Idea for a highly, but smoothly tactile switch

[bhtooefr]

I'll call it the "ball switch".

There's a linear spring off to one side, to provide added resistance and the return force, but the tactility is generated by the ball action.

The slider acts on a bellcrank to rotate the motion of the slider 180 degrees, to push a ball (sitting on a plunger) straight up. This ball is in a chamber that's angled towards one side at the top, so that an upward force on the ball, at the top, pushes it to one side.

The ball will eventually be forced into a contact chamber that's angled down (and it will fall into that chamber, off the slider, causing a sudden, large, yet smooth (unlike buckling spring) drop in force). However, a tab was also pushed up along with the plunger, retaining the ball in the contact chamber.

Upon release of the key, the linear spring pushes up, and the plunger and retaining tab come down, and the ball falls onto the plunger. There's some hysteresis here.

[itlnstln]

Looks like an interesting idea. Careful spring placement would be key to ensure you don't have any off-axis force. In other words, having the spring to the side may push the slider in a non-vertical manner.

[bhtooefr]

Manufacturing could also be a real pain, but I THINK it could be done with 5 injection molds (compare to 4 molds for a Cherry blue). Assembly would be a lot more involved, though.

Also, because the ball isn't directly coupled to the mechanism, and gravity is a part of this switch's function, the board will need to be flat or results could get really unpredictable. No shaking out board chow with it plugged in!

[itlnstln]

No shaking out board chow with it plugged in!

I was hungry. Not anymore.

I would imagine that the switches might be fairly large, but I might have some mental bias from your pic.

[bhtooefr]

Yeah, I'm thinking about Cherry MX sized.

Also, I think I found some simplification that could be done, and may be necessary with the bellcrank - have the main slider go through the spring, and engage the bellcrank underneath the spring.

[indubit]

I might have misinterpreted your design but couldn't there be a minor issue if you want to tap the key really fast? Ergo if the ball haven't dropped all the way (no pun intended), you wouldn't get the same tactile feedback? Although, that might not be a big issue..

[itlnstln]

I might have misinterpreted your design but couldn't there be a minor issue if you want to tap the key really fast? Ergo if the ball haven't dropped all the way (no pun intended), you wouldn't get the same tactile feedback? Although, that might not be a big issue..

I believe that was the hysteresis he was referring to.

[Findecanor]

I am not sure that mechanism would be tactile at all. I would like to see a prototype / mockup of some sorts.

[bhtooefr]

It would be a "smooth" tactile. Force would go up as the ball was raised, and then when the ball dropped ( ) into the contact chamber, force would smoothly yet quickly go down. It'd be as if the ball fell out from under your finger, which... oh, it did.

[Icarium]

I highly doubt that any of this will work and demand an oversized lego prototype.

[indubit]

I might have misinterpreted your design but couldn't there be a minor issue if you want to tap the key really fast? Ergo if the ball haven't dropped all the way (no pun intended), you wouldn't get the same tactile feedback? Although, that might not be a big issue..

I believe that was the hysteresis he was referring to.

Oh sorry, my bad, should have read more carefully.

[bhtooefr]

So, another switch idea I had, albeit extremely fiddly to assemble, and has a freaking SOLENOID. However, it would behave just like a Selectric.

Switch is drawn huge due to the complexity, might be possible to miniaturize it, though.



So, here's what's going on here. The key is connected to a linear switch directly, and that switch sends the keystroke. This linear switch has a nub on its slider, which engages a second, much stiffer linear switch. That second linear switch has, on its slider, an element that can slide freely in the transverse or longitudinal direction, a spring holding it underneath the nub on the first switch, and a solenoid that either kicks or pulls the sliding element out from under the first linear switch, causing the switch to vertically restore, but with the sliding element caught on the side of the nub. (It will fully restore when the first switch restores.) Optionally, a second electrical switch could be present to detect this, to determine when the key is released (even after the first electrical switch restores).

Horrendously expensive, potentially very hard to assemble, potentially large, and with high power consumption, but could have fantastic feel, complete with the Selectric's hallmarks of nearly zero preload and very low post-actuation force, and the key "falling out from under your finger" on actuation.