Chuck Yung
EASA Technical Support Specialist
While shaft currents are not a new problem (papers on the subject date back prior to 1930), what is “new” is our understanding of how to solve the problem. Shaft currents have been described as shaft voltages, circulating currents, bearing currents and circulating voltages. This article will refer to the phenomenon as “shaft currents” because it is the current that causes the damage.
When a conductor is passed through a magnetic field, voltage is induced into the conductor.
It is not the voltage that damages a bearing, but rather the current. (Fuses fail because the current is too high, not the voltage.) We don’t have a practical way to measure the current through the shaft, so we measure the magnitude of the voltage instead.
We used to blame magnetic dissymmetry for shaft currents, and rightfully so. Magnetic dissymmetry just means gaps in the iron, such as segmented laminations used to build stator cores over approximately 35” (900 mm) diameter, uneven airgap, circulating currents in the parallel circuits of a 3-phase winding, or variations between bolt-in DC poles.
Since the electromagnetic field in the stator rotates around the stator bore, those dissymmetries are one source for induced voltage in the frame. Through-bolts in rotors were another cause—but more about them later.
Correcting The Problem
When the problem was truly a circulating current, it “circulated” from the frame through a bearing, along the shaft, through the other bearing and back to the frame. To correct the problem, we had only to break the circuit. Insulating the Opposite Drive End (ODE) bearing was the most practical solution. Just as turning off a light switch stops the current flow through the light bulb, insulating the ODE bearing interrupts the circuit through the bearings as indicated in Figure 1.
Another solution was to install a shaft grounding brush in parallel with the bearing as shown in Figure 2. That did not stop the current, but the brush diverted current from the bearing (Figure 3).
Parallel paths act as current dividers, with the amount of current through each path determined by the relative resistance of the paths. The lower the resistance of the brush-shaft interface compared to the resistance through the bearing, the more current was diverted from the ODE bearing. But the grounding brush did not reduce the current through the drive end (DE) bearing.
Since the DE bearing was normally the larger bearing, the grounding brush did reduce bearing failures on the ODE—at least those caused by shaft currents.
Note that the brush resistance is critical to this solution. Not any carbon brush will do. The special grounding brushes supplied for this purpose are extremely low resistance. The key to success is to provide a much lower resistance path than the bearing to divert (most of) the current. If the resistance through the brush/shaft path was equal to the bearing/shaft path, we could expect half the total current to pass through the bearing. Bearing life would be extended, but the problem was not solved.
That points out the second problem with the grounding brush solution: If the shaft was dirty, or corroded, painted, etc., the resistance drop across the brush/shaft interface would increase. And that would divert more current back through the bearing.
Even though it is current that damages the bearing, manufacturers do not publish current-carrying capacities for bearings. And since we cannot practically measure the current passing through the shaft, we rely on measuring shaft voltages. The “rule of thumb” for many manufacturers was 100 mV for ball bearings and 200 mV for sleeve bearings. (NEMA MG-1 part 31.4.4.3 suggests a limit of 300 mV measured end-to-end on the shaft.) With shaft currents caused by magnetic dissymmetry, it was rare to measure shaft voltages over a few volts.
Variable Frequency Drives
Enter the pulse-width modulated variable-frequency drive (PWM VFD). When VFDs started to become popular, one of the major issues was shaft currents. For quite a while, we treated the problem in the usual manner: Insulate the ODE bearing, or add a grounding brush. The problem was that each of these “tried and true” solutions yielded mixed results. Users disagreed on which solution was best, and many reported less-than-satisfactory results with either method. A “belt and suspenders” approach seemed to work best, but we weren’t certain why. Longer cable runs seemed to make the problem worse, as did poor grounding connections, and certain drives developed bad reputations as “motor killers.” Higher switching frequencies (20 kHz) cause more bearing problems than slower (5 kHz) drive settings, but there is no clear line above which we expect trouble.
The VFD works by rectifying AC to DC, and chopping the DC into positive and negative pulses to simulate an AC sine wave. Varying the DC pulse width simulates a variable AC sine wave and changes the frequency, thereby changing the motor speed. One problem with that is the common mode voltage. When a 3-phase motor operates from a true sine wave, the common mode voltage was always zero (Figure 4). But with the VFD, that balance no longer exists. DC is either positive or negative, so at any point in time the three phases are either + + - or + - -. Common mode voltage is essentially line voltage. Shaft currents became a significant problem for motors operating from a VFD, even in motors much smaller than had previously experienced trouble.
Figure 4: With 3 phases 120 degrees apart, draw a vertical line (that’s a point in time) at any point on the graph. Sum the voltages (above the horizontal axis is positive, below it negative), and the common mode voltage is always zero.
Other Factors
Back when most machinery was driven by flat leather belts, the buildup of static electricity could result in electrostatic discharges which could damage the bearings. Flat belts are no longer common, but static friction is still a problem for belts, paper roll winders, and similar applications.
These static discharges may have prompted the earliest use of shaft grounding brushes.
Capacitive coupling between the rotating field in the stator and the spinning induction rotor also seemed to cause shaft currents. Insulating one bearing did not cure it because the problem was no longer a circulating current. If both bearings are insulated, the insulation acts as a capacitance. The induced voltage could continue to build, possibly to dangerous levels, until the “capacitor” discharged across the insulation. In certain applications, such as hazardous locations, the capacitive discharge could become a source of ignition.
Thanks to discussions with engineers from several manufacturers, we have a better picture of the causes of, and solutions for, shaft currents.
The higher the switching frequency of the PWM drive, the more likely there will be bearing damage from shaft currents. Common grounding of the motor and drive is critical. Even if a motor and the drive case are each grounded per electrical codes, it is possible to measure up to about 30 volts potential between the two ground connections. The solution is a dedicated common ground from the motor to the drive. Stranded ground cable should be used; skin effect is a factor. (Current “travels” on the surface of the conductor, so the resistance-to-voltage flow is affected by the surface area of the conductor. Stranded cable has more surface area than a solid conductor.)
Voltage Magnitude
Compare those previously stated relatively low voltages to these: Technical papers on the subject report VFD-related shaft voltages of 25 volts. EASA’s Technical Support Department has had reliable reports of shaft voltages of 40+ volts when rebuilt armatures or rotors were returned with the combination of through-bolts and skewed slots.
It is generally accepted that we can measure the shaft voltage, while it is the current that is harmful. Shaft voltage can be measured between the ends of the motor shaft. It can also be measured from the end bracket to the shaft. Use a brass or copper brush attached to the lead that will contact the shaft. That minimizes false readings that can result from the friction between the test lead and shaft.
One old-timers’ technique is to use a length of welding cable touched to both ends of the shaft. Observe the contact region for arcing and use a clamp-on ammeter to check for current. The problem is that some of the current still passes through the bearings, so the welding cable—at best—is a parallel path. We cannot tell from this simple test how much current the bearing still carries, but if the diverted current is enough to strike an arc when the lead wire is touched to the shaft, that is more than enough current to damage a bearing.
How much current will damage a bearing, and how much of the current is still passing through the bearing compared to the welding cable, are matters for conjecture. About the only way to measure the actual shaft current is to place a current transformer (CT) or ammeter around the shaft between the bearings. That is rarely practical.
Conclusion
It is clear that there are two mechanisms for shaft voltages: circulating current which can be interrupted by insulating a bearing, and capacitive coupling which requires insulation of both bearings plus the use of a grounding brush. It is also important to provide a common ground between the motor and the drive case.
Bearings are apt to be damaged even by very low current passing through them. And with the proliferation of VFDs, grounding problems on one motor can cause shaft currents in other nearby equipment.
Until a better solution is developed, motors operating from VFDs will benefit from the practice of insulating both bearing housings and installation of a shaft grounding brush.
AVAILABLE IN SPANISH
Print