GH DESIGN CHALLENGE: force vs displacement measurement device for keyboards/sws

[mashby]

[damorgue]

Note that as earlier mentioned, one has to measure the entire function while in motion to get an accurate depiction of the force graph. Not only that, but the speed can influence the looks of the graph so we'd need some sort of average speed at which a key is depressed or make several measurements at different rates.

Edit: I'd also be interested in measurements taken of angle to see how different technologies and sliders differ in that regard. This also happens to be another source of inaccuracy if things aren't aligned properly.

[mkawa]

yah, it's basically a shock dyno but tiny.

[OldDataHands]

Does the drive wheel really need a separate piston?
Can't you just drive the switch directly with the connecting rod?
I've never met a finger which didn't put a small side-load on a switch...

Just thinking 'bout the setup... Maybe you want to acquire data while cycling at 5Hz...
200ms for one rotation of the drive wheel. Perhaps we can acquire at 10k samples/sec.
That would mean 2k samples for one revolution. Leaving some head room,
Maybe a 1024 count rotational encoder is the right answer for position info.

Perhaps this: https://www.sparkfun.com/products/11102

You could just post-analyze the data to identify TDC, or have a single absolute
Mark which comes off a hole on the wheel and an optical switch...

You could just spool the serial datastream to file and make a graph after.
Seems like one of the small load cells would be the hot ticket for force measurement,
Though calibration could be interesting. Resolution might be a challenge... Has anyone picked their favorite force measurement device?

[mkawa]

time to put the project manager hat back on for a second (oh no!). dialectic incoming

what exactly is this thing supposed to do? i claim, _ideally_, that we have exactly two goals

a) simulate an arbitrary human typing on an arbitrary keyboard

b) imagine a incompressible, arbitrarily thin observer with arbitrarily small mass sitting on top of every key. we want to continuously report the force vector exerted on the observer by the key, its position relative to TDC of the key at rest, and probably the force vector exerted by the human finger on the observer as well (it is an arbitrary human, after all).

this begs a heck of a lot of questions, and undoubtedly needs to be simplified and further explicated to create a solvable problem. however, i suspect we will make more progress if we are clear on our premise before we start designing a hundred arbitrary devices.

[Gimmel]

i used an Electronic Force Gauge from Torbal for it, but i guess it's worth it price when you would use it more often. Anyway - your design is preaty neet.

[Soarer]

Does the drive wheel really need a separate piston?
Can't you just drive the switch directly with the connecting rod?
I've never met a finger which didn't put a small side-load on a switch...

Yeah, I guess if the rod were long enough, the angle would be insignificant. The weight of the rod would mean it was always 'hanging' from the crank, thus hiding any slop at that joint, at least at low speeds. (I don't think the speed is critical; it just has to be moving).

Maybe a 1024 count rotational encoder is the right answer for position info.

Seems like overkill! As long as the flywheel has decent mass the rotational speed will be pretty steady and only a few counts per rev would be needed.

[Soarer]

i used an Electronic Force Gauge from Torbal for it, but i guess it's worth it price when you would use it more often. Anyway - your design is preaty neet.

Trouble is: response time. Anything that processes the raw signal from the sensor will likely be averaging and therefore introducing lag. For this measurement we still want to average, but the values would be from the same displacement on successive cycles, rather than consecutive readings from the sensor.

So what I labelled as 'scales' in my diagram might indeed be a set of scales - since it's a cheap way to get a mounted sensor etc. - but modified to tap off a fairly raw signal from the sensor.

[HoffmanMyster]

This is really interesting, I'd like to contribute to this.  We'll see if I can produce anything useful. 

[mkawa]

ELECTRONIC FINGER

i had a thought the other day that what we need is a completely enclosed actuatable piston on a swingarm. the piston measures force and actuates a keyswitch. it sits on swingarm and you have to either clamp it to something or it screws into a large weight. the idea is that you're measuring the keyswitch in situ. anything else is a weird microbenchmark which doesn't really give you a good idea of what the switch _feels_ like.

[mashby]

ELECTRONIC FINGER

i had a thought the other day that what we need is a completely enclosed actuatable piston on a swingarm. the piston measures force and actuates a keyswitch. it sits on swingarm and you have to either clamp it to something or it screws into a large weight. the idea is that you're measuring the keyswitch in situ. anything else is a weird microbenchmark which doesn't really give you a good idea of what the switch _feels_ like.

I know just who to call to get one of these...