Research Associate, Union Carbide Corporation, R&D Technical Center, Somerset, NJ
Principal Engineer, Air Products & Chemicals Inc., Allentown, PA
Electronics Engineer, Union Carbide Technical Center, Bound Brook, NJ
(Received 31 March 1986; accepted 13 February 1987)
The most common and popular way to determine the amount of smoke generated during combustion of materials, including electrical insulations, is to measure the decrease in light transmission due to the accumulation of smoke in a closed chamber. The test equipment most used for this purpose today is the commercially available NBS Smoke Density Chamber.
Since the first published description of this apparatus in 1967  several modifications to the chamber have been suggested and alternative methods for analysis of the data obtained have been proposed. Some have been evaluated and a few adopted, but in general they have tended to make the operation and analysis more time consuming and complex.
In order to (1) simplify operation, (2) improve precision, and (3) automate data reduction, the complete procedure has been computerized. This includes operation of the chamber, range changing of the photometer, insertion and withdrawal of the optical filter, accumulation of light transmission values as a function of time, and subsequent calculation and printing of smoke parameters. Replicate specimens are used and corrected specific optical density results can be previewed, truncated if necessary, and averaged. The final data are stored on disk in the Data Interchange Format (DIF) for subsequent use in graphics plotting programs.
An APPLE II Plus computer with an ISAAC interface (Integrated System for Automated Acquisition and Control) manufactured by Cyborg Corporation is used for data collection, reduction to optical density, and analysis. Software is written in BASIC. A specially designed electromechanical system which contains relays for photometer range changing and a small motor for moving the filter operates upon signals from the computer. The standard NBS parameters (Ds after various times, Dm (corr), and times to reach specified values of Ds) are printed as well as SOI, SI, and MOD [1–4].
An estimate of overall test precision and a few typical test results are presented, although they are not intended to reflect or predict possible or relative hazards of any material under actual fire conditions.
Paper ID: JTE11022J