Ernest Chladni (1756-1827) discovered that, when introduced to different frequencies, solid plates are excited and vibrate in different locations. He developed a method of viewing the areas in which the excitement is occurring by sprinkling a fine aggregate on the surface of the test piece and then introducing it to a certain frequency. Chladni discovered a way of doing this by drawing a bow across the edge of the plate until it vibrated, however this method has since been mostly replaced by using a speaker and a signal generator. As the frequency is adjusted the pattern will change dramatically and it will become increasingly complex as the frequency increases. Visible patterns form as the aggregate settles onto the dead (not vibrating) areas of the test piece. These areas are known as nodes and the areas where piece if vibrating are known as anti-nodes. As the frequency changes the nodes move around the face of the plate creating different patterns. On a small metal test piece the location of the nodes is mathematically predictable, however something as complex as a musical instrument top would be far too difficult to mathematically predict.Click here to see how it works on a perfectly flat, square surface.
For the Chladni tests we used an amplifier, speaker, and signal generator. We used this setup to test our guitars at each step of the construction process. The top, top with sides(rim), and the complete body.

Equipment
Elenco Audio Generator- Model SG-9200
7 in diameter- 8 Ohms Studio Speaker
Pioneer Stereo Amplifier SA-410- plugged into “phono” setting

At the beginning of the project we tried to use a Peterson VS-II stroboscopic tuner as the signal generator to provide us with notes found in the A440 scale. This became somewhat discouraging because we were having trouble picking up many clear readings no matter how high or low of frequencies we were projecting at the plates. Eventually we switched over to using a signal generator with the full range of frequencies, however, and this allowed us to hone in much more precisely on exact patterns which the plates were clearly generating. We found that sometimes a change of only 1 or 2 Hz could change the pattern dramatically. Other times a plate might hold basically the same pattern over a range of 10 Hz or more.
Another observation we made was that it was critical that the sponges which are suspending the plate over the speaker be soft and in full contact with plate and speaker housing. Often they can be softened by moistening them slightly. The location of the sponges can also slightly change the patterns being produced. We elected to use three sponges on the guitar tops (one on each side of the lower bout and one at the top) so that we had the least amount of contact possible.
The exact distance which the plate is suspended above the speaker didn’t appear to alter the patterns in our tests as long as the plate was vibrating enough to move the aggregate. Usually between 90 and 120 decibels when measured at 6 inches away from the speaker was sufficient to drive the plate. Higher decibel levels will force the pattern to develop faster but with time the same patterns will form at lower decibel levels as long as the plate is vibrating somewhat. Especially at high frequencies (and therefore more complex patterns) we had to repeatedly cover the surface with poppy seeds in order to show the fine node lines. Otherwise the nodes can sometimes be overdriven and won’t appear in the pictures.
. In order to test whether or not outside sound could interfere with Chladni patterns we built a noise isolation box and ran some tests with all the test equipment inside of it. We then ran the same tests with a radio playing at 100+ decibels 6 inches away from the plate and found no difference in the patterns we were getting.

(Frequencies have been distorted by fast forwarding)