01 · The hypothesis
Can PBDee reduce the initial shock before it reaches the grip?
PBDee places compressed Sorbothane around the paddle neck, directly in the path between the paddle face and the player’s hand.
How we expected PBDee to work
1. Start with the impact
When the ball hits the paddle, the paddle face flexes and produces a brief mechanical shock. Part of that vibration travels from the face, through the neck, and toward the handle.
2. Place damping material in that path
PBDee holds compressed Sorbothane around the paddle neck. Sorbothane is viscoelastic, meaning it has both elastic and viscous behavior. It deforms under load, but it does not return all of that mechanical energy in the same way a simple spring would.
3. Dissipate part of the mechanical energy
As the Sorbothane deforms and recovers, internal friction known as hysteresis converts a small portion of the mechanical energy into heat. The expected result is a smaller initial shock continuing from the neck toward the lower grip.
4. Test the prediction directly
We did not treat the proposed mechanism as proof. The upper accelerometer measured the input from each bounce, and the lower accelerometer measured what reached the grip. The testable prediction was straightforward: for a comparable upper-sensor input, the lower-sensor peak should be smaller with PBDee installed.
Mass, stiffness, boundary conditions, and resonance may also contribute. This experiment measures the practical outcome, but it does not isolate the contribution of each physical mechanism.
02 · Sound experiment
Did PBDee change the sound of a controlled impact?
A suspended pickleball repeatedly struck a secured paddle while a microphone remained in a fixed position.
The reduction was concentrated near 1.6 kHz. It was not a 10 dB reduction across the entire sound spectrum.
Full apparatus, procedure, and interpretation
Apparatus
A pickleball was suspended as a pendulum and released toward a secured paddle. A microphone was held in a fixed position approximately 0.5 m from the ball and paddle contact point.
Procedure
- Secure the paddle and keep the microphone position and recording settings fixed.
- Release the suspended ball toward the same general contact region.
- Record approximately 10 impacts with PBDee removed and approximately 10 with PBDee installed.
- Compare the recorded impacts in the frequency domain using ΔL(f) = LON(f) − LOFF(f).
Result
The comparison showed an approximately 10 dB lower spectral level near 1.6 kHz with PBDee installed. A 10 dB difference at one frequency corresponds to about one tenth the sound intensity at that frequency under equivalent measurement conditions. It does not mean the complete impact sound was 90% quieter or that total perceived loudness fell by 10 dB.
03 · Vibration experiment
Two sensors measured what entered the handle and what reached the grip.
The upper accelerometer measured each impact’s input. The lower accelerometer measured the transmitted response farther down the grip.
Manual impacts vary. The upper sensor allowed us to account for how strong each hit was instead of treating every bounce as identical.
Full apparatus and procedure
| Sensors | Two ADXL375 three-axis accelerometers |
|---|---|
| Sampling | 800 Hz; 1.25 ms interval; nominal 400 Hz bandwidth |
| Capture | 160 samples spanning 200 ms per impact |
| Trigger | Upper sensor at 20 m/s²; trigger index 15 |
| Pre-trigger buffer | 16 samples, approximately 20 ms |
| Calibration | 500 stationary samples before each run |
| Order | 100 PBDee ON, followed by 100 OFF |
Procedure
- Mount and calibrate both sensors.
- Dribble the ball while maintaining approximately the same grip, paddle angle, impact region, and sensor placement.
- Record 100 valid impacts with PBDee installed.
- Remove PBDee without intentionally moving the sensors, then record 100 impacts without it.
- Verify 160 samples, trigger index 15, and zero recorded I²C errors in every trial. All 200 final trials passed.
04 · Result
PBDee was associated with a smaller initial peak toward the grip.
The primary analysis adjusted the lower-sensor response for the strength of the upper-sensor input.
This result applies to the 200-impact paddle-dribble protocol after accounting for measured input strength.
Full processing, regression, and numerical results
Signal processing
For each axis, the median pre-impact baseline used samples at t ≤ −5 ms. Samples at −2.5 ms and −1.25 ms were excluded. Vector acceleration was calculated as a(t) = √[x(t)² + y(t)² + z(t)²]. Upper and lower peaks were the maximum values from −2.5 ms to +20 ms around the trigger.
Primary model
ln(Lᵢ) = β₀ + β₁ ln(Uᵢ) + β₂ ONᵢ + εᵢ. Ordinary least squares used HC3 heteroskedasticity-robust standard errors. Percentage difference = 100 × [exp(β₂) − 1]. Reported R² = 0.3379.
| Measure | OFF | ON |
|---|---|---|
| Mean upper peak | 29.8186 m/s² | 30.8298 m/s² |
| Mean lower peak | 20.3099 m/s² | 18.3127 m/s² |
| Mean L/U ratio | 0.6670 | 0.5901 |
| Median L/U ratio | 0.6683 | 0.5466 |
| Adjusted effect | −13.396% | |
| 95% confidence interval | −20.704% to −5.414% | |
| HC3 p value | 0.001387 | |