Can Noise from Race Cars Break Glass?

David Braslau - david@braslau.com
David Braslau Associates, Inc.
Minneapolis, MN

Popular version of paper 3aNS6
Presented Wednesday morning, May 20, 2009
157th ASA Meeting, Portland, OR

Everyone has seen the television commercial of an opera singer breaking a wine glass. In real life this is very difficult to do, although it is possible in a laboratory setting where a sufficiently high sound level can be directed at the glass.  When the frequency of generated sound is the same as the natural or inherent vibratory frequency of the glass, it can increase the vibratory motion of the glass, similar to how a child continually pumps a swing to increase the height of the swing.  The natural frequency of the glass can be determined by rubbing a finger around the top rim of the glass until the glass starts “ringing.”  This tone is the frequency at which the glass will vibrate and radiate sound and is therefore called its “natural frequency”.

In the study discussed here, a large tempered glass window in a spectator box at northern Minnesota speedway with a dirt track shattered during a race by the World of Outlaws Late Model Series, injuring a spectator sitting below the window.  Legal counsel for the spectator claimed that noise from the race cars had shattered the glass and later provided a video in which a pane was shattered in a test setting by a loudspeaker facing the glass.  The claim was based on the assumption that defects at the edge of the glass when it was installed enabled the otherwise strong tempered glass to break when excited by high energy sound waves.

In response to this claim, an evaluation of physical properties of the window and measured sound level data from a similar speedway and a subsequent field test with a loudspeaker simulating sound based upon data from an actual race of World of Outlaws cars were used to determine whether the window could be shattered by race car noise.

Most race grandstand spectators are probably unaware of the level of noise to which they are exposed by “Late Model Series” cars.  Figure 1 shows the sound level time history of an evening of races at a similar track measured at the top of the grandstand, approximately 155 feet from the center of the track.  This shows a typical increase in sound level over the evening as more and more powerful (and louder) cars are used.  During the final race, the level of 103 dBA at the top of the grandstand was equivalent to 110 dBA for persons sitting at the front of the grandstand.  The A-weighted decibel (dBA) is commonly used in the evaluation of noise impacts and is a single number measure that combines all frequencies into one value.  The federal Occupational Health and Safety Administration limits sound level exposure to 115 dBA.

Figure 1

The following questions were addressed:  Why didn’t the window break earlier with other noisy cars?  Are the World of Outlaws cars that much louder that they could shatter a window?  Only one race by this class was to take place before the end of the season and was rained out.  Therefore, a theoretical analysis using the maximum level in Figure 1 was used to evaluate the window. Since the box at the speedway being studied was closer to the track than where the Figure 1 measurements were taken, the level was adjusted upward based on relative distance and a maximum level of 110 dBA was estimated at the window.  The frequency spectrum associated with the maximum dBA level shown in Figure 1 was compared with published curves on when windows, walls and floors begin to rattle.  It was found that the speedway box window rattle limit would be exceeded by 27 dBA at 63 Hz and would therefore rattle, although no conclusion as to breakage could be established.

The theoretical analysis did not provide any conclusive evidence one way or the other on breakage, so it was determined that levels from any actual race should be taken.  Campanella Associates, another member firm of the National Council of Acoustical Consultants, was contacted and asked to take measurements of the final race of the season in Ohio.  After considerable obstacles and some rain, the World of Outlaws race finally started at 1 a.m.  Detailed sound level data were acquired with a high quality sound level meter and emailed to this author for analysis.  A number of adjustments in data had to be made due to acquisition with a different meter, some different meter settings, and different distance from the track.  Comparing the adjusted World of Outlaws data with the Late Model Series data, sound levels of the World of Outlaws at critical frequencies of 160 Hz and below exceeded those from the Late Model Series only at 63 Hz, and only by 2 dB.  Thus, one would have expected the window to have broken in the previous two years of exposure to other race cars.

In the meantime, a video on DVD of an indoor test by opposing counsel was received showing a pane of tempered glass with an induced “defect” being broken by sound from a loudspeaker placed immediately in front of the glass.  However, supporting information was not sufficiently detailed to draw any conclusions other than the fact that continuous loud tone was generated until the glass shattered, similar to the wine glass experiment.

It was determined that a test of the replacement window with an induced defect similar to the one used in the indoor test would have to be made to determine once and for all if noise from race cars could have broken the window.  The time history spectrum of the World of Outlaws race was then simulated by Synergistic Design Associates with a digital signal processor and played back through a large loudspeaker 10 feet in front of a reinstalled window (see Figure 2).

Figure 2

Two simulations of the window as currently installed were performed with a maximum level of 110 dBA, but varying the level from 98 dBA to 110 dBA and back to 98 dBA each 15 seconds to simulate cars circuiting the track.  No detrimental effect on the window was observed.

The third test was performed with all quarter-rounds removed from the interior of the window.  For this test a maximum level of 116 dBA was used, which is not quite twice as loud as the 110 dBA level, and higher than could be expected at this location with vehicles powered by internal combustion engines.  No detrimental effect on the window was observed.

The fourth test was performed with an intentional defect created in the window similar to that described by the Plaintiff’s counsel.  The maximum level of 116 dBA was again used and no detrimental effect on the window was observed.

Finally, a light weight accelerometer was taped to the center of the window and acceleration levels measured.  The shape of the vibration spectrum was similar to the measured sound spectrum outside the window and did not reflect any significant displacement of the window that might occur at the natural frequency of the window.  Moreover, no rattle or significant movement of the glass was observed from inside the box.

The calculated lowest natural frequency of the window was 13 Hz, well below race car frequencies with maximum energy.  The window can also vibrate at multiples of 13 Hz and could probably be driven to shatter if excited with sufficient energy at an appropriate frequency for a sufficient length of time.

However, like a swing being pumped by a child, it takes a continuous input of energy to get the swing higher and higher, just as it takes a continuous input of sound energy to get the glass (or a window) to vibrate more and more.  In the case of race cars, the maximum level lasts only several seconds that is not sufficient to excite and shatter the window.

Finally, a few observations are needed to put window breakage from sound, blasting or wind pressure into perspective.  The static pressure from a 100 mph wind is equivalent to 156 dB or 600 times the sound pressure in pounds per square inch.  Whether the glass breaks under wind load will depend upon its area, shape, and other factors.  Frequency and duration of a 100 dB sound pressure are much more critical than a static wind pressure of the same magnitude. However, a rapidly rising acoustic or blast pressure of 130 dB is commonly used as a design limit for blasting to limit window breakage.  Glass usually breaks when the blast level exceeds 150 dB.

In summary, while only one window was tested, physics and our results suggest that noise from race cars cannot break windows.  However, an extensive research project would be needed to extend this conclusion to a wider range of conditions.