2011 Outage Handbook – Generator Exciters

How to prevent collector-ring fires

Flashovers in generator exciters have gotten considerable attention in the last couple of years as powerplant owner/operators have stepped up efforts to make their facilities safer workplaces. Today, the industry generally recognizes that collector-ring fires can be prevented by closer observation of exciter operation on rounds and proactive maintenance.

But with few, if any, O&M personnel at a given plant having more than very general knowledge of generators, close visual inspection and online maintenance often are intentionally avoided. You really can’t blame someone for being uncomfortable around a rotating shaft and electrical sparking.

Part of the problem is the inaccessibility of some exciter brush gear: Doghouse structures can be too “tight-fitting” to permit the checking of all brushes, holders, etc, without getting what many roving operators consider “too close for comfort.” Also contributing to the reluctance of staff to maintain brush gear is the unfriendly design of some equipment in service.

One O&M manager responsible for two 2 × 1 F-class combined cycles installed during the bubble told the editors during a recent plant visit that his units run base-load and equipment reliability is very important. He admitted personally feeling ill at ease performing relatively minor, but necessary, brush and brush-holder maintenance online. His solution was to retrofit all six generators in the plant with third-party brush gear to facilitate online inspection and maintenance. Series of videos at http://www.ccj-online.com/exciter-brush-maintenance/ shows how simple this can be to do.

Collector-ring concerns were front-page news at the Turbine Generator Users Group (TGUG) meeting, held in Houston, last January. A workshop at that event was dedicated to generator-rotor, collector-ring, brush, and rotating-exciter condition assessment.

David Cutsforth, who has more than four decades of experience with brushes, brush holders, and collector rings was one of the featured speakers. The founder of Cutsforth Inc, Bloomington, Minn, authored EPRI Guideline 8.13, covering brush, collector ring, brush rigging, and excitation maintenance.

What follows is a summary of Cutsforth’s remarks on the importance of a proper root cause analysis (RCA) following any collector-ring incident, plus suggestions he had for maintaining collector rings and brush apparatus in top condition.

Cutsforth began with a review of excitation-system components and their functions. If you need a refresher, access the 7F Users Group report in the 3Q/2009 issue of the CCJ at www.ccj-online.com and scroll to the sidebar “Is your exciter brush system meeting expectations?”

The root cause of a collector-ring fire usually is not properly understood and rarely diagnosed correctly (Fig 1), the speaker said. In many instances where collector-ring fires have occurred, damage is extensive and visual assessments are difficult to make. Investigation into root causes, he continued, often leads to erroneous conclusions because the timing sequence is not understood or considered. This suggests another corrective actions were not taken to prevent it.

Cutsforth said every collector-ring fire that has occurred in the industry was started by arcing associated with brushes losing their proper electrical connection with the ring. During a fire, if air is deflected by excitation system components to the arc location, the additional oxygen creates a “blowtorch effect” which causes metal to melt and insulation to burn in that area (Fig 2).

A physical characteristic of the torch effect is a molten or “flowing” surface on the metal. The metal flows because burning takes place over a period of time; it is not an explosion. This effect can mislead those determining the root cause, because the area of flowing surface appears to be the hottest spot and the assumption is that the hottest spot is the source of the ring fire.

Cutsforth pointed out that while investigating one recent fire, the initial point of failure was thought to be the terminal leads (Fig 3). But on closer examination it became obvious that the source started with the loss of connection between the brush and collector ring of multiple brushes. When sufficient arcing is present, a circle of fire forms around the collector ring.

The torch effect also can be misleading because the temperatures reached during a collector-ring fire can be high enough to melt entire brush boxes (Fig 4).

In some ring fires, a secondary flashover—one caused by the ionizing of air from the arcing or insulation failure—can occur. This is an instantaneous explosive-type action with a phase-to-phase flashover or second ground being developed. The first ground won’t create a flashover. The appearance of the damaged area from a flashover of the first type looks more like a “gouge” than flowing metal.

When both flowing metal and an explosion characterize the damage, the ring fire with the flowing appearance must be the cause because it takes time to develop. The explosive damage is instantaneous and cannot be followed by a ring fire.

Causes of collector-ring fires

Before discussing the causes of collector-ring fires related to system components, Cutsforth noted two problems attributed to poor maintenance practices. The first maintenance error, he said, is letting brushes get so short they no longer have sufficient spring force to maintain the brush-to-ring connection.

The second error is not positioning the brush leads in a manner that allows them to avoid contact with the top edge of the brush box, which can restrict brush movement. Leads also should be kept away from the constant-force springs to prevent them from inadvertently cutting off brush leads.

The brush holder, Cutsforth said, is the excitation system component most often associated with collector-ring fires.

No. 1 cause of ring fires: carbon deposits in the brush holder. Deposits form over time and eventually prevent the brush from maintaining contact with the collector ring. Occasionally,they can even lock the brush in the holder (Fig 5).

Yet carbon deposits often are overlooked as the cause of fires either because they are considered inconsequential or the brush still can be moved quite easily in the holder. Not true. Side pressure from ring rotation and from the constant-force spring restrict brush movement and cause low spots to develop in the ring surface by electrical erosion from arcing under the brush. Arcing, of course,results from the poor connection caused by the brush restriction.Perhaps the best way to easily verify the existence of brush-movement restrictions from carbon deposits is to inspect the sides of the brush. Carbon deposit marks are very evident in Fig6, but the user was distracted from the obvious by deceiving hot colors in the thermal image photos in Fig 7.

The incorrect assumption: High temperature was causing low spots to develop in the rings, thereby causing them to go out of round. But there was no discoloration of the brush lead in this case, evidence that would have suggested carbon deposits might not have been the primary cause of the ring fire.

Carbon deposits were the cause and the RCA revealed that the deposits formed on new bronze holders even though they had been plated. To avoid problems in the future, investigators suggested truing of the rings at nine-month intervals.

No. 2 cause of ring fires, brush binding, typically is caused by use of brush holders with constant-pressure springs that are shorter than the brush. The springs exert side or lateral forces on the brush which cause the brush to bind on the top edge of the brush box opposite the spring-clip side (Fig 8).

Cutsforth pointed out that carbon deposits form at pressure points so binding of the brush in the holder increases the possibility of carbon deposits, which would increase the risk of the brush losing contact with the collector ring.Some holders are designed with a plastic guide that slides on a steel plate to control spring side pressure. This works until dirt gets embedded in the plastic, creating yet another spring-pressure restriction.

No. 3 cause of ring fires: low spring pressure. As the spring ages, it loses tension and prevents the brush from making good contact with the collector ring. Springs also can weaken over time from conducting electricity. There also have been instances of cracking and of damage during maintenance (Fig 9).

No. 4 cause of ring fires, poor terminal connection, causes a brush to drop out of the circuit, thereby reducing the number of brushes carrying current. Look for a discolored lead (shunt) at the terminal end; it indicates that a high-resistance terminal connection has created enough heat to discolor the lead (Fig 10).

A poor terminal connection in bronze or brass holders also can permit current flow through the brush box instead of the brush lead. Result: Electrical erosion of the brush box from arcing. The rough surface created restricts brush movement thereby compromising contact between the brush and slip ring (Fig 11).

Contamination. Another contributor to collector-ring fires is air contamination. Dirt and oil in excess can cause the brush-to-ring contact to degrade as contaminates collect under the brush surface. This then leads to arcing and the progression toward a ring fire begins. Dropped foreign objects (tools, for example) are also capable of causing ring fires.


Selective action. A characteristic of carbon is that its electrical resistivity decreases as its temperature increases. Most materials, including copper, have the opposite behavior. This is important for understanding the term “selective action,” which describes an excessive movement of current load from one brush to another in the same circuit.

To illustrate: As one brush heats up from friction and current, the carbon becomes a better conductor. Under certain conditions this continues until the copper gets too hot and resists the current flow. At this point, a brush that may have been carrying very little current becomes the better path.Such changes in resistivity initiate a chain reaction that results in more brushes being dropped out of the circuit. Brushes begin to arc as the functioning ones become overloaded and rapidly degrade. This scenario can progress to the point where an arc is created from the ring to the holders as the air is ionized. Eventually, it can arc to ground or to the opposite polarity, causing a catastrophic explosion-type failure.

The voltage regulator senses the unstable current and reacts to the demand to maintain generator terminal voltage. In many instances, the generator trips because the maximum excitation limits are exceeded. If the condition is sustained, the damage can be severe enough to open the circuit, resulting in a loss-of-field trip.

Example: Consider a generator that has oil-contaminated cooling air coming up from the bottom against the rings. The lower brushes would pick up the contamination first and start dropping out of the circuit. Uncontaminated top brushes would carry more and more current until the capacity of one lead was exceeded.

Eventually, that lead would burn off and the next best connection would start taking the current until its lead failed, and so on. As more brushes come out of the circuit, arcing increases and the current path starts jumping from the holders directly to the ring, burning up the holders. Air blowing around in the compartment then starts burning other parts via the blowtorch effect described earlier (Fig 12).

This progression can occur over a period of months (with rapid deterioration at the end) or within a few minutes. When inspecting your generator look for indications that selective action may have occurred (Fig 13). If it has, take immediate action to reduce the risk of a collector-ring fire.

One sure way selective action will occur is if the ring wears down to the point where the helical groove disappears. Be sure your inspection plan calls for measuring and recording groove depth every time the unit is removed from service. Groove depth should be maintained at 0.035 in. (Fig 14).

Any of the causes of selective action, alone or in combination, eventually results in the loss of brush-to-collector ring contact. In turn, the brushes begin to arc, the arc causes rings to go out of round, more arcing ensues, and the progression toward a catastrophic ring fire is well under way.

Warning signs

An early sign of impending problems is the presence of visible arcing. It typically progresses slowly from difficult-to-see trailing pinpoints to significant arcing around all edges of the brush. Regular visual inspection of all brushes, if possible, is important. Remember that arcing is the indicator of a problem, not the cause.Another early warning sign is the presence of photographing (also called foot printing or ghosting) on the ring surface (Fig 15). This condition can develop on a ring that is very smooth if there are stuck or restricted brushes present. Some units are prone to photographing, which can show up within days on a freshly ground ring, and even after installing new brush holders.

Load spikes and air contamination are other causes of photographing. The important thing is to resurface (“true”) the ring online as soon as possible. Also inspect for deposits, eroded brush holders, spring problems, brush lead restrictions, and brush binding.

Discolored leads are the best visual check for the presence of selective action. As shown in Fig 14, discolored leads are easy to recognize. Taking amp readings of the individual brushes reveals the degree of selective action.Another warning sign is increased brush vibration or movement. This typically is measured with a vibration analyzer equipped with an insulated probe.Cutsforth relaxed for a bit to relate a fitting story. There was a plant in Tennessee, he said, where brush vibrations were up to 60 mils displacement and there was no visible arcing.

Also, looking at the ring with photographing or ring discoloration.

But, as the rings were “trued,” the low spots became evident. The “alarm” for this plant was noticing that the brush springs were breaking. The unit was very close to a ring fire, but there was very little visible evidence of the problem—other than brush movement.

As more and more brushes lose contact with the collector ring, Cutsforth stressed, the risk of a ring fire increases dramatically, jeopardizing operational reliability and operator safety, and ultimately causing component failure as a result of selective action.

Suggestions for preventing fires

There are two ways to prevent collector-ring fires, the speaker continued, summing up his thoughts: Keep brushes from losing contact with the collector rings, and maintain the proper depth of the spiral (helical) groove in the ring. These goals can be achieved by preventing the problems already mentioned, as summarized in the bullet points below:

  • Be proactive about removing carbon deposits. Most brush-holder designs do not permit this activity while the generator is operating. Some do, however. Consider evaluating the economic and safety benefits of a retrofit.
  • Keep in mind that carbon deposits typically redevelop very quickly—within hours in some cases—at the same spot from where they were removed. Deposit buildup can be difficult to eliminate in many instances because the underlying problem is rooted in the design of the brush holder.
  • Eliminate poor spring tension as a cause of poor brush contact by changing out springs regularly. However, the majority of OEM holders, and many aftermarket holders as well, are not designed for easy spring replacement during operation. Again, retrofit of user-friendly brush gear might be worth considering.
  • Eliminate poor terminal connections from the holder to the brush. This is not easy to do with some OEM holders while the generator is in operation. If this is the case at your plant, inspect connections thoroughly during every outage and change them frequently. In cases where the connection is not bolted, be aware that it is the spring portion of the quick-disconnect connection that fails.

You can eliminate terminal-related problems by installing a holder and brush that have the spring portion of the connection attached to the brush.

Other suggestions include the following:

  • Eliminate contaminated air by fixing oil leaks and changing air filters regularly.
  • Maintain the spiral (helical) groove at the proper depth to prevent selective action.
  • Inspect for uneven spring tension.n Change-out short brushes to avoid development of light spring tension.
  • Check for proper brush-lead positioning during every brush inspection.
  • Monitor and record brush amp readings periodically to verify proper brush-holder performance. Increase the frequency of checks if there are indications that selective action has occurred or is occurring.
  • Check amperage readings prior to changing brushes under load, especially when indications of selective action are present. Removing a brush carrying most of the load could trip the unit. The fewer the number of brushes in service, the more critical this is.
  • Take brush vibration readings quarterly at the same phase angle. Data plots will alert plant personnel when the collector ring is going out of round because of poor brush-to-ring connections that have caused electrical erosion.