Acoustic Soundboard: A Personal Design Study, Pt. 2
A side-by-side comparison of the original Martin Thinline 332 and the second-generation model shows some similarities, but there are important differences in saddle design and the size, shape, and layout of the piezo elements
As promised in my last column [“A Personal Design Study, Pt. 1,” October 2013], we’re going to go under the hood to take a close look at the design details of the original Martin Thinline 332 pickup, and then see how it differs from the second-generation version.
Both versions share the same design intent: sensing the change in force on the saddle created by the vibrating strings. Both also share several basic features: They are sized to fit under the saddle in a traditional 3/32" wide slot. They both have six piezo-ceramic elements aligned in an array that places each individual element directly under each string (where the individual elements are electrically connected in parallel). And the arrays for each are wrapped with conductive material connected to ground.
So at first glance, the designs appear quite similar. However, they are completely different in many important ways, resulting in different mechanical and tonal performance.
The most obvious feature differentiating the versions is the original’s split-saddle, phase-canceling feature. Reversing the polarity of half the elements and isolating the two groups by splitting the saddle cancels most of the common-mode body vibrations (think mechanical humbucker). This technique can be used to fight feedback, but it also cancels some of the essential resonant character that defines the sound of an instrument. For me, the inherent mechanical instability of the split saddle made this design choice a nonstarter.
With the original design as a benchmark, I constructed several prototypes using a variety of piezo-ceramic materials, altering element sizes and testing the impact of the shape of those materials. Evaluating the new prototypes was very informative, because I learned that designs employing sensing elements similar in size to the original 332 tended to feed back at even moderate volume. However, designs employing significantly smaller elements were far more feedback-resistant. It became clear why the designer of the original 332 turned to the phase-canceling approach.
As it turned out, uneven loading of the larger sensors was allowing them to oscillate in the slot and become the source of much of the acoustic feedback. Understanding the various methods of manufacturing piezo ceramic materials was critical to solving the loading issue, especially as it relates to using individual sensors for each string.
In the original Thinline design, the sensing elements are rectangular. Ceramic parts of this size and shape are generally made by dicing up a larger sheet of the material into smaller parts. These large sheets are manufactured via a process called “tape casting.” This technique spreads a slurry of wet ceramic material onto a moving belt, and then uses a device called a doctor blade to control the thickness of the slurry. This slurry is then fired in an oven to create the hard ceramic. The problem with this approach is that these larger sheets often exhibit significant thickness and density variation across the sheet. This resulted in wedge-shaped parts that were the root cause of the feedback due to uneven loading.
Ultimately, I decided to use very small, disc-shaped elements. These elements are formed by “slug casting” a rod created by pressing wet ceramic into a tube-shaped mold. After firing in the oven, the rod is “centerless ground” to a very precise diameter and then sliced with a precision diamond gang saw into extremely flat and density-consistent discs. The smaller discs have the added benefit of increased electrical output due to the higher stress concentration in each element.
Further Thinline 332 refinements included removing the spacers that positioned the elements in the original design and conductively bonding the disc-shaped elements to a carbon-fiber substrate. By separating the sensors in the second-generation Thinline, I eliminated the phase-canceling crosstalk generated by transverse vibrations introducing shear forces into the adjacent elements. The single-sided bonding also turned the discs into pseudo bending mode sensors, further increasing their output.
Another improvement concerned the shielding. The original design was shielded with copper tape backed with a soft and fairly thick pressure-sensitive adhesive. For the second-generation design, I formed the shield with conductively treated paper held together with a very thin and very hard thermosetting film adhesive. This material change greatly improved the sonic performance by eliminating the high-frequency damping effect of the original’s softer materials.
There we have it: Two acoustic guitar pickups that initially look quite similar turn out to be very different in design and function. There is a common misconception that all under-saddle acoustic guitar pickups are basically the same, but nothing could be further than the truth. As with magnet/coil electric guitar pickups—where you have literally hundreds of very different-sounding pickup options that may look identical at first glance—there are a multitude of under-saddle piezo pickups for acoustic instruments that differ widely in terms of touch sensitivity, balance, dynamic range, feedback resistance, and tone.
To wrap it up this month, let’s not forget the quote from the great Charles Eames that I included in my last column: “The details are not the details. The details make the design.”