[sdiy] Emulating keyboard action with magnets; new synthesis method

Sun Jan 24 20:27:44 CET 2016

Is that a "screw type" linear servo actuator?

The basic idea is to be able to program the action such that when the
key is traveling (the distance of travel x is between 0 for nothing
and 1 for max), the force needed to push the key down should be
programmable (so there could be a table such as d(x) listing the force
needed to press down at the key travel x) and similarly the force
needed to press up (we could call that u(x))

the values of u and d should be between 0 and 1, where 1 means you
can't press the key down unless you press fairly hard.

A more complex system would also take into consideration the velocity
and acceleration of the key, x' and x'', so you'd have u(x, x', x'')
and d(x, x', x''). Note that with a weighted key, you can push on a
key and take your finger away before the key ends travel, which is the
whole point behind a weighted keyboard that lets players play really
quick passages.

Does anyone know what the force is a player exerts on the piano key to
press it down on a steinway grand middle C? (Steinways are fairly
standard and they are also fairly "heavy")

I see there are a few options, with their positive and negative aspects.

a. metal pole in a voice coil
+ may be the simplest to control and manufacture
- might be unable to perform fine control
- might be unable to create enough force to make it useful
b. "comb" plate like on a stepper sandwitched between two other combs.
Whether the two other combs use the same spacing, or deliberately use
spacing that'll have a "moire interference pattern" to the one in the
middle, is a good question, maybe in the interference pattern design
the single "teeth" of the comb could be turned on and off separately
to facilitate fine control
+ the combs may be simple to manufacture as PCBs
- might be unable to create enough force
c. gear at the fulcrum of the key, connected to a motor
+ simple to buy, simple to manufacture
- motor might be much too big or motor with the right dimensions might
be difficult to get
- might have to add gearing to create enough torque and fine control
from a motor of appropriate dimensions
d. screw type linear actuator
+ might be able to provide a lot of force
- might be too slow to react - the piano keys go really fast and it
doesn't seem like the linear actuator could really keep up
e. hydraulic piston with voice coil valve
+ might be able to completely stop the key from moving at all
- messy (hydraulics)
- might be too slow
f. brake calipers. Like on a car, where the brakes hold onto some sort of rail.
+ can stop the key completely
+ not only on/off, depending on the pressure of the calipers it could
present different force. The jaw could be out of rubber so that as you
press harder it presents more surface.
- cannot actuate the key
- slow
- u(x) and d(x) are the same
g. force feedback vibration mechanism (FFB)
+ very fast (the fastest of all)
- cannot stop the key at all. it can only be felt. (so no DC control, only AC)

regrading the magnetics (1-3), wouldn't they interfere with each
other? How much shielding would one need (how thick would the steel
have to be)?

It seems to me that each of those methods has different bandwidths
that it can serve. I would say this is the order from slowest to
fastest:

d. screw type linear actuator
f. brake calipers
e. hydraulic piston with voice coil valve
c. gear at the fulcrum of the key, connected to a motor
b. sandwitched comb plates
a. metal pole in a voice coil
g. force feedback

it would seem to me that the way to go might be to use multi-way
actuation. For example, a weighted key connected via an elastic medium
(say neoprene) to a very lightweight cover. that lightweight cover has
a strong force feedback mechanism in it. The weighted key is connected
using another elastic medium (say a springed harness) to a screw type
linear actuator, or somehow to a fulcrum gear, and all of that
reinforced with a brake caliper combo.

There are those two questions that come to mind:
1. what is the analog bandwidth of each of those systems?
2. what is the bandwidth of the sense of touch of a person's fingers? That is,
2a. at what frequency do we just start sensing "ok it's vibrating, but
that's all"?
2b. At which point do we stop sensing the vibration completely?

I assume the answer to question 2a is a number counted in m*Hz, that
is, if the frequency is high enough to not feel it any more, and then
you increase the amplitude (measured in meters), you will feel it
again. Since this is linear oscillatory motion where the velocity is a
sine wave, the unit m*Hz shouldn't be denoted as m/s.

As an aside, I wonder if there's a good system for making a weighted
key using flywheels.

On Fri, Jan 22, 2016 at 6:41 PM, nvawter <nvawter at media.mit.edu> wrote:
>
> Yes, I remember reading a paper on this a long time ago, perhaps even around
> 1996.
> I believe it was from Stanford.  They had made a small bank of keys, each
> one with servo actuator and
> feedback mechanism.  They were able to emulate several different pianos and
> other systems.
> I tried a quick search just now, but I'm at work and can't go in-depth.  I
> did find this, which has
> some inspiring pictures:
> http://web.stanford.edu/class/me327/lectures/lecture04-sensors_actuators.pdf
>
> There's also other work people have been doing in haptics in the
> high-frequency range like the comb
> teeth you mentioned.  My friend is working on one.  You can see the group
> project he's started here:
> https://hackaday.io/project/8446-weft-electrovibration-demo-board
>
> It's got some links to some papers at Disney, such as:
> http://www.disneyresearch.com/wp-content/uploads/REVEL.pdf    and
> http://www.disneyresearch.com/wp-content/uploads/teslatouchuist20101.pdf
>
> Anyway, just because it's been done once or twice doesn't mean it's over at
> all!  This is a humungous
> space to explore.  I know I've been spending months trying different methods
> of controlling synthesis
> based on FSRs.  The leap from piezo sensor to FSR is humungous.  There are
> so many possibilities and modes
> of expression, it drives me crazy.
>
> It would be nice if there were some kind of test platform for this stuff.
> What would it look like?
>
> Maybe a high-quality servo with a piano key on it and an FSR on the end of
> it...  then connected to some
> kind of block-diagram configurable interface like an Axoloti or analog
> modular?
>
> -Noah
>
>
> On 2016-01-21 18:09, cheater00 . wrote:
>>
>> Reading up on the new Nord Piano 3, I was wondering if anyone ever
>> tried to emulate the feel of a piano keyboard - how the resistance of
>> a key changes depending on the state of the hammer mechanism, by using
>> electromagnets or hydraulics.
>>
>> Either have electromagnets directly control the height of the key (in
>> opposition to the force exerted by your fingers) or use hydraulic
>> pistons feeding a reservoir, and have a small transducer constrict the
>> flow into and out of the piston (at which point you might even control
>> in and out separately).
>>
>> In addition to piano keyboards one might emulate all sorts of
>> different feels. Say, how it would feel if you were scraping a wooden
>> stick across a comb, or the feel of silly putty. This could feed a
>> synthesizer model - each physical interaction model could feed a
>> different synthesizer model.
>>
>> Also you wouldn't necessarily have to have the keys return
>> automatically. At which point you might want to think about making the
>> keys extend beyond the fulcrum - so that they can be pressed on one
>> end to press them inwards, and pressed on the other end in order to
>> depress them. For example, for ambient sounds, one might have a sort
>> of keyboard action that only very slowly returns the key - and the
>> height corresponds to the loudness of the note being played - while
>> letting you quickly stop a sound by pressing on the other end of the
>> keyboard.
>> _______________________________________________
>> Synth-diy mailing list
>> Synth-diy at dropmix.xs4all.nl
>> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy