Methodology and Standard Weights for Accuracy Testing on Four Mechanical Balances

by David W. Davenport, January-February 2000

Purpose: To determine how three mechanical "pocket scales"—Accura, Calibrön, and Deering 10 gram —compare in accuracy to the extremely accurate Mettler AE 100 analytical balance (Swiss, electronic).  Also determine how the laboratory-quality Ohaus Dial-O-Gram 310 Model mechanical bench balance compares to the Mettler AE 100 analytical balance.  Compare and contrast all four mechanical scales for accuracy, features, and ease of use.

Method: A calibrated Mettler AE 100 analytical balance (accurate to 0.0002 grams = 0.2 mg) was used to acquire three replicate weights of objects used as mass standards.  These standards consisted of Ohaus calibration weights (from Model 218-16 set, 10 mg to 500 mg, ANSI/ASTM E617, 1991, Class 6; as well as an Ohaus 1.000 g mass), and common United States Mint coins (penny, nickel, dime, quarter).  These objects provide a practical range of "standard weights" for determining the accuracy of three mechanical pocket scales (Accura, Calibrön, and Deering) with ranges of 0 to 4 grams, or 0 to 10 grams.  In addition to the pocket scales, the Ohaus Dial-O-Gram 310 Model mechanical scale (a top-of-the line mechanical bench balance) was tested in its 0 to 10 gram range (this scale measures up to 310 grams, with readability to 0.01 gram = 10 mg).

Procedure: The Mettler AE 100 balance was leveled, then calibrated using an ANSI/ASTM E617 Class 1 100 g, ± 0.00025 g mass at the start of the weighing project.  (The Mettler balance was professionally serviced by a certified calibration technician, just four months ago.)  The balance was tared prior to each replicate weighing of each "standard mass."  The four mechanical balances were placed on a flat, non-skid surface, then the beams were leveled, and the "zero" checked prior to each replicate weighing.  The weighings were conducted in a room with ventilation fans turned "off," to eliminate "swinging" by the beams.  Forceps were used to manipulate all Ohaus calibration masses, and latex gloved fingers to manipulate all coins—to insure the masses did not pick up any "extra weight" from handling.  Statistics were used to find the mean, standard deviation, and coefficient of variation (CV) for each set of three replicate weighings.  The mean value of each standard, per the Mettler balance, was then used for comparison to the mean value of each standard, per each of the four evaluated balances (Accura, Calibrön, Deering 10 gram, and Ohaus 310-00)—see following  spreadsheets.

Note: In the spreadsheets and graphs on the following pdf files (Accura, Calibrön, Deering 10 gram, and Ohaus 310-00), the Mettler mean weight is the "independent variable" on the X-axis; while the mean weight of the mechanical balance under testing is the "dependent variable" on the Y-axis.  A "perfect" value for the line shown on each graph is a linear equation (y = mx + b) of "y = 1.0000x + 0.0000" and an R² value of "1.0000."

Summary of Accuracy Determinations for Four Mechanical Balances

Features Comparison for Four Mechanical Scales, and Recommendations