Here’s a set of photographs of the interactive, kinetic sculpture “Blame”, as well as a writeup on the technology and techniques used in building it – by Thomas Edwards.
“Blame” is an interactive kinetic sculpture consisting of a robotic arm with a pointing hand on the end that scans the gallery, and when it finds a viewer, the arm stops with the hand pointing at the viewer, and then the sculpture proceeds to blame the viewer for some horrible crime against society. After that, the arm returns to scanning for a new victim to blame.
“Blame” on exhibition at Fraser Gallery, Bethesda, Maryland.
The front of the hand, showing the ultrasound proximity sensor hung directly below it.
The hand of “Blame” is a plaster casting of my own hand made using a hand casting kit from Casting Keepsakes. The hand is attached to a PVC pipe arm using a masonry screw through a PVC end cap, bolstered with epoxy gel. The end of the arm that is not attached to the hand has a counterweight attached for balance. It uses thin cylindrical weights from an adjustable ankle weight set from Sports Authority. The arm is driven using a surplus Saturn windshield wiper motor from BG Micro. The motor runs at 41 RPM on the slow setting, and draws about 1A at 12V. 41 RPM is too fast to have the arm move, so I used a pulse-width-modulated (PWM) signal into a motor controller from Hobby Engineering that uses an L298 dual H-bridge IC.
The “brains” of “Blame” is a PIC 16F648A flash-based, 8-bit microcontroller from Microchip Technology which orchestrates all the elements of the sculpture. It produces the PWM and direction signal to the motor controller, monitors sensors, and drives the audio system.
To detect the viewer, I used an ultrasound proximity sensor package hung from the arm just below and behind the hand. The sensor sends out an ultrasound pulse, then measures how long it takes for an echo to return in order to determine how far away a viewer is. It is fairly directional, and can only detect objects within a tight cone in front of the sensor. Ultrasound distance sensors also make a “ticking” noise, which in this case helps the sculpture seem more menacing.
The arm moves in a 180-degree arc. To detect when the arm has reached the end of the arc, a magnetic reed switch from Radio Shack (used in home security systems to detect the opening of windows) is used as a limit switch. A magnet is attached to the bottom of the arm, and the two reed switches are located at the end of the arc. The PIC microcontroller has several digital inputs with “weak pull-ups” that pull unconnected inputs to a digital high signal, but weak enough that inputs connected to ground can still be pulled down to a digital low signal. One side of the reed switch is attached to a PIC input with weak pull-ups, the other side is connected to ground. When the PIC sees that the arm has reached the limit of its arc, it toggles the direction signal to the motor controller.
The inner board of the IMP3 MP3 player, removed from the case, with power wires and the play button wires soldered on.
The main logic board which contains the PIC 16F648A and M-957 DTMF decoder.
Saturn windshield wiper motor.
L298 based motor controller.
A magnetic reed limit switch.
The IMP3 MP3 player, from the outside.
The system to split the stereo output of the MP3 player into left (audio) and right (control) channels.
The impendance matching audio transformer.
The front of the ultrasonic proximity sensor.
The back of the ultrasonic proximity sensor.
The main logic board, controller board, and MP3 player boards all connected together.
In previous sculptures, I’ve used Windbond Chipcorders for the audio system. They are a single chip audio recording and playback system with non-volatile memory. Unfortunately, they always seem to sound like a recording, and don’t provide high-fidelity audio. For “Blame,” I decided to use an actual MP3 player. The IMP3 from Flash Memory Store uses SD/MMC cards to store audio, and also has a USB port to connect to a PC to transfer MP3 files onto the player. The IMP3 “play” button toggles audio playout on and off. In order to operate the IMP3, I soldered wires to each side of the play button. That way, the IMP3 can be controlled just as if someone was pushing the play button to start and stop playout. The wires connected to the sides of the play button are attached to a relay that is controlled by the PIC.
In order to know where one audio clip ends and another begins, I used a technique of self-timing. The left channel of the MP3 files is used for audio and the right channel is used for control using DTMF tones (dual-tone multi-frequency, also known as touch tones). The audio out from the IMP3 is split into the left and right channels using audio splitters from Radio Shack, and the right channel goes into a Teltone M-957 DTMF decoder chip (from Electronics Parts for Less), whose output goes to the PIC. When the PIC sees a certain DTMF tone, it knows that the audio clip is over, and the PIC then energizes the play button relay to stop the IMP3 player. The MP3 files were edited using the open-source program Audacity.
I found the default audio output of the IMP3 is a little weak when I hooked it into a standard powered computer speaker system, and rather than trying to interface to its pushbutton volume system, I use an audio output transformer from Radio Shack. The transformer helps to match the low-impendence output of the IMP3 intended for headphones to a high-impedance signal appropriate for the powered computer speaker system.
I enjoy making interactive kinetic sculptures using PIC microcontrollers. In some cases it might be easier to use a PC for control, but I’ve found PCs to be notoriously undependable in long-term art installations. Nothing is more embarrassing than having an artwork crash halfway through a show. The boot-up time of a PC can be a negative in an artwork as well. The PIC microcontroller always starts up immediately when power is applied, and it takes up much less space and power than a PC as well.
In January, 2006, I exhibited “Blame” as part of a group technological art show at the Fraser Gallery in Bethesda, Maryland. And it sold!
Reference Web Sites:
Electronis Parts for Less
Flash Memory Store
10 thoughts on “Interactive, kinetic sculpture “Blame” – how it was made”
I for one welcome President Bush’s new supreme court nominee.
Instead of using a relay to trigger the external sound player, I was able to use a NPN transistor (which is cheaper and more reliable.) This also works on many children toys. The buttons on the external boards seem to all have 2 wires, one connected to ground, and the other pulled up to Vdd. So connect the collector pin of the transistor to this wire that is pulled up. Connect the emitter to ground (just like the other switch wire.) Make sure you connect the ground from your circuit to the external device. Then connect your output pin to the base pin of the transistor with a 10kOhm resistor (to limit the pin output current so you don’t destroy your microcontroller). When the output pin is low, the transistor has a super high resistance and the switch wire is still pulled high. When the output pin is high, the resistance of the transistor is greatly reduced so the wire will be pulled low and trigger the device. This will save you a few dollars each time.
If you can get an NPN transistor or a FET to work as the switch, you should!
I often run into problems on these MP3 player switches though. MP3 players are engineered to be ultra-low power portable devices, so they often run on voltages below standard 5V logic, plus the switches are often not simple Vdd/GND affairs.
If one had the schematics, one could arrange a solid-state control interface between the microcontroller with the MP3 player, but lacking the schematics, the relay solution always works :)
There is another issue, the risk of coupling noise on the microcontroller line into the MP3 player. Besides digital noise, there is 60 Hz ground loop risk due to running different power supplies for the microcontroller and the MP3 player.
Again, if you spend a lot of effort you can design a way to filter the noise out and avoid the 60 Hz ground loop problem, but the relay solution always works…
Suffice it to say I did try to do a bipolar and FET solution before I gave up and went with the relay! If someone can solve the solid-state solution to the IMP3 play button, please let me know.
Very good article!
Wow! This is really intense and the outcome is so cool! I think it would be interesting to see how zeiss conquest binoculars are made and all the high-technology that goes into them. Thanks for sharing this though!
Wow! This is really intense and the outcome is so cool! I think it would be interesting to see how
binoculars are made and all the high-technology that goes into
them. Thanks for sharing this though!
Comments are closed.