A few of the guys got together Friday afternoon and built a hurdy-gurdy prototype. The stringed instrument is constructed from a cardboard box and plywood frame. So far the instrument sounds decent when turned, but needs some work on amplification. Using material other than cardboard should definitely help the acoustics...

Fellow RobOrchestra member Andrew Burks and I spent the better part of Friday night (yay social life) working in the shop on the bass drumming robot (as yet unnamed). Andrew made an awesome SolidWorks model which we used as a reference for the parts.

The basic design uses a Bimba air cylinder limited to ~0.5" stroke under 30psi. The cylinder has pivot points at both ends. The pivot points took a lot of machining as we had some pre-existing components that needed to be modified to allow attachment points. One screw up and we were out those parts and had no spares. Fortunately, everything went off without a hitch.

Once the cylinder has pivots at both ends, one end gets attached to the "shoe" of the pedal, the other end inside our robo-leg. We have some sweet plans for decorating said leg -- but until then...

here's a truly amazing picture of yours truly, oh how flattering :-)

The RobOrchestra Snarebot is finding its legs..err arms this year. The project is a few years old and has experienced incremental improvements over its lifetime. This time around we are looking to make a dramatic improvement based on the existing design and input from drumming consultants. We want to make it as realistic at drumming as possible with good dynamics and a pleasing sound.

The old design used two large solenoids pulling on cables to actuate the drum sticks. The wires were connected to specialized rubber holders that had built in compliance to allow the sticks to rebound quickly off the drum surface. These work really well and were one of the biggest components in the success of snarebot 1. What the 'bot lacked was good control over speed and dynamics - since solenoids are either "on" or "off", that's all you could coax out of it. Since there was a single pivot point, it was also impossible to induce different velocities out of the drum strike.

This is the design I envisioned last year - a small, portable device that could fit on any standard size drum. It was actually really neat, and I built a small prototype for the shell and legs. When you rotated the center handle the legs would extend or contract depending on the direction you turned. Since there wouldn't be much force opposing the legs (all downward) they would remain extended and the 'bot would perch happily on the edge of the drum.

This worked pretty well, but only left about an 8 diameter area in which to fit the drum stick mechanism. We found out that short sticks sound really flat since you couldn't get the pleasing resonance you normally hear from a snare drum. The device made had about 4" arm and could hit every 25ms, so speed wasn't an issue.

My sketch for this years' redesign is above. Its larger offset design allows for greater movement and the use of standard drum sticks, so we avoid the resonance issues of the past design. It will incorporate cable drives, solenoids, and servo motors to provide a wide range of speed and dynamics control lacking in Snarebot 1. The drawing isn't exactly to scale as the drum sticks actually stretch across the diameter of the drum, but after taking some measurements I verified the design is doable. Now it's just a matter of finding time to sit down and CAD model it in SolidWorks.

Some people have expressed concerns about latching onto the tuning bolts, but others say we should be fine. I've designed the 'bot this way so that we can eliminate the large clunky stand. It worked out pretty well for Snarebot 1, since you could easily reposition it, but we have something up our sleeves for this design that can accomplish the same thing.

You can't see it from this drawing, but the base will incorporate motorized translation and rotation so the snarebot can target different areas of the drum. Now, we're starting to talk about a lot of motors here, but each RobOrchestra controller board has ~16 outputs, so we should be fine. Hmmm....how about some neon underlighting?