Troubleshooting seals on a running pump

Leakage can occur at any time throughout the life of the mechanical seal. To troubleshoot seals effectively it is helpful to know just when the leakage starts. This is the advantage of being able to troubleshoot a running pump or one that is still hooked up to its piping. By noting the type of leakage, and when the leakage occurs we can do a more thorough job of analyzing any seal failure. In addition to leakage we will be looking for other symptoms that are visible to the trained troubleshooter.

We will start with the different types of leakage – please look at the following diagram:

The leakage sometimes takes place at only certain times:


The leakage can be detectable visually, by odor, or by instrumentation. A strobe light can sometimes be used to determine its location. As you can see in the above diagram there are several leak paths possible. You must determine which ones you have. The seal can leak:

Between the lapped seal faces. Since they are a wearable surface the leak will probably get either better or worse. It should never remain constant. The leak started for a variety of reasons that include:

  • The spring load has been reduced because of thermal growth, axial thrust, or impeller adjustment.
  • The seal was set screwed to a hardened shaft and has vibrated loose.
  • One or both of the seal faces is not flat. Solid tungsten carbide and silicon carbide faces are often lapped on only one side. You may have installed the face backwards
  • The dynamic elastomer has swollen up and seized the spring-loaded face, preventing it from remaining in contact with the stationary face.
  • The outside springs in a dual cartridge seal were painted during routine maintenance.
  • The product prevented the lapped seal faces from remaining in contact.
    • Dirt has gotten into the sliding components.
    • The product has crystallized.
    • The product solidified or became very viscous.
    • The product is vaporizing across the seal faces, expanding, and blowing the faces open.

At the static and dynamic elastomer locations. In the illustration the O-ring between the inner sleeve and shaft is static. The O-ring between the inner and outer barrel is dynamic.

  • This type of leak tends to remain constant and will often stop when the small opening clogs up with solids. The leak can be caused by a damaged elastomer or damage on the surface where the elastomer seals. In some instances the elastomer is not seated properly. It is twisted because of poor installation, excessive shaft movement, or high-pressure extrusion.

At the gasket between the stationary face and the pump stuffing box.

  • This is the easiest leak to detect because it is very visible and does not change with shaft rotation.

Between the sleeve and the shaft (not shown in the illustration). Some pump shafts have a sleeve installed for corrosion resistance or to act as a sacrificial part for seals that frett and damage shafts.

  • This is a problem with double-ended pumps, where the sleeve is used to position the impeller and there is no method of sealing the sleeve against the impeller.

Between the seal face and its metal holder.

  • The leakage frequently increases as the product temperature increases because the metal face holder has an expansion rate three times that of the carbon or hard faces.

Coming out of the quench connection in an API (American Petroleum Institute) gland.

  • Be sure to explain to the operator that this leakage is normal.

Through fretting damage

  • The damage is caused by spring loaded dynamic O-rings, Teflon® wedges, chevrons, U- cups etc.
  • You can’t miss the frett marks. They’ll be located on the pump shaft, pump sleeve, or inner sleeve of the mechanical seal where the dynamic elastomer contacts the metal part. Many sleeves terminate inside the stuffing box making this leak hard to detect.


  • The stationary seal face has been over tightened against the face of the stuffing box causing the hard face to go out of flat. When the pump is stopped the carbon will relax and re-adjust to the distorted hard face and stop leaking.
  • The clamping is not equal and opposite across the stationary seal face. Look for different width gaskets at the front and rear of the stationary face (see the illustration). The carbon will readjust when the shaft is not turning.
  • Between the face and the face holder. The holder heats up and expands faster than the pressed in face. The leak will begin when the metal holder comes up to temperature. Remember that metal expands three times faster than a carbon seal face.
  • Cryogenic (cold) service will harden the elastomer. Be sure to check the lower temperature limit of the elastomer that was selected.
  • Misalignment will cause excessive movement of the seal components.
  • The shaft is bending and not allowing the seal to move freely. This occurs if the pump is operating off of its best efficiency point and the shaft L3/D4 is not small enough to resist the bending.
  • The product is vaporizing across the seal faces.
  • Cavitation, slip stick, harmonic or some other type of vibration is bouncing the faces open. Check the lugs or drive pins for sign of excessive wear.
  • The seal was installed without enough compression, or the impeller was adjusted after the seal was installed and thermal expansion of the shaft is opening the faces.
  • A discharge recirculation line is aimed at the seal faces or some other critical point and the faces are being forced open.
  • A non- concentric seal, bad sleeve installation or an out of balance rotating assembly is causing the rotating portion of the seal to run off the stationary face.
  • A bent shaft can cause the rotating portion of the seal to run off the stationary face.
  • If you are sealing a non lubricant or a poor lubricant you can have “slip stick” vibration problems while the pump is running
  • The rotating portion of the seal is hitting a stationary object. Look for:
    • A protruding gasket or fitting.
    • A foreign object that has worked its way into the stuffing box area.
    • A stationary portion of the rotating equipment, such as a close fitting bushing.
    • A thermometer well extending into the stuffing box.
  • At elevated temperature the product thins out (the viscosity decreases) and is leaking through an elastomer. It will not leak at the cooler temperature when the product viscosity is higher.
  • High temperature is causing the lapped seal face to go out of flat.


  • The seal is also leaking while the pump is running, but the leak is vaporizing and not visible. Hold a piece of white paper over the seal area while the pump is running and see if the paper becomes damp.
  • A meniscus caused by centrifugal force and liquid surface tension had formed at the inside diameter of the seal faces.
  • You are using high temperature grade Kalrez®. It is too hard at ambient temperature and will soften at operating temperature.
  • The pump is running under vacuum and while it is running air is being pulled into the system. The fluid leaks out when the shaft is static. This can occur if an open impeller that was designed to be adjusted against the volute, has accidentally been adjusted backwards against the back plate. The impeller “pump out vanes” can then pull a vacuum in the stuffing box. This is a common problem if you use a lot of Duriron (Flowserve) brand pumps and then bring in a few of another brand.
    • The problem can also occur if the pump is equipped with a repeller that is causing a vacuum in the stuffing box while the pump is running.


  • Face distortion caused by a high-pressure surge that was created when the pump was started with the discharge valve shut.
  • The shaft is bending and interfering with the seal movement.
    • The shaft is running off of its best efficiency point (BEP) because the pump is started with the discharge valve throttled or shut. Operators shut the discharge valve at start up to save electricity and prevent cavitation.
    • The same problem can occur if the pump is started with the discharge valve wide open, and because of the lack of discharge resistance, the pump will run to the right hand side of its curve. In some cases you could also burn out the electric motor.
    • The product has changed state and becomes a liquid again when the pump comes up to operating temperature. The product had :
      • Crystallized
      • Solidified
      • Became viscous
      • If the crystallized fluid or any type of solids remain on the shaft outboard the seal, the solids can restrict the movement of the seal as the rotating face has to move to compensate for wear.
  • You are getting excessive axial shaft movement at start up. This is a common problem with sleeve bearing equipment.


Look for re-occurring events that initiate the leakage. They can include:

  • Flushing the lines at the end of a batch or season.
  • Alternating pumps in a multiple pump arrangement.
  • An additive is being put into the product.
  • Batch operations are beginning or ending.
  • The cooling water is passing through temperature cycles.
  • The outside ambient temperature has changed dramatically. I ran into a situation where a supplier was oiling the bed of his truck to prevent phosphate ore from sticking to the truck bed in the winter. This oil ended up in the finished product and attacked the ethylene propylene O-ring in the seal.
  • Hard water is being used as a flush and it is gradually restricting the flush lines or cooling jacket.
  • A filter or strainer is clogged in a flush line.
  • The flushing water pressure drops at certain times of the day because of demand.
  • The boiler or cooling tower is being blown down.
  • There is a control valve in the pump discharge that is causing the pump to occasionally operate too far off of the best efficiency point (BEP).
  • The stuffing box is cycling between a positive and negative pressure.
  • Vortexing can occur if the pump suction falls too low. This also occurs in mixers and agitators.
  • You are quenching a high temperature application with water. As the quench water vaporizes it leaves dissolved solids outboard of the seal, restricting axial movement as the seal faces wear.
  • The pump is cavitating on a regular or intermittent basis. Here are a few possibilities:
    • The suction level falls too low
    • The tank vent freezes.
    • The velocity is too high on the suction side of the pump.
    • A suction strainer is plugged up.
    • A stuck or broken check valve in the pump suction piping.
    • A temporary loss of discharge head.
    • A booster pump has shut off.
    • A suction eccentric reducer was installed up side down allowing slugs of air into the suction of the pump.
    • The fluid is vortexing in the supply tank. The level is too low for the pump capacity.
    • The pump is lifting liquid and the foot valve is sticking.
    • The impeller is too close to the cutwater.
    • Air is entering the system through the pump packing.
    • A lower “specific speed” impeller as been substituted.
    • The pump was specified with too low a “suction specific speed” number
    • Reaction bonded silicon carbide can crack if the system lines were flushed with caustic solution


The leak rate is changing – It gets better or worse.

  • This type of leak is usually associated with seal face leakage because the seal face is a wearable surface.
  • The carbon seal face is not flat.
  • The seal face was damaged at the time of assembly.
  • Dirt or solids are imbedded into one of the faces
  • Coke (over heated oil) or some other solid has formed on the seal faces causing them to separate.
  • The rotating face is hung up on the shaft.
  • Outside seal springs have been painted during routine maintenance.
  • The hard stationary face has been installed backwards. These faces are usually lapped on only one side

The faces spit liquid or vapor.

  • The product is vaporizing at the faces. Check the fluid vapor point. When using balanced seals the stuffing box pressure must be at least one atmosphere higher than the product vapor point. Unbalanced seals require a much higher differential pressure.
  • The rotating face is running off of the stationary face.
  • The stationary seal face was not centered to the shaft.
    • The seal is not concentric with the shaft.
    • The rotating assembly is out of balance.
    • The shaft is bent

Fire hose type leakage. The leak is following shaft rotation.

  • Product has solidified on the seal face and a piece has broken off. A high temperature between the faces often initiates this.
  • The rotating face is cracked.
  • The hard surfacing, or coating, is lifting off of the rotating face.

Intermittent leakage.

  • Temperature changes or pressure surges are altering the face flatness within the elastic range of the material.
  • The stuffing box is alternating between vacuum and pressure
  • The movable face is sluggish and not able to follow run out.
    • The product is viscous.
    • The product has started to solidify.
    • The shaft/ sleeve is too large in diameter and it is restricting movement of the seal. Spring loaded dynamic elastomers such as Teflon® wedges, U- cups, chevrons and spring loaded O-ring designs are very sensitive to this problem
    • Dirt or solids are clogging the seal and preventing it from following shaft run out. Designs that use multiple springs in the fluid are prone to this problem.
    • In a non O-Ring version, the spring load is too high causing the elastomer or Teflon® to stick to the shaft. Some designs use spring loaded O-rings that experience the same problem.
  • The product is occasionally vaporizing between the faces.
  • There is a leak between the face and the holder that becomes visible only when the unit comes up to operating temperature.
  • A bending or bent shaft is causing the seal outside diameter to contact the inside diameter of the stuffing box, or some other stationary object.
  • The pump is running with too high or too low a head. This causes a shaft deflection that may be excessive. Check the pump curve against actual operating conditions.
  • The application is cycling between ambient and cryogenic temperatures causing the elastomer to harden on the cold cycle and the faces to go out of flat.

The seal area is damp. There is no visible leakage.

  • There is a leaking flange or fitting above the seal that is dripping close to the seal location.
  • The product is vaporizing. Hold a clean piece of white paper over the running seal, and check for leakage. The paper will become damp.
  • Any condition that could cause intermittent leakage will cause this problem.

Constant dripping. It gets neither better nor worse. This leak cannot be caused by a damaged seal face because seal faces are a wearable surface and the leak rate would have to change.

  • The elastomer (rubber part) is cut or nicked.
  • The shaft/sleeve is damaged at the elastomer location.
  • There is damage in the O-ring groove. Maybe the O-ring was removed with a sharp metal instrument and this has caused a scratch in the O-ring groove.
  • There is a leak path between the carbon and the carbon face holder.
  • Leaking between the cartridge sleeve and the shaft.
  • Leaking between the shaft sleeve and the shaft.
  • Leaking between the gland and the stuffing box. This leak path is very visible in most applications
  • Leaking between the stationary face and the seal gland.
  • The seal faces are stuck open.
  • The elastomer has swollen up due to chemical attack by either the product, the flush, what ever is being used to clean the lines, or by the lubricant that was put on the elastomer to help the installation. This attack usually takes place within one week of exposure to the non-compatible lubricating fluid.


Too much heat is being generated at the seal faces. Remember that unbalanced seals generate more heat than balanced seals.

  • The carbon is being insulated by a gasket of some type and cannot dissipate the heat.
  • You are using high friction face materials. Two hard faces usually generate more heat than carbon vs. a hard face.
  • The carbon-graphite seal faces are positioned too far away from the circulating barrier or buffer fluid in a dual seal application. Carbon-graphite is a poor conductor of heat compared to either silicon or tungsten carbide hard faces.
  • The faces are running dry.
    • The stuffing box has not been vented. This is especially important in vertical applications.
    • You do not have a barrier fluid between the seals in a dual seal application.
  • You have lost an environmental control.
    • Flushing has been shut off.
    • The quenching connection is not operating.
    • The cooling jacket is clogged or not functioning for some reason.
    • The discharge or suction recirculation line is clogged.
    • The barrier fluid has stopped circulating in a dual seal application or you are using oil as a barrier fluid. Oil has a low specific heat and poor conductivity making it a poor choice as a heat transfer medium. If you must use oil as the barrier fluid you may have to forsake convection and go to a forced circulation system or a pumping ring.
    • An API (American Petroleum Institute) type gland has been piped incorrectly
  • Poor heat conductivity of the hard face. Silicon carbide is better than 99.5 ceramic.
  • There is too much spring load on the seal faces:
    • A wrong installation measurement.
    • The impeller was adjusted after the seal was installed. Any pump impeller that adjusts against the back plate has this problem. Durco pumps are a good example.
    • Excessive axial movement of the shaft.
    • Thermal expansion

A seal component is rubbing the inside diameter of the stuffing box or against a product that has attached its self to the inside of the stuffing box. Hard water forms calcium deposits and petroleum fluids can form coke solids at elevated temperature.

  • The seal is not concentric with the shaft.
  • The shaft is out of balance.
  • The shaft is bent.
  • The pump and driver are misaligned
  • There is a pipe strain problem.

The sleeve, shaft or rotating seal is hitting a stationary component.

  • A protruding gasket or fitting.
  • A bushing in the bottom of the stuffing box.
  • A thermometer well could be penetrating into the stuffing box.

A foreign object is loose in the stuffing box.

A suction recirculation line was used to lower stuffing box pressure. The high velocity re-circulation in this line is heating up the return line.


  • The product is cavitating in the pump. There are five types of cavitation:
    • Vaporization.
    • Internal recirculation
    • The Vane Passing Syndrome
    • Turbulence
    • Air ingestion
  • A rotating component is rubbing.
  • The bearings are bad.
  • The seal has come loose from the shaft. The soft stainless steel set screws have vibrated loose from a hardened shaft or sleeve.
  • A foreign object has entered into the stuffing box or was left in the stuffing box during the last seal change.
  • The sleeve is hitting an API (American Petroleum Institute) disaster bushing.
  • The seal faces are running dry. They will make a whistling noise.
  • You have hit a critical speed.
  • Coupling misalignment.
  • The noise is coming from the motor or some near by equipment.
  • “Slip stick” at the seal faces.


The convection tank

  • It is running backwards. There are multiple causes of this problem:
    • The inlet and outlet ports are not drilled properly. Centrifugal force is working against convection.
    • A cartridge dual seal has not been centered properly.
    • Connections have been made to the wrong fittings.
    • The piping fitting has bottomed into the gland fitting shutting off or restricting the flow.
  • The pressure or level in the tank changes.
    • Temperature change of the barrier or buffer fluid.
    • No air pocket was left in the top of the tank.
    • One of the seals is leaking. The pressure or level change should tell you which one. If the inboard seal is leaking the level will increase. An outboard seal leak will cause the tank level to drop or empty completely.
  • The level in the tank drops a little and then holds.
    • The outboard seal is leaking. Without a vent on the top of the convection tank the dropping fluid level created a vacuum in the tank stopping the level at a new height. This is similar to the level you can hold in a drinking straw if you keep your finger on one end of the straw.
  • The convection tank is not convecting at all. The tank was installed incorrectly. The minimum and maximum dimensions for the pipe locations were ignored.

Flow meter not indicating.

  • The meter is broke.
  • Line clogged.
  • The flow is not high enough.
  • The gage graduations are too large for the desired flow.

No flow through the quench and drain connections.

  • You are piped to the wrong connection. Most glands that have been drilled for a quench connection, have a flush connection also.
  • A valve is not open.
  • The pipe or a fitting is clogged.
  • Operators sometimes shut this quench connection off to stop what they think is a seal leak coming out the drain line.

Loss of jacket cooling. The incoming and out going lines are at the same temperature.

  • A layer of calcium has built up on the inside of the cooling jacket.
  • A discharge recirculation line is connected to the stuffing box (it may be hidden inside the insulation and no one can see it).
  • Someone has shut off the cooling water or steam.


There are several different types of vibration that include:

  • Mechanical causes of vibration
    • Unbalanced rotating components. Damaged impellers and non-concentric shaft sleeves are common.
    • A bent or warped shaft. This often happens during the removal of the shaft sleeve, the bearing, or the mechanical seal.
    • Pump and driver misalignment.
    • Pipe strain. Either by design or as a result of thermal growth.
    • Thermal growth of various components especially shafts.
    • Rubbing parts.
    • Worn or loose bearings.
    • Loose hold down bolts.
    • Loose parts.
    • The product is attaching to a rotating component, probably the impeller.
    • Damaged parts.
    • There is not enough mass or weight in the pedestal. If you weigh the pump and its driver there should be a least five times that mass in the pump pedestal.
    • The pedestal is not wide enough. If you drop a vertical line from the center of the motor, two lines radiating out thirty degrees from this centerline should pass through the base, not the sides of the pedestal.
  • Hydraulic causes of vibration
    • Operating off of the best efficiency point (BEP) of the pump.
    • Vaporization cavitation.
    • The impeller vanes are running too close to the pump cutwater.
    • Internal recirculation
    • Air is getting into the system through vortexing etc.
    • Turbulence in the system (non-laminar flow).
    • Water hammer.
  • Other causes of vibration.
    • Harmonic vibration from nearby equipment.
    • Operating the pump at a critical speed. Watch out for this problem in variable speed and pulley driven pumps.
    • Seal “slip stick” at the seal faces.
    • The product is vaporizing at the seal faces
    • A pump discharge recirculation line is aimed at the seal faces. Each pass of an impeller vane gives a pulse to the fluid going to the stuffing box.


  • On February 18, 2018