Noise dosimeters: on measurement reliability and instrument accuracy

Abstract

Sporadic reports of substantial variations in measurements of exposure noise dose in the workplace performed with contemporary dosimeters have generated concern in the community of users. Measurement differences ranging up to many hundreds of percent have been reported. Differences have been observed when several dosimeters of different makes have been used simultaneously as well as in instances where apparently identical dosimeters have been used. Moreover significant differences have been observed in day to day tests on a given worker. These tests have been made with the same dosimeters in an apparently invariant working environment. Careful investigations of the properties of the ambient noise fields and limitations of microphones worn on the body explain, and sometime mitigate, some of the measurement variations. Remaining differences have been attributed by some users to inadequate “accuracy” of the dosimeters. ANSI S 1.25 stipulates that type 2 SLM performance is required for dosimeters. Some individuals have proposed changing to type 1 dosimeters (of even type 0 SLM's). The problems are quite complex and cannot be resolved by simple upgrading or resorting to SLM's. This paper identifies several sources of measurement variability which are introduced by generic configurations of dosimeters. These include: detector performance in the presence of impulsive noise when a 5-dB exchange ratio is used, and uncertainty in achieving threshold cutoff. It is concluded that present standards and applications practices must be changed to achieve improved measurement reliabilities.

Letter to the Editor

“…In addition, calculation of TWA and LA eq8hr at the time that the reviewed studies were conducted (30–40 years ago) was accomplished without the aid of instruments capable of integrated measurements. The authors’ conclusions do not reflect that the over-prediction of hearing loss could have resulted from inaccurate estimates of actual exposure, due to a prediction error from methodological/instrumentation differences (Earshen, 1980, 1994). More recent research has shown that the adequacy of the exchange rate depends on both noise intensity and noise kurtosis (Lataye and Campo, 1996; Davis, Qiu, Hamernik, 2009), and that the accuracy of L eq predictions of hearing loss risk depends upon duration (hours per day), intensity, and type (continuous vs. intermittent) of noise exposure (Hamernik, Qiu, Davis, 2007; Danielson et al., 1991).…”

Response to Suter and NIOSH

“…However, neither of their cited references (two abstracts by Earshen 1980Earshen , 1994 suggests systematic overestimation of exposure by earlier sound level measurements. We clearly indicated in our paper that hazard estimates for exposures rich in impact or impulse noise might require additional considerations, such as the measurement of kurtosis.…”