|
|
IN PROCESS TESTING FOR PM MOTORS James A. Fisher Fisher Data Products, Inc.
Kettering, OH 45429
Motor manufacturers are increasingly faced with the battle between improved product performance and lower manufacturing costs. As with any production process, a balance is determined to closely satisfy both requirements based on available equipment technology and currently accepted methods. The process outlined offers solutions for both improved quality and lower costs by eliminating the need to couple the motor shaft to obtain performance data.
It has long been a practice in the manufacture of electrical motors to test and publish specific performance characteristics for the motor designs so that the motors can be used most effectively and in appropriate applications. Two relationships that provide useful performance characteristics are motor speed versus torque, and motor current versus torque.
In conventional testing arrangements, the shaft of a motor to be tested is coupled to a secondary device during measurements of motor characteristics including speed, current draw and torque. Additional devices, such as torque transducers and brakes, are also often used for the testing. Unfortunately, the use of couplings can lead to problems including shaft binding, misalignment, and of course the added time and work required to connect and disconnect the coupling from the motor shaft. Such problems with couplers and the devices connected to a motor shaft for testing can affect the test results and can also cause heat build up within the motor, particularly within the armature. Such heat can be sufficient to skew the results of the test. These problems cannot only affect the test results, but also can prevent tests from being repeated consistently.
The method described herein recognizes that testing a motor without the use of a coupling device and other secondary components can result in more reliable motor performance measurements, notably stall current, stall torque, vibration measurements, and motor noise measurements. Further, the complexity level of fixtures required for testing can be greatly simplified, thus resulting in a reduction of cost and testing time.
Traditional Uncoupled Testing of PM Motor - All permanent magnet motors exhibit a nearly straight line (or linear) performance curve when run in an uncoupled test condition.
As with any linear equation, if two points are known, the function can readily be determined. In the case of a PM motor curve, the no load (zero torque point) is easily obtainable thru simple instrumentation such as a D.C. current meter and an external tachometer.
A second point can be determined for torque by using a known law of physics:
T =
SJ Where T = Torque
t
S = Change in speed (Sf - Si)
J = Moment of Inertia
t = Time Duration (tf- ti) (Eq. 1)
The midpoint torque may be computed by measuring the amount of time it takes the motor to reach maximum speed (SNL) from zero speed. In this case
S = SNL. The midpoint torque is the computed torque at half speed, (half of SNL).
Once the midpoint torque and no load speeds are determined, any other point may be computed on the curve thru interpolation for points above the half speed or extrapolation for points below half speed.
A second current point may be determined thru the use of a high-speed peak detector circuit monitoring motor current at the instant that power is applied. The highest current is always at zero speed (stall).
After determining the no load and stall currents, any other current point may be interpolated between the two known points.
Although this fundamental method has several advantages over other technologies such as dynamometers and inertial loading with encoders, the high torques and relatively low inertia associated with most PM motors result in extremely low acceleration time which may cause inaccuracy with peak current and time to full speed measurements.
Reverse Inertial Testing - The reverse inertial testing an improved method for uncoupled PM motor testing is to determine stall torque by measuring the time required to brake the rotating armature until it reaches zero speed. This is accomplished by applying reverse motor voltage to the leads with respect to the rotational direction that the shaft is turning.
The same law of physics applies to this approach:
T =
SJ
t (Eq. 2)
In this case
S is still a positive value if the initial speed is considered negative and the final speed is zero. Now T is the actual stall torque rather than an average or midpoint value.
Figure 1 demonstrates the theoretical torque vs. speed curve for a motor that is reversed and allowed to run thru zero speed on up to full no load speed.
Figure 1 - Theoretical Current vs. Time
Figure 2 demonstrates the theoretical current vs. time plot for the a motor which is run in reverse until the applied voltage is reversed (at ti) and then allowed to run thru zero speed (at tf) and then up to full of load speed.
Figure 2 - Theoretical Torque vs. Speed
Determining tf, - Of particular interest is the current change at tf. It is this point where current begins to decrease because the motor is now accelerating in the forward direction. By monitoring the current and utilizing a simple algorithm, tf is determined with reasonable accuracy.
In determining ti, the value of ti is the exact point in time where the motor voltage was reversed. Using solid state switching components, which provide almost instant state change assures that ti is very accurate.
In determining
S, the initial (reversed) speed of the motor may be determined by off the shelf proximity or optical detectors aimed at the motor shaft. To simplify test fixturing even further, a high resolution snap shot of the current over a period of time, analyzed via FFT software can also determine the speed of the armature. Generally, the highest amplitude in the spectrum is at the frequency that the commutator bars make and break electrically with the brushes.
In determining J, the moment of inertia of the armature may be determined in a lab or closely calculated using the equation:
J = MR2
2 (Eq. 3)
This step need only be performed once for each armature type to be tested. There may be small variations of J for each part tested due to wire overlay patterns, however, these differences will have minimal effect on accuracy.
If actual current vs. time were plotted for a PM motor, which has been reversed, a slight slope would appear between ti and tf due to the back EMF of the motor which is opposite in polarity to the applied voltage. As the motor slows to a stop, the back EMF falls proportionally. Additional software algorithms are used to compensate for the back EMF component of the waveform.
The ripple on the top portion of the waveform may be used as a measure of commutation quality and / or fusing quality. A poorly fused tang, or an improperly installed brush spring, will result in high ripple content. These problems can easily be detected in the motor while turning at high current.
The advantages are:
1.Simplified Test Fixturing - A hold down mechanism and lead hookup terminals are all that is required.
2. No Torque Calibration - Using the laws of physics and the precision crystal clocks found in all computers, eliminates the need for external calibration.
3. Fast Cycle Time - The total test time is limited only by the acceleration times (forward and reverse) of the motor.
4. Commutation and Fusing Quality - This analysis may be incorporated without sacrificing cycle time.
5. Vibration and Noise Analysis - These tests may be performed in the same station after the motor settles out at no load speed. The elimination of shaft coupling allows for this option.
In conclusion, motor performance testing is a critical part of the manufacturing process. Due to the simplicity and speed of this method, the implementation of several types of performance analysis may be combined into one test station without sacrificing accuracy. Benefits include reductions in labor, test time, equipment costs and maintenance expense. As a result, the motor manufacturer adds quality to their product while reducing costs.
