD. Now with three 'prototypes' processing in parallel, I can dive into CAD knowing that end of next week, I'll be able to send 'final' DXF (based on reality) and also be able to start installing lasercut samples in my 3D printed enclosures. Look for shortcuts. #fastbuildlabs #design #engineering

Designing hardware products. The idea that you can simply create 1 or 2 prototypes and then just jump into full production is a big myth. The reality is very different, Using arduino/raspberry pi/STM32 and having something to show is just the beginning of the long journey in designing hardware products. What you have at this stage is usually more or less a proof of concept(POC). More often than not, you cannot ship a product to consumers with this design . A POC prototype cannot simply be bought and be brought to market. This often also most likely goes for many 3D printed designed projects. They are useful for helping to quickly build something and seeing if the idea makes sense at first. The next stage will involve designing a preliminary production design. You will need to come up with a system level block diagram and design your circuit diagram and come up with your PCB design. You order your PCB prototypes, 3D print the mechanical parts and buy other electro-mechanical components. Remember at this stage, you have not mass-produced anything. The PCB prototype and mechanical design fabricated would most likely not work as expected in the first try here. You will go through several iterations, fix bugs and run to other issues you least anticipate. Once you pass this stage, you can design your 3d Model custom enclosure which you may 3D print or use CNC machining as a prototype. Some people call it a Works-Like Prototype or you can call it a pre-production prototype. It is usually a little close to the final product your customers will see. The next stage is now testing and validation. At this stage, you will need to produce more units, about 10-50 units of your product. You may have to start working with your manufacturer and industrial designer at this stage. The goal of this stage is to validate that the working-prototype meets the functional, performance, and reliability specification. It is called Engineering Validation Testing (EVT). The next stage is design validation. You need more units here (usually more than 100 units).You obtain electrical certifications and necessary safety certifications. Depending on the country you want to sell, the certification differs and you need to obtain all this.  You validate it meets necessary design and environmental specifications. An issue or problem observed at this stage can take you to the beginning of the process again to correct your electrical or mechanical designs. Once you are done with these 2 stages, you can proceed with mass-production through manufacturing to create the final product. This is a simplified process for any consumer hardware product you may see. It holds true for robotic systems and products too. That is why you should validate and can even pre-sell with your potential customer first before any prototype. Don't start finding customers after you have a production-ready product. You are more likely to fail by doing this. 

What's the right number of prototype iterations in MedTech? Hint: the answer is not 1. That's like expecting a hole-in-one on a long distance golf shot. Just ain't gonna happen. Instead, focus your prototype iterations on answering specific questions: ➡️ Prototype 1: Does it work on the bench? Simplified proof-of-concept prototype that addresses key questions related to technical performance. ➡️Prototype 2: Does it work (pre)clinically? Early prototype aimed at collecting data, preclinically (for significant risk devices) or clinically (for non-significant risk devices). ➡️Prototype 3: Will people use it correctly? Usability prototypes (or mockups) aimed at evaluating user interfaces, usability, and possible misuse through human factors studies. ➡️Prototype 4: Does it achieve target COGS? Alpha prototype integrating industrial design and engineering, while designing for production materials and processes. ➡️ Prototype 5: Does it meet the requirements? Beta prototype addressing shortcomings of Alpha, and used for engineering verification testing (before V&V). So, minimum.. 5 prototype iterations. Often many more. Stage these prototype iterations so that each one gains the benefit of the prior. If you isolate these risk factors, your prototypes can be much simpler, faster and more cost effective to design, produce, and test. Prototyping is a mindset -- it's about learning, quickly and effectively. > Identify the right questions to answer > Build simple prototypes focused on the key questions > Run the tests, learn, and iterate. #medtech #medicaldevices #prototyping