Novel Tubular Bell Player - Part 1

This is very much "work in progress" in the development of an experimental tubular bell playing device.

In contrast to the usual method of striking a tubular bell to raise a note, I am attempting to use electromagnetic impulses. Each of the aluminium tubes has a steel annulus attached at the lower end. A nearby electromagnet is fed with a signal at the tube's resonant frequency; this causes the tube to vibrate in sympathy with the signal. The 16mm diameter tubes are suspended at a stationary point (about 22% from base) using a brass plug and 2mm steel wire. The 7 tubes are tuned (using a file!) to the notes D5, E5, F5, F#5, G5, A5,and B5; I plan to have 16 notes in the final version.

The resonating signals are generated using an ATMEGA328 microprocessor on an Arduino board. The ATMEGA has a nice 16-bit PWM which provides great control over frequency (better than 0.2Hz). The single PWM signal is sent through a demux (74HC4514), then a TIP120 Darlington transistor to drive each electromagnet. The electromagnets are made from boxwood with a 4mm mild steel core and a winding resistance of about 4 ohms.

I do a little envelope shaping of the PWM signal to avoid harsh sounds generated by fast leading or trailing edges to the PWM bursts. I (naively) thought I could damp notes by phase changing the signal at the end of a note - this has proved very difficult. The bells resonate for some time after they have been activated causing a rather "muddy" sound. I have considered developing a mechanical damper for each tube but made little progress with this.

The note sequence (music?) is held in memory but I plan to interface the controller to a simple keyboard, much like on some manual carillons.

Polyphonic playing should be possible by "dithering" the PWM signal between bells. However, I have found that if a signal is sent to an already resonating bell, it must accurately be in phase. This raises quite stringent timing requirements if polyphonic playing is to work. I expect another approach to polyphonic playing would be to use separate processors (or simple oscillators) for each pipe.

The rather uninspiring noises generated so far will not concern Mike Oldfield! However, I'm working on it.

Category:

Science & Technology

Tags:

• tubular
• bells
• carillon
• automatic
• microprocessor
• music
• electromagnetic
• impulse
• pulse-width
• modulation