Stationary cartridge seals. How to solve the problem of seal movement 11-10
Everyone agrees that cartridge seals are the only way to go. The good news about cartridges includes:
- The installation is much quicker than conventional seals that have to be set screwed to the shaft.
- There is little chance to damage the lapped seal faces during the assembly process.
- No prints are needed at assembly. There is nothing to measure. The cartridge makes it automatic.
- With a cartridge seal you can adjust impellers to compensate for the inital setting, thermal growth and impeller wear.
But then there is the bad news, and in this instance the bad news is “really bad”!
In another paper I discussed the advantages of using stationary seals and you will recall that their advantage rested upon the seal designer’s ability to position the rotating face perpendicular or square to the shaft.
As long as the rotating face is square, or perpendicular to the rotating shaft there will be no “back and forth” axial movement of the stationary face.
Please take a look at the following sketch. Here you can see that the rotating face is positioned square to the shaft because of the clamping arrangement of the rotating face. The clamp is manufactured on a mandrel and the faces are machined perpendicular to the mandrel, making them in turn, perpendicular to the rotating shaft.
If the rotating face had been set screwed to the shaft it would have “cocked” the rotating face relative to the shaft. This would cause the stationary face to move “back and forth” twice per revolution of the shaft, causing the same problems we experience with the rotating version of a mechanical seal.
The seal movement problem starts when we try to put this rotating face on a cartridge sleeve. Take a look at the next line drawing:
The drawing is exaggerated to emphasize the point. As you tighten the sleeve set screws to the shaft the sleeve will “cock or tilt” and, although the rotating face stays square to the sleeve, it is no longer square to the rotating shaft. This will cause the stationary seal to act like a rotating seal and you lose all of the advantages you gained with a stationary design.
- The spring loaded stationary face will move back and forth axially twice per shaft revolution.
- Any solids in the fluid could lodge in the sliding components of the seal and open the lapped faces.
- The moving elastomer will frett and damage the stationary face depending upon the amount of movement and the seal materials involved.
If we look at the detail of the stationary face on the cocked sleeve we will see:
1. The A.P.I. (American Petroleum Institute) recommends a tight tolerance fit between the pump shaft and the seal sleeve to prevent the sleeve from cocking when the set screws are tightened.
2. The next line drawing describes a design where both the stationary and rotating faces are spring loaded. In this design you are running a rotating seal against a stationary seal.
3. The following sketch describes the double O-ring method for keeping the rotating face square to the shaft.
4. The next drawing describes a three point contact similar to what you would find on a three jaw chuck used on a lathe or drill press. This arrangement is called a “cloverleaf” design by one of the major seal manufacturers.
- Three set screws positioned at 120 degrees apart, deform the sleeve to the shaft outside diameter to insure squareness of the rotating face.
- An additional three set screws go through the sleeve and lock the sleeve to the shaft. These set screws are positioned 120 degrees apart and are located between the set screws that are centering the sleeve to the shaft.
#1 The A.P.I. (American Petroleum Institute) version:
- The tight tolerances required to get the slip fit are expensive.
- There is wide variance in the tolerance used on the outside diameter of conventional pump shafts. If you adopted this method to get “squareness” you would have to rework or replace many of your existing shafts or shaft sleeves.
- Close fitting shaft sleeves are difficult to remove. The necessary heating and banging will almost guarantee a bearing replacement along with the new seal.
#2 Two spring loading both faces:
- Centrifugal force is working for you. The higher the centrifugal force the stiffer the system.
- The centering of both faces is critical. If the hydraulic balance lines are not exact the faces could cock. This is a difficult problem to over come.
- Building two spring loaded faces is expensive. You are actually running a stationary seal against a rotating seal
#3 The double O-ring system:
- This design requires a lot of axial space. When ever possible you will want to get the seal faces as close as possible to the pump’s inside or radial bearing.
#4 The three point contact method:
- At this writing this is the lowest cost of the four solutions.
- This design takes a very short axial length, making the cartridge assembly no longer than a conventional cartridge design.
- Replacing the seal components is low cost and easy with this design.
The cartridge mounted stationary seal doesn’t make any sense unless you are using one of the above solutions, or some other comparable design that corrects the problem of “cocking or tilting” the rotating seal face.
If you are approached by a seal salesman with some other method to insure squareness, have them attach their design to a shaft and measure the squareness with a dial indicator. It will either be square or it will not, you can tell quickly.