The book mentioned that one of the studies found that 3.0 mm was optimal, but that other studies just suggested a range that includes that length. Out of curiosity, I measured the throw of my favorite switch to date, the Alps on my Apple Extended Keyboard. Using my Mark I eyeball and a ruler to take multiple measurements, I consistently got a throw depth of 3.0 mm. That seems to back up that idea.
I think the typical Alps switch has about 3.5mm of travel to bottom out.
But it also seems (based entirely on feel) that the Alps activates at 2.0 mm, just like the Cherry. That makes the total throw length unimportant for Cherry users that don't use the full 4.0 mm of travel.
Alps switches definitely have noticeably less travel to the actuation point then Cherry MX switches. Hopefully I can measure this precisely for a variety of Alps switches in the near future.
It got me wondering how to test various throw lengths while avoiding the problems of varying switch types. It seems to me that the throw and activation points of Cherry MXs—especially ones like the clear—could be modified by adding material to the top and bottom of the slider. Of course, I wonder if its even economically feasibly to produce such sliders for that kind of test.
To increase the post-actuation travel distance, you’d need to take material off the bottom of the slider, and also add material to the top of the keycap mount. To decrease travel-to-tactile-point, you’d need to move the little bump on the slider upwards. Actually changing the actuation point may be a bit tricky. I haven’t thought enough about the inside of MX switches.
I think some other type of switch would likely be better for doing this kind of testing. The one paper which tested a bunch of possibilities built their own single switch with some parts they could tweak. (Obviously that has it’s own problems, as typing is very different from repeatedly pressing a single switch in a single position).
It also seems to me that the jailhouse blues mod is exactly this type of shortening of the throw. It makes the activation point sooner and the total throw shorter. Has anyone done any testing to determine if this kind of shorting is beneficial or just preference?
I personally think it’s beneficial to have the actuation point higher (say 1 or 1.5mm into the keypress), but maintain a substantial amount of post-actuation travel distance.
Alps is a pretty reasonable actuation point I think. I wish Alps switches had a softer bottom landing.
I find that the Model M and Model F have an actuation point which is lower than I’d prefer, and less post-actuation travel than I’d prefer. Other than that the Model F seems very very good, in terms of the way the tactile point feels, the overall weight of the switch, etc.
Cherry MX switches don’t have a very good tactile feeling in my opinion (I don’t like the plastic-on-plastic friction, or the shapes of the force curves). Their actuation points are lower than my preference.
I actually think the important measure may be amount of work required to actuate, rather than force to actuation or distance to actuation, per se. I find that the SMK alps-mount (“monterey”) switches have about the same pre-actuation travel as Cherry MX switches, but because the first part of that is very low force, the actuation point feels subjectively higher. They also have a sharper click than MX switches.
Heck, I wonder if Cherry has done that kind of testing. They would certainly be better equipped to conduct such studies than most.
Probably, but considering their target market has mainly been POS keyboards, and keyboards for office and industrial applications, where reliability and longevity are more important criteria than typist preference or productivity, they could basically coast for the past 15 years, without much substantive innovation (I guess they did introduce Cherry MX red switches ~6 years ago, but those are just a mashup of their existing switches, using the linear slider with a lighter spring). I wonder if Cherry still has any employees working directly on testing new types of keyboard switches or changes to their existing switches. I somehow doubt it’s very many people; their main recent innovation seems to be producing the same switches with transparent-plastic housings.
Unfortunately, there’s not much research going into keyswitches these days, as far as I can tell.. unlike in the 1970s/80s when there were dozens of companies working on it. New mainstream computers seem to have settled down to either rubber domes or scissor switches.
