Testing Saves: Break Something… Before Users Do
The models and expectations you have built may fail to capture all the interactions of the system that you need in order to ensure a successful product enters your users’ hands.
Real-world testing saves us from ourselves. Testing reduces long-term costs, prevents late-stage schedule failures and, more critically, create superior products that serve the users.
I was once very frustrated at a palm tree. I was in college taking my first physics class. I never had one in high school because I took biology. But I was a mechanical engineering student now, we did physics, and a tree befuddled me.
I knew about the basics of Newtonian motion yet knew nothing of material science and the interplay of chemistry and the molecular world relative to the simplifications of Newtonian modeling and the ramifications at the macroscopic level.
I understood that energy was conserved, and I lived by the assumption of non-compressible rigid bodies, which could not explain a palm tree gently swaying in the breeze. Within weeks of Intro to Physics I understood elasticity and the macroscopic effect of covalent bonds. My knowledge had been incomplete previously, and I had failed to consider the complexity of the interaction of physical forces.
Capturing all of the details and creating testable procedures is challenging when time and resources are limited, which they always are. You are left with what is possible.
On that note, I was also confounded by a skateboard. I developed one in college and wanted to calculate a real world impact force for an average skater when they stomped on their board. I was setting a metric for our composite design and thought I could model it on paper.
I was convinced the engineering tools I had mastered would cover the physical definitions.
After visiting my dynamics professor I found that the power generation of the human body was beyond accurate assumption. I needed empirical data only supplied by cutting edge equipment of the time.
Instead, I made some design assumptions on weight and structural support. We put laminated maple boards into a force cell and found the failure strength. This became our target for the composite board we designed and tested.
We couldn’t capture all the details so we tested the devices available to us.
Now my work involves myriad factors. The medical industry has one of the most challenging environments for a manufactured device: chemical exposure, multiple daily users, biological contamination and the breadth of human morphology to accommodate. All of this is compounded by the rigorous standards, specifications and technical demands of the field.
Why Medical Products Need Physical Testing
The only way to coalesce many complex issues into empirical evaluations is physical testing typically of the destructive variety:
My catch fits the design guidance and can be molded. The flow analysis confirmed it. A drop test could shear all the catches off the housing and leave my internal working in a pile. I may now add material thickness to avoid this before my production tooling is cut and finished.
The selected material claims durability to a cleanser through its spec sheet. My chemical testing showed minimal cosmetic damage but a 50% loss of tensile strength. Time is lost revising calculations, updating CAD models and drawing notes, adjusting specification and evaluating new material. However, no devices will be recalled tarnishing reputations, endangering the users and overrunning resources.
The button face of the device is push tested and looks fine. The button no longer works. We can shore up the supports in the mold or on the PCBA before the intended users break the switch first.
The assumptions of function must be confirmed when all the constituent parts are present. The number of ways a device can fail exceed any modeling on hand. The ways parts interact with each other and their environments are so complex that only real world testing will shake out all the gremlins.