Winding temperature is an important concern when evaluating any motor design or drive application. Although the nameplate indicates the intended temperature limit, actual operating temperature is normally lower. How much lower, and therefore how much thermal margin may exist for overloading, can be determined only by test.

Thermometer and thermocouple readings were once the standard measure of winding heating. Today, the change-of-resistance method is the basis for motor ratings. The IEEE 112 test standard allows a delay between motor shutdown and the first resistance reading, because of the time required to connect a resistance bridge to the safely deenergized winding. That can give a misleading result, depending upon whether winding temperature falls or briefly rises immediately following shutdown.

Also, bridge readings themselves are subject to error. The calculated temperature change involves two such readings. plus the measured ambient temperature. A 1965 study showed that the total error in winding rise could be plus or minus 9 percent.

Regardless of test precision, temperatures may differ within a group of supposedly identical motors. These variables are responsible:

1. Total heat-producing losses within stator and rotor.
2. Heat transfer from those sources to the external cooling medium.
3. Ambient temperature. (Winding heating will be greater in a standard 40 degree ambient than in a 20 or 25 degree ambient, because resistance will be higher.)

Losses are subject to many sources of variation, such as tolerances on wire dimensions (affecting resistance); on lamination steel metallurgy; and on the air gap. Typically, those can result in winding temperature rise variation of at least 10 percent. Heat transfer (particularly from stator core to its enclosing frame) can add another 5 degrees C variation.

Although all those effects seldom occur to the same extent in any one machine, measured temperature should never be considered accurate within 1 to 5 degrees C. Tests on identical motors, built at the same time in the same factory, have shown differences as great as 20 to 25 percent. Ten percent is not uncommon. Caution is therefore warranted when using any one test as the basis for judging a motor's thermal margin.

From  "How Consistent are Motor Temperatures?" ...by Richard L. Nailen, EA Engineering Editor  - 
Complete article, with graphs and photographs, published in Electrical Apparatus December 2004  © 2004 Barks Publications, Inc.  All Rights Reserved.
Copies of this issue are available from our online webstore