deskthority

Since virtually all keyboard switches have springs of some kind to provide tactility, click, and return force, I was thinking about a different type of action that may prove smoother than spring-oriented switches, as I am typing on my beam spring, right this moment.

My idea surrounds using a button, fitted with a natural magnet that levitates above an electromagnet that is, in its unpressed state, so driven as to repel the button and make it afloat. When pressed, the button triggers some kind of mechanism that flips the polarity of the electromagnet below, which makes the electromagnet attract the button towards it; this makes the button literally fall down after the press is registered by some means, perhaps the same or another pair of magnets with hall effect. After the button drops to a certain level, the polarity is again flipped, which repulses the button upwards to its initial state. The flipping of the polarity would provide tactile feedback, and a click could be added by giving the button some sort of a tail that would smack into the electromagnet.

The idea is that the button would have minimal friction, if it is only held in place by a very loose coil spring; in fact, the only source of friction would be air rushing pass it, which may be further decreased by aerodynamics.

Let me know what you guys think of this crazy idea!

mmmh possibly related

http://www.google.com/patents/US3713056

Looks like someone has this crazy idea fifty years ago.

You forgot the bit about capsense so you can track where the sliders are…

Why capsense? Seems overly elaborate to make a magnet a capacitive element Oo .

So you can track its position, analog and in real time, to vary the magnet's power and play some interesting games. Doesn't have to be capsense I suppose, but that's how I've imagined it.

I think I've heard of a single magnet capsense board before, but it seems excessive to me. Why not make them Hall effect, magnetic reed or magnetic valve instead?

Also, I suppose you'd need a really long magnet for stabilisation purposes.

You could use an LVDT for each key if you are going for overkill!

Muirium wrote: You forgot the bit about capsense so you can track where the sliders are…

I thought I said Hall effect.

Brilliant idea, as you have noticed it's been done before in more that one format. Now while a magnet solution is a great approach to achieve minimal friction and hence abrasion over a very long time there are other interesting switch solutions in regards to minimal friction like Fujitsu Leaf Spring:

wiki/Fujitsu_Leaf_Spring

keyboards-f2/fujitsu-limited-n860-2505- ... 13721.html

review-f45/fujitsu-sanyo-keyboard-revie ... 13970.html