Quantitative analysis of key-weight perception

[daniel0731ex]

Ever considered the possibility that perhaps it is the area under the force diagram that has a greater influence on the perception of the "weight" of a keyswitch than the actual peak force?

[Shadovved]

Technically, the area under the graph is Nm, which corresponds to the change in momentum of the switch (torque is not considered as there is no angular movements)

In changing the momentum of an object, inertia is most frequently encountered 

so in perception, may be true, but scientifically incorrect 

[daniel0731ex]

Technically, the area under the graph is Nm, which corresponds to the change in momentum of the switch (torque is not considered as there is no angular movements)

In changing the momentum of an object, inertia is most frequently encountered 

so in perception, may be true, but scientifically incorrect 

Not incorrect if I put "weight" in quotation marks. What I really meant is the fatigue felt by the user for prolonged typing.

In fact, the area under the curve is the work done to depress the switch, which relates directly to fatigue experienced by the user to depress the switch.

The reason for my use of the word "weight" is to point out that the community have long been focusing on the wrong mechanical aspect of keyswitches; the force-travel diagram assumes equilibrium motion, but the fact is that it is not at all a realistic representation of how keyboards are used. The effort expended by the user to type on a certain keyswitch type should be analyzed through collision and energy analysis rather than simple high-school-level Hooke's law.


I have posted this concept on this forum a long time ago, but was dismissed by one of the forum's self-proclaimed keyboard scientists as "just another McRip effect" (or was it called something else? I don't really follow all these geekhack definitions). But after all this time of reflection, I have come back being even more certain that I am right.

[Shadovved]

True, work done as well.

I judged the units of the eventual area, LOL, so momentum might be wrong........ or maybe not, since the change in momentum correlates to work done 

[daniel0731ex]

True, work done as well.

I judged the units of the eventual area, LOL, so momentum might be wrong........ or maybe not, since the change in momentum correlates to work done 

Joules = Newton x meters.

BTW, do you know whatever happened to Ripster?

[Shadovved]

True, work done as well.

I judged the units of the eventual area, LOL, so momentum might be wrong........ or maybe not, since the change in momentum correlates to work done 

Joules = Newton x meters.

BTW, do you know whatever happened to Ripster?

I heard he got banned. Oops, momentum is Ns. Argh, damn, my professor will be extremely disappointed...

[daniel0731ex]

True, work done as well.

I judged the units of the eventual area, LOL, so momentum might be wrong........ or maybe not, since the change in momentum correlates to work done 

Joules = Newton x meters.

BTW, do you know whatever happened to Ripster?

I heard he got banned. Oops, momentum is Ns. Argh, damn, my professor will be extremely disappointed...

I see. Has keyboard science improved or deteriorated since then?

[Shadovved]

Dont really know, its only been awhile since I joined GH........

[limmy]

Ripster is banned for good. Not only on GH but also on DT. You can find his keyboard science at reddit/r/mechanicalkeyboards. I find some of them useful, but some of them misleading.

About the topic, I think there is more than integration of force curve when tactile feed back is in play, though. Oh.. and I think best way to find out about switches is to try them all. Research is most useful when you cannot try them yourself. Now mechanical keyboards has become pretty cheap, so I say trying it out yourself has become less risky than ever. Also, there is pretty big second-hand market where you can unload unwanted keyboards.

[SurgeonKyle]

In relation to the above post, yes, he's kicking around on Reddit:

[Soarer]

Good luck finding anything useful there.

On topic... I've thought about it before, but since most switches have similar activation points and starting forces, the peak force before activation will correspond very closely to the area under the graph, at least in terms of ranking. If there's a notable difference in activation point or whatever, that's usually worth noting separately.

[fohat.digs]

All good points and solid science, but I tend to thing that there is a fuzzier component that could be much harder to quantify.

How do "YOU" actually type?

Personally, I suspect that my fingers hover a bit higher over the board than many other users, and therefore I have about a centimeter of "dive" with my fingertip before I contact the key. This added distance/speed/momentum means that I can tolerate a higher activation force.

In my martial arts training, we do a "snap kick" which means kicking to and through the target, whatever, but then stopping the foot, drawing it back, and "re-setting" it as we step down. You can step forward, back, or to the same point, but you stay in balance and have control. Beginners kick as hard as they can, then basically fall forward, off balance.

My guess is that I am using a similar stabbing "snap press" technique on the keys to strike down onto them, then immediately retract, similar to the kick. I like a precise tactile point, high up, and am less sensitive to force, although in general I like light springs better than heavier.

Other typists, like my wife (a magazine editor who types thousands of words a day and thinks that the best keyboard in the world is the thinnest newest Apple board), probably float just over the tops of the keys and make the slightest possible downward "dip" press to activate the key. So flat keytops and minimal travel are what she enjoys most.

My technique is probably not ideal, but I learned to type in 1969 when my father bought me a manual typewriter when I was a senior in high school. Activation force might have been measured in kilograms rather than grams on that thing - it had to be "attacked" with real force and momentum!

[TotalChaos]

The effort expended by the user to type on a certain keyswitch type should be analyzed through collision and energy analysis

I have been saying this for 9 years but nobody listens to me.

Sure the force required to actuate a switch is important and I love force graphs.  I wish I could find force graphs for all the switches.

But nobody ever talks about the force of the crash, or even if there is a crash.

When I became horrifically disabled with hand pain on April 27, 2003,  I tried searching for "zero impact keyboards" and found nothing.  I tried searching for "springloaded keyboards" and only found buckling spring keyboards which I definitely did not want, even though they are better than rubberdome over membrane.

How does one measure the force at impact for a typical key activation?

Whey ppl type they do not press down with the exact amount of weight required to reach a point.  They always press with a little more and they have acceleration as they go down.

Then there is the hardness of the thing they are banging into to consider.

My fingers are burning right now because of the simple fact that cherry switches only allow me 2mm to stop after I reach actuation.  How am I realistically expected to do that?  Yeah ok I am practicing to be better and all that but Cherry isn't exactly making it easy for me.  The difference in force between actuation and CRASHING into a steel plate is only 15g over 2mm.

If Cherry or someone would just make a switch with 8mm of travel, actuation at 2mm, using a longer version of a Cherry red spring then all my probs would be gone and I could type up a storm today.

[fohat.digs]

I would like that, too!

Having a deeper "well" would make it much easier to stop before you hit bottom. Every single mm would help.

I still think that O-rings are great, regardless of how they effect travel. It is a bit disconcerting to bounce into one that is very close to the activation point, but the cushioning is worth it, to me.

[alaricljs]

Are you asking for a measurement of the spring force at full depression or the force of your finger smacking into the  bottom?  The force charts tell you the spring force throughout the stroke, but nothing can tell you how hard you're hitting the board unless you go get yourself measured.  That's an individual thing based on how hard you're attacking and how long it takes you to back off.

[TotalChaos]

I would like that, too!

Having a deeper "well" would make it much easier to stop before you hit bottom. Every single mm would help.

With a deeper well you get more than just time to stop.
With a deeper well you get more than just distance to stop.

With a deeper well you get more and more and more resistance from the compression of the spring.   So that even if you crash, you are moving very slowly and the impact is not so much.

Springs are awesome natural shock absorbers!

But they need to be long enough to do their job.

The reason we all resort to buying O-rings and crashpads is because Cherry switch springs are toooo short.  The force graph goes up gradually for 4mm then suddenly jolts straight up into the sky.

[rootwyrm]

I would say the perception of key weight isn't the peak force, or the past-peak point. It's actually the ramp to peak force. Think about it a moment. Peak force is the activating point more often than not. Now if your ramp looks like this:

/^\
That's going to feel a lot lighter than a ramp that looks like this:
--\

(Hooray bad ASCII art, but you get the idea.) The ramp going to peak is where you feel it most - a flatter ramp provides less feedback indicating you're going toward peak. Which is a problem for most folks, because they're used to a bell curve ramp. (As an example of a flat ramp, I'd say look at flat membrane. Not the roll up stuff. The stuff like on ATMs and industrial equipment.) So my take is that you shouldn't be looking at peak force except as the stopping point, you should be looking at the ramp to peak.

Now personally, I'm one of those people who bottoms out buckling springs. This is despite the fact that I'm also a pianist who has no problem with mezzo pianissimo on a synthesizer. It's a matter of feedback. The tone doesn't sound right for activation till I actually bottom it out. And I write thousands upon thousands of words a day, easily. Why do I bottom them out other than the feedback? Because buckling spring essentially has the same bellcurve pattern as Cherry MX and so on. I actually think it could be argued the bell curve is actually a major design flaw in keyboards and contributing to RSI problems.
What's needed is a switch where the force increases past peak activation, if you want to stop bottoming out. (Which O-rings do, except with a stop that I'd argue is still too abrupt. And yes, I agree, the movement length of Cherry is too short.) The problem is: getting a spring that does that seems to be exceptionally difficult or expensive, neither of which is conducive to mass production.

[fohat.digs]

I don't think it would be hard. As Total Chaos pointed out, simply making each part a couple of mm longer would be a huge help, with the added length going to making the top of the key sit higher relative to the backboard.

But the manufacturers won't do it because they would have to make new equipment such as molds and dies.

[rootwyrm]

I don't think it would be hard. As Total Chaos pointed out, simply making each part a couple of mm longer would be a huge help, with the added length going to making the top of the key sit higher relative to the backboard.

But the manufacturers won't do it because they would have to make new equipment such as molds and dies.

Sure, increasing travel is actually fairly easy. You're correct in that it would require new tooling, which is why they're likely resistant. It also would require redesigning the keyboards themselves, and they would have to be thicker - which is contrary to the popular push to make things thinner. (If you increase travel from 5mm to 8mm, that means the upper section needs to be +3mm.) It also makes things more expensive by simple virtue that you require more raw materials, because you've added 3mm of height.

Probably should have been clearer - what I meant was difficult is a spring with a flatter curve that increases resistance past peak point.

[rknize]

Why not modify the Cherry stem to move the activation point near the top?  This won't give you more travel, but the travel you have is weight towards what you are trying to accomplish.

[TotalChaos]

And yes, I agree, the movement length of Cherry is too short.) The problem is: getting a spring that does that seems to be exceptionally difficult or expensive, neither of which is conducive to mass production.

Why do u say such a thing?

Did you know that they had springs back in the 1800s?
Its an ancient technology.

To make a longer spring they just need to
A: Heat up steel to the appropriate point
B: Run it thru their little springmaker machine twice as long
C: Profit

They haven't done it out of sheer laziness.

I wanted a cherry red keyboard in 2003.  I desparately needed one.  I could not find one.  As near as I can tell they only invented cherry reds a couple of years ago.  What on earth were they doing all those decades?

Maybe we should think about doing a group buy to bribe that chinese company that makes cherryclone switches to make some double length switches.

I have already practiced at my keyboard and I could easily handle an extra 12mm of height.  So an extra 4mm is nothing.

I donno what the perfect length for the spring is.  All I know is an extra 4mm would be wayyyyyy better.  I would have loved an 8mm or 12mm of travel cherry black way back in the 1980s.  But they didn't make them.

[TotalChaos]

Sure, increasing travel is actually fairly easy. You're correct in that it would require new tooling, which is why they're likely resistant.

Making new tooling didn't stop cherry from making 10,000 different weird switches and weird keyboards.  They love to make new tooling for specialty items.  And a full travel switch is not a specialty item.  Its something everyone can appreciate.

It also would require redesigning the keyboards themselves, and they would have to be thicker - which is contrary to the popular push to make things thinner. (If you increase travel from 5mm to 8mm, that means the upper section needs to be +3mm.) It also makes things more expensive by simple virtue that you require more raw materials, because you've added 3mm of height.

Take a look at a Das Keyboard or a Topre Keyboard.  Do you see all that plastic they are wasting?  They are wasting more plastic (and deskspace) than the extra 4mm of plastic would take to raise the height of a keyboard.  And going up does not waste deskspace.

Ok the switch itself would use more plastic, true.  And the spring would use more steel.  True.  But these are not insurmountable obstacles.