Michael Ang

Hydrobot Mark I is a robot that makes music using water. Mark I sprays water pressurized by a pump onto metal resonator plates that are attached to a contact microphone. Laser-cut acrylic disks mounted to waterproofed motors spin and interrupt the flow of water to create rhythm. The sound of the contact microphone is amplified and broadcast to the audience.

Hydrobot Mark I made his debut performance at the ITP TOOL show at Tonic Saturday night. Everything worked great though in the final sound check I turned the water pressure down just a tad too much so the effect was a little more subtle than I had wanted. Mark II will have greater (and dynamically controllable) dynamics!

Video exists, but hasn’t been captured yet, so here are the pictures from final construction at ITP and setup at Tonic. The show was great, with many fantastic performances and installations.

Pictures from Hydrobot Mark I debut peformance at Tonic NYC

Hyrobot Mark I makes his big debut! Saturday April 29 at Tonic! Hydrobot Mark I will perform “Hydrobot Learns to Surf” with visuals by Ilteris Kaplan.

TOOL: An evening of novel performance technologies

Performances from the graduate Interactive Telecommunications Program (ITP) at NYU grace the upstairs stage at Tonic with interactive installations down in subTonic. Projects from classes such as “New Interfaces for Musical Expression” and “Live Image Processing for Performance” put the “live” back into electronic music performance. Presented by ITP instructors Jamie Allen, Gideon D’Arcangelo and Luke Dubois. Installations in subTonic from 7pm. Stage show upstairs begins at 8pm.

Saturday, April 29

TONIC
107 Norfolk Street
New York, NY 10002
$12, free with NYU ID
212-358-7501
http://www.tonicnyc.com
Tonic is located at 107 Norfolk Street between Delancey and Rivington Streets in Manhattan’s Lower East Side.

Performers include:
Jeff Gray
Todd Makinen
Mike Ang
Nick Sears
Langdon Crawford
Ryan Holsopple
Josh Knowles
Michael Harari
Britta Riley
Chris Kucinski

For further information on the show, contact jamie at nyu dot edu or gideon at nyu dot edu.
For more information on the Interactive Telecommunications Program at NYU: http://itp.nyu.edu.

Working in the lab. Here’s a quick and dirty drill guide for the Airpax LB82773-M1 stepper motor.

Airpax LB82773-M1 drill guide (.ai)
Airpax LB82773-M1 drill guide (.png)

Ah, April, when a young ITPer’s thoughts turn to… oh! crap! so! much! to! do!

Pictures from Hydrobot interrupter disk testing

Last night I tested the Hydrobot interrupter disks attached to stepper motors. The KP4M4 stepper was way to weak, but the ones I got from Jameco should work well. They can go pretty fast, but I will have to start them slow to avoid missing steps. The “spider” shaft couplers (rubber cross between two metal pieces) I got from Jameco are nice but using them means I need some way to trap the interrupter disk shaft axially since the coupler doesn’t provide any axial attachment.

The current plan to control the flow of water from the nozzles is to have spinning disks which interrupt the flow. Last week I cut some 5″ disks with 1/4″ mounting holes from acrylic using the laser cutter at NYU. I bought some Airpax LB82773-M1 5V 0.8A bipolar stepper motors from Jameco and some SN754410NE dual h-bridge drivers from Sparkfun. I got 7.5 degree step motors which should give a chunky step pattern.

Here are the test patterns. The final disks may be in acrylic or aluminum. I will probably bend the vanes to direct the flow of water down. I still need to design the housing for the motors. Either custom built acrylic or adapted from an existing water-tight box.

Illustrator source file

The Hydrobot performance will take approximately five minutes and represents a single cycle of the ocean from high tide to low tide and back again. It should give the feel of a good surfing session with high tide being dull, building waves as the tide ebbs eventually leading to poor conditions at low tide, then better conditions again on the way to the end of the day. Each of the four jets of the hydrobot can be thought of as different points (underwater rocks with the potential to create breaking waves) on a beach which faces an incoming ocean swell. Each point is assigned its own height value, which determines if incoming swell will pass over the point, break on it, or be closed out (break in a turbulent way).

If an incoming wave is more than one wave height above the point, it passes without breaking. If a wave is less than one wave height above the point, it will “break”, and the jet for that point is open for a period of time proportional to the height of the wave over the point. If the height of the wave when it breaks is too high, the wave is closed out and the jet will pulse rapidly and irregularly.

As the piece begins at high tide, the points are submerged and the waves do not break. The water from the jets is in this case constant. As the tide ebbs, the points will begin to break. As the tide lowers further, the points will become closed out, and the piece will be cacophonous with rapid bursts and irregular rhythms. As the tide rises, the piece returns to rhythm and then to a steady stream.

I made two short videos of the HydroBot for my NIME midterm presentation. (Need the latest QuickTime for the videos… sorry.)

Walkthrough video (high quality .mov 28MB)
Demo with sound from contact mic (high quality .mov 9.5MB)
Walkthrough and demo together (medium quality 14MB .mov)

Here is a sketch for what one of the nozzle arrays will look like. Right now I’m thinking two sets of these for 8 nozzles total.

And here’s a pic of a commercial valve manifold:

Some progress on the Hydro/SquidBot. I’m trying out different mounting techniques for getting the water spray to hit the bar perpendicularly. The bar wants to hang vertically, so that means the nozzle has to be parallel to the ground. The bar gets pushed around from side to side when the water is flowing, so I may need a more rigid mount. It seems like putting the bar on a pivot would work, or possibly I can make an arrangement more like a marimba/xylophone, where the bar is sitting on top of a support.

Since the water jet hits the bar perpendicularly there is a lot of spray. I kind of like the effect, but I need to figure out some way to contain the spray if this will ever be an indoor instrument! Or maybe SquidBot needs some pontoons and an outdoor pond!

I had considered trying to modulate the water stream using the solenoid valves to achieve different tones, but it looks like it will be more of an on/off kind of thing. Once I hack the valves for microcontroller control I can experiment, but they seem pretty slow.

The sound from the bar is still quiet relative to the pump, so I’m going to try using a contact mic to amplify it. Also I’m going to try connecting the bar to a low impedance metal resonator (sheet metal) to make a physical amplifier. The tone is also much higher than the fundamental of the bar. I may want to include some heavier bars (probably steel).

I’m trying to figure out how to move the nozzles to vary the tone. Ideally, the nozzles will still be perpendicular to the bar while they move. I considered hydraulics/pneumatics but possibly I can get away with servo motors and pushrods.

SquidBot testing pictures

UPDATE: More testing pictures

Some progress on the Hydrobot front. My thinking now is that I want to move the water more, and have the movement be more visible. The goal is to make the bot be itself a performer, and that requires the audience to understand what the bot is doing and have a feeling of tension. The audience should be amazed when the bot pulls off the performance.

The new idea is a bot in the form of an octopus or squid. The sound would be generated by spraying high pressure water onto tuned metal pipes. I’m not sure if this will actually work — experimentation is necessary.

Here is the concept drawing (click to view images fullsize):

Closeup of one of the pipe resonators. A solenoid valve controls a spray of water onto the resonator. A return hose brings water back to the pump.

Squid configuration. Resonator are arranged around a central pipe and empty into a single catchbasin/return system. The water dripping/spraying down from the tubes will be visible to the audience.

The main question at this point is what it will sound like. Time to rent a pressure washer and scrounge some pipe!

HydroBot: Proposal For A Water-Based Musical Robot

Primary Goals:

• Create mechanical instrument that incorporates water into its mechanism
• “Essence” of water should be clearly sonified
• Mapping should be as physical as possible
• Instrument should be self-playing (robotic, though under control/programming of musician using MIDI)
• Create a tonal instrument (if percussive, then with tonal qualities)
• Create an instrument that is musically interesting

Secondary Goals:

• Generate “watery” sounds or modulate pure tones such that effect of water is clear
• Create an instrument that can be used in ensemble with other musical robots or performers
• Mapping should be visually obvious
• Technology reuse for WaterLight

Constraints:

• Must not leak
• Transportation issues (e.g. tank length)
• Should be relatively easy to fill/drain

Design Process:

• Research existing instruments (e.g. water in wineglass, water drums, WaterPhone (demo), Mocean, others?)
• Consult with others (e.g. Eric Singer/LEMUR, NIME class)
• On paper design
• Practical research
• Test/prototype/refine

Design Ideas:

• Is it possible to use water to transmit audible tones directly to air (no interleaving material) loud enough to avoid amplification?
  • Experiment with metal plates and speakers excited under surface of water
  • Whale songs are created using air passed through structures in the body
• Suggests to use water to modulate another another body already vibrating (metal in contact with air)

Preliminary Sketches:

• Wave tank
• Underwater spinner
• Water Lily
• Overflow tank