Because Over-Engineering is Underrated
Rapid Replacement Part Production
While they work faster and do a better job of actually getting clothes clean than a more modern washing machine, maintaining a late 1990s to early 2000s washing machine is not without its challenges. One of those is the lack of easy availability of spare parts for the things that will inevitably break. Washing machines of this age are recent enough that the parts are there, but old enough that the one part that breaks right before the weekend is unlikely to be stocked locally. Jenny is dubious about the value of custom manufacturing readily available items, but she's even more dubious about piles of dirty laundry that smell vaguely like other dogs.
PROBLEM STATEMENT
The Maytag Atlantis series of top loading washing machines use a lid safety switch that disengages the spinning drum when the lid is lifted. That's a reasonable safety precaution - the internet is full of horror stories about everything from crushed fingers to broken arms from trying to reach into a spinning machine. The safety switch on this model is located next to the rear of the lid. When the lid closes, a cam actuator clipped to the rear of the lid (the technical term used by engineers for this is a "pushy-in thingy") pushes a plunger in that activates the safety switch and allows the drum to spin. Lifting the lid moves the cam actuator off of the safety switch assembly, and the drum stops spinning.
Unfortunately, the part is plastic, and degrades with time and exposure to the heat and vibration of the washing machine. Also, they are designed to fail after 5pm immediately before a weekend, which is when this part failed, and needed to be replaced. Fortunately, replacing the lid cam actuator is a 2 minute job requiring zero tools - if the part is in hand. Otherwise, the washer is deadlined.
AoA: ANALYSIS OF ALTERNATIVES
At first, this seemed like a procurement issue, not an engineering problem. Custom engineering a replacement part normally only makes sense when the part isn't available through normal channels. A quick online search indicated that the part was available from multiple vendors, at prices ranging from $9.45 (which seemed reasonable) to over $25 (which seemed astounding.) While the nominal cost appeared modest, the actual cost of getting the part within a few days skyrocketed into the $35-$50 range. Most vendors offered two or three shipping options, ranging from one-day shipping at astronomical prices (which, coming into a weekend, really meant two or three days, once they heaved themselves out of their 1970s style padded rolling office chair, picked the part from their warehouse, and sent it out), to ground shipping which was expected to take somewhere between two weeks and right before the heat death of the universe.
So ordering the part was out. Having it in a few days would cost about 35% of the $100 cost of the washer in the first place, and waiting for weeks to wash clothes was a non-starter. Jenny insists on freshly laundered towels still warm from the dryer after her frequent baths (to remove the smell of dead things she has inadvertently rolled in while on squirrel patrol), and Steve is extremely fastidious almost to the point of obsession (he may be an engineer, but he's still a cat.)
The field expedient solution of using a small screw drilled through the lid to replace the cam actuator was also considered and ruled out. There was the risk of damaging the switch assembly, leading to another repair, and it would have scratched the appliance paint leading to (more) eventual rust.
Another alternative we considered was opening the washing machine top cover and bypassing the switch, or, even more simply, just jamming a small metal washer into the slot to hold it down, while we waited for the part to arrive. Steve was adamantly opposed to the HomeSHA violation - he deeply disapproved of any exposed rotating machinery that was above his line of sight, but within easy leaping distance. So that alternative was discarded.
Unexpectedly, custom engineering a replacement part was the best alternative.
The OEM part that our part vaguely resembles.
SOLUTION DEVELOPMENT
Fortunately, we were able to recover most of the fragments of the original part to take measurements. Between tracing the actual cam "nose" onto an engineering pad, using a caliper to take measurements of the lid width and the mounting holes in the lid itself, using a feeler gauge to determine the distance between the body and the rear of the lid when closed, and using a graduated rod to determine the actual available plunger travel, the measurements took longer than actually designing the part - especially since it didn't become apparent that we needed some of them until we were in the design phase. Even with the iterative measurement requirements, though, the part came together quickly.
Because we weren't attempting to duplicate the injection-molded geometry of the OEM part exactly, we could make the replacement somewhat stronger, albeit somewhat less flexible. We printed the part out of PETG, upped the bottom and top shells to 6 and the wall loops to 5, and used a 50% gyroid infill. It flexes easily enough to snap into place, but stays snugly in the existing holes. Because it had parts sticking up from both sides (the cam actuator on one side, and the clips on the other), it would normally have required supports to print. Both to provide a stronger print orientation and to save us the trouble of removing supports and sanding, we printed it in two parts - the actuator slips through the body with the clips. We glued it into place to make installation easier, but it would have worked fine without that: the body of the part holds the cam actuator component snugly against the lid, so even if the glue fails, the actuator wouldn't separate from the clip body once installed. Also, the original Maytag part is white to match the lid, but we printed ours in black, because it's what we had in the AMS at the time.
Those measurements led us to add 2 mm to what we think was the original cam length to ensure reliable operation (without creating additional stress by bottoming out the plunger.) After a quick review of the mechanical operation of the actuator and plunger and how they engaged with each other (thanks to YouTube, we didn't have to feel tempted to take anything apart to see how it worked) we did keep the asymmetric shape of the original cam.
Because our goal was to create a functional equivalent instead of an exact duplicate of the Maytag part (W10814230, if you're curious), we created a more compact design with clips facing inward to snap around the lid fitting and a slightly longer cam. (In researching the replacement, we found commenters on Reddit who noted that slight lid deformation can prevent the stock cam from fully depressing the plunger. Because we at Spinning Jenny trust random Reddit comments implicitly, we measured both the gap between the rear of the lid and the washer body, and the available plunger travel in the switch assembly. )
RESULTS
Unusually, the part fit correctly on the first try - because it was so small, we skipped the usual PLA prototype. It works as well as the original part, and it took less than an hour from start to finish - less than 30 minutes to design (even with walking the length of the house between the laundry room and the OpenScad computer several times to confirm measurements), about 15 minutes to print, and 5 minutes to install. Less time than it would have taken to drive to the local parts store, back when there were local parts stores, and back home, and at a printing cost of about twenty-six cents.
The part works as well as the OEM part (and we could replace it more than 35 times for the cost of the OEM part, even ignoring shipping costs) and we were back to washing clothes the same day. Jenny took a certain amount of satisfaction from a simple engineering task done quickly and correctly - not, mind you, the same sense of satisfaction that barking triumphantly while standing on a mound of squirrel skulls would bring - but satisfaction nonetheless.
