book collections email follower instructable user

The soothing effects of chimes normally depend upon the wind, but this wind chime is Arduino powered.

Rather than wait for the wind to blow, it is possible to excite a series of tuned pipes using solenoids controlled by an Arduino program (sketch). I will discuss my design and the many factors that entered into its creation.

Yes, by adding a speaker to the Arduino, a simulated wind chime can be created, but nothing sounds as good as the real mechanical version.

Step 1: The chimes themselves

Picture of The chimes themselves
test fixture.jpg
Available notes.jpg

My original prototype chime player utilized some anodized aluminum tubing purchased from a Chinese source that were all pre-cut to a fixed length and drilled with a hole to hang them. They were all 17 cm in length and therefore had to be cut to various lengths to produce the correct notes. A web site that discusses ALL aspects of wind chime design is by Lee Hite . It is an amazing collection of everything you ever wanted to know about chimes and provided much insight into my design.

I determined that my chimes would primarily be excited in their first natural frequency and made a simple frame to allow me to "ring" each chime and measure its frequency. A free computer program (Audacity), allows recording of the sound of a "struck" chime and you can easily determine its fundamental frequencies. A great reference chart for frequencies of the equal tempered scale can be found here.

For the 17cm long tube I had, I determined that its first natural frequency was 1661 hz. which was G#6 in the key signature of "C". I decided to tune my first chime player to a different key signature so that the first chime would represent an "A" note. Since I wanted to use only 8 chimes, the subsequent chimes would be tuned to B,C,D,E,F,G,&A.

Rather than go through lengthy calculations, I found it easier to simply cut a tube at a different length and measure its fundamental frequency. Then some simple math comes into play.

The basic equation for the vibration of tube looks something like this...

Freq= (Blah,Blah..Blah) / (Tube length) squared.

Since the tubes were of the same material and only the length varied, a simple equation can be made to determine the length of tube required to produce any required frequency.

(Length 2)= (Length 1) * Sqr Root (Freq 1/ Freq 2).

Easy to compute on a spread sheet.

While it appears possible to compute all the required lengths at once, I found it better to continuously recompute the next required length from the previous one. Rather than go into details regarding the 8 note chime player, let me go into details of the 16 note player.

K & S Engineering produces thin wall aluminum tubing, 5/16"diameter, available in 36" lengths. It is sold in a package of 4 from Amazon for about $10.50 plus shipping. It is enough material to produce16 tuned chime tubes.

While the 8 note player was nice, there were many songs that had too great a range or included sharps or flats that the 8 note player couldn't handle. The 16 note player was tuned in the Key of C with the lowest chime at a frequency of 2637 hz which is E7.

The note/frequency/tube length list is shown below:

Note Frequency Length (cm)

  1. E7 2637.02 22.3
  2. F7 2793.83 21.6
  3. F#7/Gb7 2959.96 21.0
  4. G7 3135.96 20.4
  5. G#7/Ab7 3322.44 19.6
  6. A7 3520 19.2
  7. A#7/Bb7 3729.31 18.6
  8. B7 3951.07 18.2
  9. C8 4186.01 17.5
  10. C#8/Db8 4434.92 17.0
  11. D8 4698.63 16.5
  12. D#8/Eb8 4978.03 16.0
  13. E8 5274.04 15.6
  14. F8 5587.65 15.1
  15. F#8/Gb8 5919.91 14.6
  16. G8 6271.93 14.2

The tubes were measured, marked to length and cut using a simple tubing cutter. After each tube was cut, its frequency measured and if close to the desired frequency (within +/- 50 hz) set aside. If the tube was outside of limits, the length was adjusted and another tube cut. (The rejected tube was reused to make a higher frequency chime so there was little waste.)

The measurement of frequency using the Audacity program involves supporting the chime at approximately its "node" points which are located at 22.4% from each end. A microphone feeds the sound into the audacity program and the tube is "struck" several times while the signal is being recorded. The program then allows you to view any section of the produced wave signal and analyze the signal by plotting its spectrum. The various peaks of the spectrum are easily read off the graph, and you quickly can determine which peak represents the first fundamental. I suspended the tubes using 6 lb test nylon fishing line. It's what I had. Heavier lines may also be suitable and possibly easier to see, but this light line had very little effect on the tubes.

Also, the test frame evolved; with the tube first being struck by a tap from a screwdriver, into a solenoid driven striker actuated by a battery and push button. This also gave me a test bed to test my solenoid system.

gada8883 years ago
Thanks,nice project.very informative
albenbrewder (author) 3 years ago


I have added a little more content to this project as described in the new step 12.

I added the real time clock (RTC), and developed a couple of simple programs.

The first plays a song every hour (Watermark). The second plays the BIG BEN (Westminister) chime sequence every quarter hour.

albenbrewder (author) 3 years ago

Thank you

Always nice to receive a nod of approval from a master.

Lee Hite3 years ago

Congrats on an terrific accomplishment. I am constantly amazed at the creativity and well designed projects by site visitors, and this is certainly one of them.. The sustain time is noticeable and adds a wonderful character to the overall presentation. Very nice