Watching my father-in-law struggle for about an hour to set up this year’s holiday train set gave me another opportunity to reflect on manufacturing at scale and specifically the quality aspect of a product.
Every year, for the past three years, my in-laws have purchased a plastic train set at a bargain price for my son, who is 4 years old. Having over 20 years of manufacturing experience, I cannot help but look at the product from a manufacturing at scale/engineering perspective. And I am usually seeing a lot of areas for improvement. So this year again, I opened the set, eager to witness the improvements of the latest generation. The last two years, the rails connectors broke after one use. Would that be the case for Gen 3? Or will continuous improvement bring this year's set to a more acceptable level?
In the previous generations, the combination of design for the locking mechanism and material did not agree with each other. The assembly was difficult because the clips were soft and bent when not properly aligned, often causing damage to the parts. Once in place, no satisfying clicking sound was heard (partly because the material in use was not stiff enough). Disassembly was nearly impossible without breaking the parts (human fingers did not seem fit to reach the locking mechanism to unlock the rails, which is either a big omission or proof that the intention was to create a single use product).
This year though I noticed a change (for the better) in the type of locking mechanism. It is not "clipping" in place but using gravity and long contact surfaces to keep the rails aligned and close to each other. Given that there is no snapping action, there is no satisfying clicking sound, so using a soft plastic is not an issue. It may even open the possibility to widen the raw material specs, allowing a lot of regrind and/or recycled material to go into the final product. Not having acquired a Differential Scanning Calorimetry machine yet, I cannot be sure. But I am toying with the idea of burning a sample in a semi-controlled manner to measure the solid content (wouldn't it be nice to find some glass fiber in there, displaying an attempt to increase the stiffness?)
Another disappointment of previous generations was that the rails were distorted. When fully assembled, there was a mismatch at the junctions. I am confident that it looked great on the CAD model. But I am also thinking that shrinkage during the cooling of the part in the mold (most likely injection molded) was not taken into account nor was it controlled. I understand that it can be a bit tricky, especially for the curved sections which have a bit asymmetry. Also the thickness of the bottom part of the rails varied from that of the cross ties (a.k.a sleepers), causing stress and deformation. The shrinkage is material dependent. So even if the factory controlled its injection molding parameters tightly, the likely variation in raw material properties caused by the use of recycled material or regrind would make control nearly impossible. This year though, it seems that the thickness has been kept more even throughout the rails. Whether that was enough or not, the alignment is better.
"Resist the urge to over-engineer your product"
Despite my disappointment at seeing a product on the market that was not designed with great quality in mind, my son enjoys playing with the train. The end customer is therefore satisfied. So in addition to a fun exercise of holiday time reverse engineering, I am also reminded this rule of product design: Resist the urge to over-engineer your product or you might miss the opportunity to sell thousands of products, risk not delivering in time, and maybe even endanger the success of your company.