I wonder if direct rotation drive is feasible? You can get a lot of power out of a motor, and as much throw as you want! Also, the force is constant, unlike with a solenoid, which varies as the bolt is surrounded by less coil.
We need angular acceleration for this. Moments of inertia. My A-levels feel very long ago…
Let’s model the drumstick as a simple rod with the axis at one end. 0.5m long, 0.05kg in mass
So:
People who know about this sort of thing: There’s no way I can generate that much torque from a standing start in just a few milliseconds is there?
I suppose you could have a continuously rotating drive with a flywheel, and some sort of trigger mechanism that engages the drumstick when a hit is required. Potentially, you’d only need one motor for the whole machine, if you lined up the instruments along a common shaft.
Further thoughts:
The “triggers” could be part of the rotating drive, like the pins on a music box cylinder. But instead of putting a pin where you want a note, you just engaged the instrument. That would guarantee perfect synchronised timing, for free, assuming you can maintain a constant rpm (or you can feedback a reading of the rpm and use it to drive the bpm of the sequencer
Isn’t this one of those magic triangles where you cant have everything (Pick two out of three etc).
if its over spec you pick the time, speed or mass and half one. so as the mass is fixed, you need to either go slower or accelerate over a longer period.
Oh for sure. I just thought we should be able to do better than the human body!
Speed has a required maximum. And given that we can’t (realistically) have more than about half a circle’s worth of run-up, acceleration has a lower limit too.
However, Mass and Length aren’t neccesarily fixed. If I used drumsticks half as long, the torque is 25%, the stroke length is 50% at the cost of double the rpm.
You can send the flyback current back into the coil the “wrong” way?
Instead of me dithering about here, could you or courty recommend a solenoid that’s got enough oomph I should buy to experiment with?
Yeah that’s the sort of thing. Although I think I’d rather have it sprung in reverse. If the hitting stroke movement is synched to the drive, which has a constant RPM, then I can guarantee a hit on time, and the rebound will help the reset the stick. Relying on spring + gravity to always exert the same force is going to end up with timing errors.
The dangers of a mechanical trigger system are that it could make as much noise as the drums themselves. It’s also subject to wear, which leads to sloppy/innacurate timing or explosive failures.
Plenty of time for the former. Little time for the latter!
Instead of just dumping the flyback to V+ you clamp it to the maximum that the driver can withstand.
In some dot matrix printer’s the coils were alowed to flyback to 120V
As for picking a solind I would take a scuk it and see approach and buy a selection and have a play
Remember that you can do a lot with the linkag as well
I want to minimise linkage complexity. Fewer hinges = less slop and simpler to build.
With this sort of arrangement, each individual stick would only get activated once per beat, with one stick deicated to each 8th note. Pull the pivot point of the sticks towards the cam wheel to engage it.
I’ve trawled through youtube and there are a bunch of solenoid drumkits. None of them make anything close to a loud noise. This is the closest thing I’ve found to what I have in mind in terms of volume and timing control:
Not a solenoid in sight. Beefy steppers instead by the looks of things?
I’ll jump for joy and buy you a beer when it can do the Amen break. Probably not simultaneously, that wouldn’t work.
For what it’s worth, robot drumming was a thing a few PhD friends did as part of their research. If you wanted to talk through your ideas with someone who has walked that road (with different criteria, of course), they’d likely relish the chance to unload over a coffee.
Told a musician friend about the project they pointed me at this https://vimeo.com/blancali/robotmaywadenki. Not the same brief but some interesting forms and mechanisms
Another really successful iteration of this idea was made (or at least commissioned) by Pat Metheney. It was focused on responding to the nuances of his playing and so more responsive than most. Interestingly he references the dynamic limitations of using solenoids in the FAQ.
FYI the DADAMachines automat box converts MIDI to control voltages so that may solve one part of the problem while you work on the mechanics. They’re just about to ship their Kickstarter kits and I think it’s going on general sale after that - not sure what the schedule or price is.
I’d be really interested to see how you get on with this as it’s my kind of thing. I’m looking to do something related with really big instruments; e.g. I have a 1.2m diameter circular saw blade gong that I’ll be turning into an interactive installation some time soon. I’ve made a smaller scale project called Curio that has a capacitative touch interface - you can see snippets of it here about 1 minute in - I still need to make a proper video It uses 42 instruments though so I can only use the cheapest solenoids and simplest mechanisms so that I can actually get it finished in reasonable time and at reasonable cost.
I also run a group called Hackoustic - we’re always looking for new projects to show at our events so you’ve definitely got a gig if you want one once it’s made
Do you not think steppers might work? Better control of speed and therefore volume. If the duty cycle is too low you can double up on motors and beaters.
It uses 42 instruments though so I can only use the cheapest solenoids and simplest mechanisms so that I can actually get it finished in reasonable time and at reasonable cost.