Chuck Yung
EASA Senior Technical Support Specialist
There are times when a DC motor or generator experiences a catastrophic failure and the customer wants to know why it happened. One type of failure that seems to stimulate lively conversation is when the failure involves dramatic damage to the brushholders and commutator. The term “flashover” describes the appearance of the failure; the very name conveys an accurate mental image of the failure. See Figure 1.
The questions that arise next are predictable: “What caused this?” and ”What can be done to prevent a recurrence?” Or, if the motor was recently repaired: “What did you do to my motor to cause this?!” The purpose of this article is to help you answer those questions.
The causes of a flashover can be partially explained by the insulating properties of air, and Ohm’s Law. Air is an electrical insulator, although the dielectric breakdown voltage of air is low compared to the insulating materials we use in electric motors. Inside an operating DC motor, we find heat, carbon dust and other contaminants, and perhaps even humidity. Each of these will reduce the dielectric strength of air.
As for Ohm’s Law, E/R = I; winders use this frequently to evaluate shunt fields and to extrapolate the temperature rise of those fields. But it also applies to the armature circuit.
At the moment a DC motor is energized, before the armature starts to rotate, the armature current is limited only by the available kVA of the power supply.
Consider the example of a 500 hp motor, with a 500V armature circuit. Static resistance of the armature-interpole circuit measured only 0.02 ohms, so the short circuit armature current could reach 25,000 amps if the drive has sufficient kVA: 500/0.02 = 25,000 amps.
Effects on armature
Fortunately, drives ramp up the armature voltage, rather than applying it instantly. As soon as the armature begins to rotate, the inductance provided by the armature becomes a factor in suppressing the armature current. Paraphrasing the now-defunct IEEE Standard 66: When voltage E is applied across a circuit consisting of a resistance and inductance L in series, the maximum rate of rise is given by the equation di/dt = E/L amperes per second; where E equals volts, and L equals henrys. In other words, the armature current decreases rapidly as the armature speed increases.
Every DC motor can be used as a generator, by driving it mechanically and applying current to the fields. When operating as a motor, there are times where the motor might be driven by an overhauling load (e.g., a loaded conveyor running downhill; or a hoist lowering a heavy load). When that happens, the counter-emf (electro-motive force) produced overcomes the applied emf, and flashover is likely. In layman’s terms, operating conditions cause the armature current to increase rapidly, and generated voltage/current trigger the flashover.
A list of operating events that can cause a flashover is included in Table 1.
If the interpoles are not correctly adjusted to maintain brush neutral throughout the operating load range, the shifting neutral results in arcing as the load increases outside the black band region. That can, in and of itself, trigger a flashover. (The black band region can be described as this: Weakening / strengthening the interpoles, independent of all else, until the brushes begin to spark produces a band within which no sparking occurs. That band is referred to as the “black band.” For more information, see the Assembly and Final Test section of Fundamentals of DC Operation and Repair Tips.)
Preventive measures
Working to help your customer understand the basics of how a DC motor operates can go a long way towards helping them avoid problems. One of my most vivid “triggers” of a flashover is the customer who installs a newly rebuilt compound motor with more than 50% compounding. (The percent compounding describes the percentage of total field flux contributed by the series fields, at full load.) They check rotation and discover that the motor needs to be reversed. We all know that the correct way to do this is to swap the A1 and A2 leads (the large wires that are thoroughly taped). But, says the customer, it is so much easier to swap the shunt field leads (they are smaller, and probably held in a terminal strip by screws) instead. That shortcut has worked in the past — on straight shunt motors.
With a compound-wound machine, this time-saving shortcut changed the motor from a cumulative connection to differential. The motor runs fine unloaded, and even with a moderate load. But when the load is increased to the point that the series overpowers the shunt fields, catastrophe occurs. Since this is a newly rebuilt motor, there is a very good chance that your customer will blame you. After all, you just rebuilt the motor. So it is important to educate the customer to avoid just such a situation. (And yes, I have had many, many calls where a newly installed motor failed exactly as just described.)
If someone blames a flashover on “drive settings,” that implies that the drive is accelerating or decelerating the motor too rapidly. If so, a competent drive technician should be able to adjust that to reduce the chance of flashover. Blaming the drive may instead mean that the motor is in an application calling for a regenerative drive, but the customer replaced the drive with a less expensive model that cannot handle the regenerative mode. (And the customer might not admit having done so until you press the issue.) One example would be a compound wound motor driving a roller coaster. When the cars are coasting downhill, the regenerative mode is used to prevent dangerous over-acceleration.
A compound wound motor, in such an application, requires a drive that has connection points for the shunt, armature and separate series field leads. This is to permit the motor to operate with a cumulative connection in both directions of rotation. If a compound wound motor is operated from a drive with only shunt- and armature circuit leads, in a reversing application, it will be cumulative in one direction but differentially compounded in the opposite direction. The higher the percent compounding, the greater the risk of speed instability and/or flashover. See Table 2.
Specific to any DC motor, there are several preventive measures to reduce the chance of a flashover. The first of these is to simply chamfer the end of the commutator bars. Voltage stress varies exponentially to the inverse of the radius. Chamfering the customary square corner at the end of the commutator to a 1/16” (1.6 mm) radius reduces the voltage stress to approximately 15%, significantly reducing the opportunity for flashover to occur. See Figure 2.
Add flashover protection
If a customer has chronic issues with flashover, take a lesson from the traction motor industry and add flashover protection. Install four equally spaced short lengths of angle iron in line with the end of the string band area. The bolted connection must be electrically sound and the edge closest to the commutator must be bare metal (no paint or other coating). The bare metal provides a reliable path to ground, if an arc is to occur, thus minimizing damage to the costly brush boxes and commutator. See Figure 3.
Flashover detection is commercially available and reliable. It has long been known that, at the moment a flashover begins, the field polarity reverses. Automated instrumentation, by monitoring the polarity of the field current, can shut the motor down before the fault current causes damage.
If the application is a fan, blower or downhill conveyor, where the motor might start while the load is free-wheeling in reverse, the solution could be a brake – either mechanical or otherwise, interlocked with the drive to release the brake when the motor starts. One option the end user might consider is to use the shunt fields as the dynamic brake. If they do so, the field current should not exceed 1/3 of the rated shunt field current. Otherwise, the shunt fields might overheat and fail prematurely.
The manufacturer has more latitude than we do as repairers, so it is common to see larger machines designed with a compensating winding (a.k.a. “pole face bars”), imbedded in the face of each field pole to effectively extend the influence of the interpoles. Those compensating windings, just like interpoles, must be connected correctly so as to yield the correct interpole strength. Misconnected interpoles or compensating windings (i.e., the wrong number of circuits) radically change performance and are much more likely to spark and/or flashover.
Print