deskthority

I'm trying a rather ambitious project involving magnets, Cherry MX mounts, Hall effect sensors, and some 3D printing, and it's got me asking questions about keyswitch design.
1. What are the major sources of scratchiness? Generally, I'd think of it as interaction between the slider and the housing, but everything has a degree of interaction.
2. What are the major sources of key stability? By which I mean the resistance to wiggling the keycap around, and the ability to press a key off-axis and still smoothly actuate the switch. Again, I figure slider and housing interaction design, but what about the design causes it? Like, what makes a Honeywell Hall Effect or SKCC so damn smooth?
3. Does anyone have definitive numbers on the force/travel/actuation point of an IBM Selectric? I've heard 30g / 6mm / ???, but I'd like to ask you folks.
4. What would you think of a 6mm travel switch, either in the abstract, or if you've typed on one?
5. Do you think there would be any significant downsides / upsides to a switch that had an extremely light spring that served essentially as merely a return mechanism, with most of the resistance being from the tactile event itself?

1) Amount of friction and lubrication. More roughed up not lubricated area is going to be more scratchy than minimal contact and good lube.

2) WHERE the contact is over WHAT size area?

3) No idea.

4) I can not possibly imagine sorry.

5) unfortunatley the spring needs to be strong enough to return the slider if it has some sort of tactility. EG a 25g spring won't be able to push up something that needs 35g to get over the tactility.

atrere wrote: ...
5. Do you think there would be any significant downsides / upsides to a switch that had an extremely light spring that served essentially as merely a return mechanism, with most of the resistance being from the tactile event itself?

Had this for a little while when I made some ErgoClears with too light springs. I found it very unpleasant, the tactile bump becomes much too prominent and even with properlubing larger keys frequently didn't make it back past the tactile bump.

A lesson from this is that whatever switch you design you always must make sure the spring is strong enough to lift even heavier (large, stabilized SA profile keycaps ...) keycaps all the way back up. And that "return" movement hast to be lightning fast, basically immediately. It feels very odd when switches sometimes fight a brief moment overcoming a bump on their way back and you press it again before it fully returned. Greatly disturbes my typing experience.

So unless you find a switch with basically ZERO return force, you will have to add a spring with compression force. While I can technically imagine switches that have a unidirectional tactility, I can not come up with mechanics simple and reliable enough for a keyboard switch.

1. Surface with microscopic imperfections sliding against a spring under tension. I think the spring amplifies the imperfection. Dust and grit getting into the mechanism worsen things.
I think it would be hard to avoid imperfections in 3D-printing. Maybe you could try printing with the barrel and slider lying down along an axis of your printer.

2: The slider is in a rail. More stable switches have longer sliders/rail contact surface along the axis of movement. Another way is to have a built-in stabiliser - some Topre keyboards use stabilisers even for 1.75 wide keys.
Scissor switches are stabilisers without sliders and therefore often very stable. If you could construct a high-travel scissor mechanism then that would be awesome.

4,5: My favourite switches are highly tactile switches such as Cherry MX Clear, (and clones such as the smoother purple Zealio).
These are tactile, but with a large rise in force after the tactile bump and valley. The large rise in force is due to the spring being hard: it cushions the stroke, but I sometimes bottom out anyway. If the travel was longer but feel otherwise the same then I would bottom out much less and that would be great.

I also don't like it when switches have a too small initial force, such as Alps that start at 30g. I much prefer MX because they start at 40g.
The much loved IBM Beam spring switch starts at 45g and has a very low slope until actuation where force drops considerably but where the slope after the actuation point is very steep. Extremely light spring: Not for me.
In general, shorter spring gives sharper slope. Longer spring gives flatter slope. Spring should be under tension in "resting" to provide higher initial force.
6mm travel: Yes, please. But actuation and tactility should still be high up.

I have been thinking that maybe it would be possible to construct a very nice switch by combining a rubber dome and a spring -- where unlike Topre, the spring's contribution is significant.
With the rubber dome providing the majority of the force before actuation, that would allow a shorter spring that is not under (much) tension before the key is pressed. Maybe that could also provide high tactility and low force after actuation.

3: I found this force graph of the Selectric on my harddrive. Not sure what it comes from, but it does not look very smooth.

1. Friction between moving parts and non-moving parts. Can be lowered by lube and smoother contact surfaces.

2. Low amouts of wiggle room. No clue about smoothness of honeywell switches.

4. Could be nice, but should probably add as little additional height as possible.

5. Could get stuck below the tactile peak. I'm a fan of heavy switches, so that's rarely a problem for me. Findecanor's diagramms remind me of those ballpoint pens where a little pin glides down a tactile shape on the way down when you press the button, locks in place and takes a linear path back up. That mechanic without the locking could eliminate this problem, but probably isn't applicable to keyswitches.

1. Friction, key switch parts in contact with each other in movement. Fewer parts in contact will result in less fritction and hence less "scratchiness". Honeywell Hall Effect is a good example, by design that switch has no parts in friction since the contact is made by the magnet resulting quite smooth in feel. There are other examples like Fujitsu Leaf Spring that achieve this by design.

wiki/Honeywell_Hall_Effect

2. General switch design and percision in build quality. Keycap mount and general stability within the switch itself in relation to the design. How much movement does the stem allow when the keycap and how much "play" does the slider mechanism have in general? All this is of course dependent on percision and the build quiality and wear and the overall keyswitch design itself. Many keyswitch designs are supposed to have some "wobble" on the cap.

3. https://www.google.de/#q=ibm+selectric+specifications

4. I'd have to try it to tell you. And the 6mm tavel alone won't make for a good keyswitch.

5. None. Some people may like / need some tactile feedback, others may not. I own switches like that, you either like the feel or not.