Approximating a Slider-Crank Acceleration Profile w/ Vibration Motors

[adishavit]

Say I have a mechanical device like this one:

If I plot the acceleration profile at (trace-) point D I get something that looks something like this:

Now, here's the question:

Can I build a device composed of one of more (eccentric rotating mass) vibration motors that when rotating will approximate this acceleration profile?
Let us assume that they can be synchronized and initialized properly (as stepper motors can).

Ideally I'd like to have as few such motors as possible.
Maybe some DCT-like transform of the profile?

I'm not a-physicist nor a mechanical engineer, so be gentle.

 

[marellasunny]

I don't know the answer,but is it a Whitworth shaper? Asking because the acceleration profile looks so skewed.

 

[adishavit]

The mechanism is as drawn in the diagram. It is not the Whitworth shaper, but it was designed to has an asymmetric acceleration profile.

 

[Baluncore]

You might synthesise that acceleration pattern from the Fourier components.
If you have sinusoidal harmonic vibrators that are synchronised, then you will need several harmonics to approximate the acceleration spike.

 

[adishavit]

You might synthesise that acceleration pattern from the Fourier components.
If you have sinusoidal harmonic vibrators that are synchronised, then you will need several harmonics to approximate the acceleration spike.

Yes. That was my thinking too.
I mentioned DCT ([Discrete] Cosine transform) which has good signal de-correlation attributes. I'm glad this idea was not so far-fetched.
Would FFT be better than DCT (or DST = Sine Transform)?

What exactly are "sinusoidal harmonic vibrators"?

 

[Baluncore]

They are something like your "Can I build a device composed of one of more (eccentric rotating mass) vibration motors that when rotating will approximate this acceleration profile? Let us assume that they can be synchronized and initialized properly (as stepper motors can)."
Stepper motors can miss steps during acceleration so they would need some form of phase trimming to keep them in time.
I would use an FFT to compute the most significant phase and amplitude requirements for emulation.

 

[Baluncore]

Seems to me that the acceleration spikes are the key and the recovery waveform is probably non-critical. So why not use an electric jack hammer or a hammer drill to make the acceleration pulses.
Then, if you need to cultivate the rest of the signal, use a simple eccentric vibrator.
How accurate does the alternative implementation need to be?

 

[adishavit]

How accurate does the alternative implementation need to be?

Not very. Indeed, the important part is the asymmetric acceleration - higher in one direction with a slower return.

Seems to me that the acceleration spikes are the key and the recovery waveform is probably non-critical. So why not use an electric jack hammer or a hammer drill to make the acceleration pulses.
Then, if you need to cultivate the rest of the signal, use a simple eccentric vibrator.

I want a miniature device - think pager-motors.
Jack hammers are a tad too big.

 

[Baluncore]

Sorry, but I misinterpreted the scale by several orders of magnitude.
I really don't like the idea of multiple motors at that scale.

This seems like an application for a speaker voice coil or a piezo electric transducer.
Use a MOSFET to connect it to the supply momentarily, then discharge slowly through a resistor, wait and repeat.

 

[adishavit]

This seems like an application for a speaker voice coil or a piezo electric transducer.
Use a MOSFET to connect it to the supply momentarily, then discharge slowly through a resistor, wait and repeat.

Thanks. I'll consider that.