Abstract
IN a paper read to the Institution of Electrical Engineers on May 1, Mr. C. C. Paterson pointed out that there is always a tendency to judge a manufactured product by its showing one desirable feature rather than from its uniformity. He said that with certain exceptions, such as the laws of gravity, there is a remarkable absence of uniformity throughout Nature. There is a tendency to admire extremes, such as the tallest mountain or the longest river. This explains why there is a certain distrust in the pursuit of uniformity. The engineer is rather apt to take it for granted. Much attention is paid to class -testing, but the same effort is not made to ensure that every product is up to the level of the class tested. In the case of electric boiling plates, for example, if the supply voltage is 6 per cent low the consumption of energy compared with the heat developed is increased 14 per cent. Variations such as these are superimposed on the manufacturers' permissible limits of variation now called ‘tolerances'. Fifteen years ago, the manufacture of glass was largely a matter of rule of thumb methods, with the result that in making lamps about 150 bulbs were used for every 100 lamps. When an effort was made to obtain glass of uniform quality with the desired characteristics, only 104 bulbs were used for each 100 lamps produced. In the early days of dry battery manufacture, great stress was laid on individual output. With the growing use of multicellular batteries, where failure of any unit meant failure of the whole battery, attention was concentrated on securing uniformity, with the result that the failures now are of the order of five in a million. In conclusion, Mr. Paterson dwelt on the importance of using frequency curves when assessing the deviation of products from the standard. Instead of using tolerances, it would be much better to adopt the coefficient of variation or the standard variation from the frequency curve by the method used by statisticians.
Article PDF
Rights and permissions
About this article
Cite this article
Uniformity as the Gauge of Quality. Nature 137, 861 (1936). https://doi.org/10.1038/137861c0
Issue Date:
DOI: https://doi.org/10.1038/137861c0