Its integrated plate, one side of the switch will always be stiffer when we are talking about a switch near the edge like left control or caps. The alphas look uniform enough but why not just make a half plate with just the bottom row and top row having the integrated plate so that the pcb still has something to be mounted onto with switches. Your simulations are doing one keystroke at a time rather than multiple subsequent (sometimes concurrent) strokes within a short time period of each other. People also dont always hit the center of the keycap/switch directly at a perfect 90 degrees. Some people type heavily and others light. The simulations work only in a theoretical setting and arent really practical. A partial plate will provide as much consistency in typing feel without having to worry about whether or not those switch cutouts will break or bend.
Overall I think the plate is over engineered and you could've taken a much simpler, cost effective approach if you wanted to achieve uniformity.
Let's go through this bit by bit. Much like with Iron165, all engineering was done for the worst-case scenario of 2N or 200g force if you want to compare it to spring weights. From speaking with those who have typed on a 165 FE, this worst-case engineering process works for creating a more consistent typing experience, even if it doesn't represent the average user. Next, in terms of forces being not normal to the plate, I invite you to try and type off-angle right now. Likely, your finger just slipped over the key as switches are constrained to follow a normal path to the plate. If any sideways forces exist, they are due to friction between your skin and the keycap, a force so minimal it isn't worth time approximating. Finally, to address the one keycap at a time simulations: the fastest typists on here are at 200 wpm or so? With an average word length of 5 letters, that's 1000 letters a minute or 16.6 letters a second. If we generously round up, that means you depress a key every 1/20th of a second. Under a 2N load, the spring rate of aluminum is considerably faster than that.
- go observe your fingers resting on keycaps in typing form. It is not a perfect 90 deg. My argument is that its not a single point like your simulations use but rather an entire area that may or may not exert its force evenly. You can't assume a perfect force and say it transitions to a practical scenario.
- 200g is way more than what is necessary to actuate a switch. Just look at the most popular spring weights which usually fall in the lower to mid 60g weighting. Lets take the point above and assume that your 90 deg theoretical typing pressure completely applies to practical typing. You can make the plate consistent for 200g, but 200g is not anywhere close to how much force is applied per keystroke. Most people bottom out their switches whether harshly or lightly. Even with people who harshly bottom out, they might increase the spring weight to the upper 70s at most. Some outliers love weights above the 70s. Those are still obscure cases since the 70-89g range is still uncommon in most custom boards relative to 50-69g range. So you are basically adjusting the feel of the board for the .001% of people who might actually constantly type with 200g of force at a reasonable WPM. Consistency != flex. Your plate can be consistently stiff and still be consistent.
- Your simulations don't even account for a pcb. This means that even if we assume your hypothetical force simulation to be a 1-1 replica of real typing, it would still be inaccurate. You aren't accounting for the fr4 pcb that will go beneath the aluminum plate and whatever thickness that you choose for it.
Sometimes consistency shouldnt always the goal, and its seriously become a fad nowadays. Perfect consistency in keyboards is extremely hard to achieve and sometimes impossible with a given mount: Your F row, arrow cluster and nav cluster won't be consistent feel wise with your alphas or mods, its just a facet of many mounting styles and how you implement flex cuts. What matters is if the board provide a comfortable typing experience on the alphas and mods where typing is most frequent. I think you are going a bit overboard with the science and not looking into the practicality of things.
Okay, let's do this one more time.
- My simulations aren't a single point, they're force over area, with that area being the switch contact surface on the plate. With the way keycaps are designed and the way the switch physically works, the difference between pushing on the left side of the cap vs the right side of the cap is negligible when it comes to the force distribution over that area.
- 2N was an experimentally determined upper-boundary during the iron165 IC. The great thing about working with non-elastic materials (like metals) is that if flex is consistent at 2N, it's also consistent at all loads less than that.
- They do account for a PCB. You're just wrong there.
- I've never pretended I want the F-row, columns, and arrows to flex as much as the alphas flex. In fact, I'm doing exactly what you're suggesting: which is prioritizing the feel of the alphas and mods. That's the reason why there are no crazy cutouts in that section of the plate. I'm going overboard with the science to show that not only is this analysis possible in keyboards but that it has practical benefits to an end-user.