Mike Howell
EASA Technical Support Specialist
The two primary reasons for performing stator core testing in the service center are (1) to verify that the stator core is acceptable for continued use and in the event of a rewind, and (2) to verify that the repair process has not adversely changed the stator core condition.
The purpose of this article is to discuss how we determine, assess and compare stator core test results. It is extremely important to understand that variance in test procedures may invalidate comparison.
Stator core testing
Service centers typically perform a loop test (also known as a ring test or core flux test) to assess the condition of the stator core. This can be done using a commercial core loss tester or a conventional manual loop test using an appropriate AC source, cables and meters. While the loop test does not provide core loss measurements that could be used for calculation of machine efficiency, it is a good tool for assessing the relative condition of the stator core. And, although other test methods for assessing stator cores do exist, this is the most common method used by service centers.
Many loop tests performed in the service center are done with no knowledge of the stator winding data. To facilitate testing many different sizes and types of machines, in different states of operability, it is logical to choose a baseline test level that provides reasonable excitation for assessing the condition of the stator core. Most of these tests are done with a target back iron magnetic flux density of 1.32 T (85,000 lines/in2).
Test procedure
If the service center is not using a commercial core loss tester, a suitable procedure for performing the loop test can be found in the EASA Technical Manual [1]. As an alternative to manual calculations, EASA has posted a simple loop test calculator on our website that is easy to use and smart phone / tablet friendly [2]. Technicians may find the ability to quickly calculate test parameters and results without leaving the test floor to be convenient and efficient. The
interface for the tool on EASA’s website is shown in Figure 1.
It is recommended to verify the actual back iron flux density developed utilizing a one-turn search coil as shown in Figure 2. Commercial core loss testers are typically equipped for this and perform the calculations automatically. The basis for this calculation is to eliminate error due to the voltage drop of the excitation coil caused by its resistance and current.
Depending on the specific test setup, this voltage drop may or may not be significant. The voltage measured at the search coil terminals is determined by the power frequency and the magnetic flux density in the back iron. So, for any measured voltage, we can easily calculate the back iron flux density.
If the power supply is capable of adjustment in small voltage increments, the operator should dial in the calculated induced voltage and therefore, the correct back iron flux density. If this technique is used, the input voltage and current should be monitored to make sure everything is approximately as expected.
Assessment of core iron
The EASA Technical Manual procedure for stator core testing recommends three criteria for assessing the core iron.
- Minimal core temperature rise.
- Even core heating (no hot spots).
- Acceptable core loss.
For criterion 1, minimal core temperature rise, the procedure states that surface temperatures typically rise 5-10°C (10-20°F) in about 30 minutes. It is noted that this varies with the size of the motor. Cores should be repaired or replaced if the temperature becomes very high or rises very rapidly. Cores are deemed acceptable if they have a temperature rise of less than 15°C (27°F) above ambient temperature.
For criterion 2, even core heating, the procedure does not include a numerical definition of hot spot. It is reasonable to define any area with a temperature of 10°C (18°F) above the average core temperature to be a hot spot. If hot spots are present, the core should be repaired or replaced.
For criterion 3, acceptable core loss, the procedure states that the core loss should compare favorably with published data or measurements taken on similar cores. A more practical included note states that good cores will typically have between 1 to 5 watts per pound (2-11 watts per kilogram) depending on lamination material grade, gage and processing. This core loss value is based on watts per pound (or kilogram) of back iron. Additionally, in the case of a rewind, it is specified that the core loss should not increase by more than 20% after the burn-off process. Invariably, it is criterion 3 that gets the most attention from industry groups, service centers and customers so it is of interest for further discussion.
What watts, what pounds?
Calculating watts per pound (or kg) is a very straight-forward process by simple division with watts being the dividend (numerator) and pounds being the divisor (denominator). The more complicated issue at hand is deciding what watts and what pounds to use.
Watts
First, let’s look at the watts or real power component. A simple model of the test circuit is shown in Figure 3 and the current I
1 = V
1 / Z where the impedance (Z) will be R
1 + jX
1 + Rc // jX
m. So, the current will have a real component and a reactive component. There are two resistive components shown in Figure 3 that contribute to the real power loss in watts, the cable resistance (R
1) and the core loss resistance (R
c). We are interested in the real power due to core loss (P
c) and the easiest way to separate them is to subtract the cable loss from the total measured input power (P
t).

It is apparent from the equation above that the cable resistance (R1) can have a significant impact on the watts loss used in the watts per pound calculation.
Let’s look at an example. The measurements required in Figure 2 are given in inches as follows.
L=1.400 D1=3.750 S=0.710 B=0.535
The power supply for this test is a 0-130V variable autotransformer rated for 20 A. Using the calculator, it is determined that 1.78V and an 11-turn coil should provide the desired back iron flux density with an estimated current of 14.7 A. The target induced voltage for a 1 turn search coil is 0.16 V to achieve a back iron flux density of 85,000 lines/in2. The actual test results were as follows.
Pt = 14 W I1 = 12.7 A Bc = 84,472 lines/in2
The cable I2R loss was 5.3 W leaving 8.7 W for the core loss. If the cable resistance is ignored in this test, the watts per pound would be 14/8.7 = 61% higher than actual. In most cases, it probably will be beneficial to separate the two components. Manufacturers of commercial core testers can account for cable loss automatically in their core loss calculations.
Pounds
When performing core testing, the watts loss is divided by the weight (mass) of the core to normalize the data. This way, the same pass/fail criteria can be used regardless of size. The units of measure (typically pounds or kilograms) can be whatever is convenient for the service center, providing the pass/fail criteria are adjusted accordingly. As previously mentioned, the procedure given in the EASA Technical Manual states that good cores will typically have between 1 to 5 watts per pound (2-11 watts per kilogram) depending on lamination material grade, gage and processing.
The
EASA Technical Manual procedure calculates the weight of the stator back iron only as opposed to entire stator core. The logic behind this is reasonable in that due to the orientation of the test coil, the magnetic flux path will only be through the stator back iron as shown in Figure 4. Although there may be some fringing into the root of the tooth, it will be insignificant.
However, at least one commercial core tester manufacturer has used the weight of the entire core when calculating watts per pound. This will of course make a significant difference in the test result with the magnitude of the difference dependent on the design of the stator in question. Let’s return to the test example to investigate this.
The calculated weight of the stator back iron is 3.5 lbs. (1.59 kg). If the teeth are included (tooth width = 0.174 inches, 4.42 mm) in the calculation, the entire core weight is 5.2 lbs. (2.36 kg). This is roughly a 50% increase in weight and the watts per pound calculated would be 1/1.5 = 67% of the value calculated with the back iron weight only.
So, if we have established that the magnetic flux is only traveling through the stator back iron, why would it make sense to include the teeth when calculating watts per pound? The logic behind this approach is also reasonable in that even though the magnetic flux path will only be through the stator back iron, eddy currents will follow a short-circuited path between laminations at any point in the stator, regardless of the magnetic flux path [3]. This is illustrated in Figure 5 where the small black dots represent shorts.
In Figure 5 (1), there are two short points between laminations creating a closed loop for eddy currents to flow. The short at the bottom could be caused by a defect, a weldment or a low resistance contact with the frame. In Figure 5 (2), there is no closed loop, but all insulation is imperfect and there would still be a path for reduced eddy currents to flow. In both cases the short locations would overheat. So, one could certainly argue that if shorts in the teeth will be identified during the test and if the current through them contributes to the watts loss, then it is logical to include the weight of the teeth in the calculation.
Watts per pound
Now, back to the simple watts per pound calculation. We simply need to take our watts value and divide it by our pounds value and compare the result to our acceptance criteria to determine if that criterion is acceptable.
But, we’ve discussed two methods of arriving at a watts value and two methods of arriving at a pounds value giving us a total of four different possible results for the same test. These data are provided in Table 1 for comparison.
As of the January 2017 revision of Section 7 (Electrical Testing) in the EASA Technical Manual, the option from Table 1 consistent with the stator core testing procedure would be teeth out / cable loss in = 4.0 W / lb. This approach is the most conservative and requires the least calculations. While oversizing the excitation cable will reduce its losses, subtracting the cable losses from the input power gives the most accurate results.
Having a solid understanding of how these choices affect results should aid service center personnel and end users in working through issues, especially where multiple test methods or test apparatus have been used on the same stator.
In the case of rewinds, it is worthwhile to reiterate the importance of following the same test procedure before and after the burnout process. It should be apparent that practices such as switching from loop test to commercial core loss test, using two different commercial core loss testers or even using a different test cable for the loop test could significantly affect the results.
References
[1] EASA, "Stator Core Testing," in EASA Technical Manual, St. Louis, Electrical Apparatus Service Association, 2017, pp. 7-17 to 7-23.
[2] Electrical Apparatus Service Association, Inc., "Loop test calculator", Available: https://easa.com/resources/software/loop-test-calculator.
[3] R. Nailen, "What Core Loss Testers Can Do," Electrical Apparatus, pp. 43-46, February 1986.
AVAILABLE IN SPANISH
ANSI/EASA AR100
More information on this topic can be found in ANSI/EASA AR100- Section 2: Mechanical repair
- Section 3: Rewinding
EASA Technical Manual
More information on this topic can be found in EASA's Technical Manual- Section 2: AC Machines
- Section 7: Electrical Testing
Print