Guitarmaking: Tradition and Technology
I assembled my first three guitars using the excellent reference Guitarmaking: Tradition and Technology by Natelson and Cumpiano. Braces were shaped to copy Martin scalloped braces.
Chladni free plate tuning
Interest in improving the sound of my guitars led me to voicing guitars using free-plate tuning with Chladni patterns. Chladni free plate patterns were performed on the guitars that I built from 2003 through 2008, including guitars from the Beta guitar to the Walnut Phi. Free plate tuning is a tool that has a long tradition in voicing violins, and over the past decades has been used in voicing guitars. The source for much of my information and guidance was provided by Alan Carruth in luthier forums. I also attended a 2007 Healdsburg Festival workshop on free plate tuning by luthier Mark Blanchard.
One of the goals of free plate tuning is to tune the soundboard and bracing to form a closed circle on the ring-and-a-half mode. My first attempts at “closing the circle” were significantly unsuccessful as seen on my Beta guitar.
A change to a double X bracing system with a bridged lower X allowed me to tune braces to get the closed circle. In hindsight, I was overbuilding the guitars even though I was able to close the circle. Bracing was 5/16” wide, soundboards were nearly 3mm thick, and the back was about 2.5mm thick.
My first fixture for free plate tuning was a speaker mounted in a sheet of plywood. The soundboard was suspended on soft foam pyramids with the speaker directly below the position of the bridge. This setup required a lot of audio power to form the patterns – sound levels that can damage eardrums. I learned in the workshop by Mark Blanchard that a better solution is to use a hand-held speaker and move it around the soundboard to find the maximum response. Note from experience: always wear ear protection.
The best media that I’ve found for Chladni particles is dried bulk tea leaves.
I use TestTone Generator software and standard external PC speakers to generate the Chladni patterns.
The Responsive Guitar
The book set The Responsive Guitar and Making The Responsive Guitar by Ervin Somogyi, along with his DVD Voicing the Guitar, opened my eyes to the importance of light construction. My understanding is that the best guitars are on the verge of collapsing from the 180# pull of the strings. The challenge is knowing when you’re near that point. While the book set provides excellent direction, I was unable to find specific information on bracing and soundboard thickness that would provide a starting point.
I decided the best way to learn was to build a guitar near or beyond the point of collapse. The experimental guitar was the Malaysian Blackwood guitar. Bracing was reduced to ¼” wide, and the soundboard was thinned to 2.2mm. I also followed Ervin’s advice to diligently tap all sections of the soundboard to get a full, consistent sound. The sound of the Malaysian Blackwood guitar was a significant step forward.
The tap response process will typically take more than a day of tapping, listening, trimming braces, tapping, listening, trimming, etc. To assist the process, I added a clamp ring to my outside mold to allow the soundboard to be removed for easier access to braces. This also allowed me to measure Chladni responses as I progressed. I wanted to observe the progress on both Chladni patterns and Responsive tuning together.
As I noted previously, the lightly built Malaysian Blackwood was a significant improvement in sound and response. And it’s not near collapsing. But I was able to further improve the sound in the next phase.
Scientific Design Methods
The book set Contemporary Acoustic Guitar Design and Build by Trevor Gore and Gerard Gilet provides a more scientific approach to voicing. With my engineering background, I appreciate the technical approach that bridges the link between science and practical methods that help focus my efforts toward learning to build a more responsive, full sounding guitar. I highly recommend this book set.
One of the key measurements is derived by tapping the soundboard with a miniature hammer while using a microphone and computer to record the resulting frequency response. Per the recommendation, I use an eraser attached to a wooden dowel as a miniature hammer to create the impulse. I use Sound Forge software to record and print the frequency spectrums. My current method is to tap five times in the middle of the bridge to capture the monopole response, tap five times on the end of the bridge to capture cross dipole response, and tap five times about one inch below the bridge to capture the long dipole response. At left is a typical chart.
We’re interested in the frequencies recorded, not the magnitudes. I’ve observed that the frequencies are repeatable as long as the guitar is in the same orientation and supported in the same way. The magnitudes of the response will vary based on the strength of the impact and location of the microphone.
A goal in voicing the guitar is to achieve a soundboard monopole response of 170 or 180 Hz, a corresponding apparent Helmholtz frequency of 90 or 95 Hz, and a back main frequency @ 214 or 226 Hz – 4 semitones above the soundboard.
Based on these measurements, I’ve modified several completed guitars by thinning braces and soundboards, including my Cocobolo guitar, Malaysian Blackwood guitar, and Rosewood guitar with subsequent improvement in sound.
A second measurement is bridge rotation. Essentially, a long wooden bar (ruler) is attached to the bridge, parallel to the strings. Measurements are taken at the ends of the bar, then the strings are loosened and the measurements repeated. A change in angle of 2 degrees is about the right amount of flexibility of the top. Much less that 2 degrees indicates a too-stiff top. Much greater than 2 degrees indicates a top that is too flexible and will likely distort over time and may lack focus. A chart of measured bridge angles of my guitars is shown, along with two premium factory guitars that I own.
A third measurement is Monopole Mobility, defined as “a measure of a guitar’s responsiveness, being the mobility of the monopole mode of vibration of a guitar.” The physical measurements involved in recording monopole mobility include the deflection of the soundboard to a known force and the uncoupled natural frequency of the top. Good guitars are typically in the 10 – 14 x 10-3 s/kg range. A chart of monopole mobilities of my guitars are shown, along with a reference of Trevor Gore’s guitar with falcade bracing. Taller bars correlate with a more responsive guitar.
There are several other voicing aspects that I haven’t yet addressed in my efforts, including intonation of individual notes.
I continue to use Chladni patterns in my construction and voicing. Before cutting out the shape of the guitar from the rectangular plate, I use Chladni patterns to determine the long grain, cross grain, and twisting vibrational frequencies for the plate. With the plate dimensions, weight, and frequencies in a spreadsheet, I can establish the target thicknesses according to the formula provided in the book. I use the information to compare and select top plates. Then, after cutting out the profile of the soundboard and back, I reduce thickness until the top and back long grain vibration frequencies are 52 and 43 Hz respectively. This gives me a dynamic measure of the stiffness of the plates.
Then while tuning braces with the soundboard and sides clamped in a mold, I use Chladni patterns to observe the monopole, cross dipole, and long dipole vibration modes to give me visual insight into how the plates are responding.
I’m tracking the coupled frequency response of guitars through the build process to better understand changes and how to achieve the target response. The chart below tracks the process of the Ziricote guitar and highlights the change in monopole responses. Observations include:
- The monopole responses have the most change. The change from “in the mold” to “assembled” is due to the stiffness and mass of the mold, and is explained in the book. Also the change due to mass of the bridge is explained.
- I do not yet understand the change due to binding.
- The cross dipole and long dipole are relatively consistent in this guitar.