menu 9

Online Course for Practical Solutions Vibration

Seminar Schedule

Online Vibration Book

Request More Information

Practical Solutions to Machinery and Maintenance Vibration Problems

Chapter 5, Unbalance

Section 16, Unbalance Due to Assembly Errors - Key Length Considerations

Something as simple as key length can be a major source for additional vibration due to unbalance. The key is part of the total rotor assembly's dynamic balance. International standards (starting in 1990) require that the primary rotor, such as an armature, fanwheel, turbine, etc., be balanced using a "half key," that is, a key that is equal in length to the straight portion of the keyway and only half the height of the finally assembled key's full cross-section. For example, the straight portion of an armature's keyway may be 8 inches (approximately 200 mm) and 3/4 inches wide. A flat, cold rolled steel bar that is 8 inches long, 3/4 inches wide and 3/8 inches high is used to fill the keyway void and taped into place before balancing. (With some modern balancing machines, using a steel bar half key is not actually necessary as the effect of the half key is programmed into the balancing machine's computer.)

Now assume that the rotor to be assembled to the armature is a coupling half with a hub length (and straight portion of its keyway) of 4 inches. It's keyway is usually of the same width as that used in the total assembly; therefore, the half key used for balancing would be 3/4 inches wide and 3/8 inches high. The principle used is that if each rotor is balanced with its own half key, the assembly of the two rotors to each other is supposed to result in a balanced assembly. This will be true only if the correct length of the final key of square cross section is used.

Most companies let the technician who is assembling the two rotors determine the length of the key used. If, for example, the technician uses the key that is often supplied with a new armature, the final key will be 8 inches long. However, that will result in added unbalance due to the extra metal protruding above the shaft that is 4 inches long, 3/4 inches wide and 3/8 inches high. If, instead, the technician uses a final key length that is equal in length to the coupling half's bore, then there will be metal missing in the shaft's keyway that is also 4 inches long, 3/4 inches wide and 3/8 inches high. Therefore, to preserve the balance of both the armature and the coupling half, the technician should follow a very simple procedure:

1. Measure the length of the straight portion of the keyway for the armature. Do the same for the straight portion of the coupling half's keyway. (Tape measure accuracy is sufficient.)

2. Add the two lengths together and divide by two, to obtain the length of the final key used in assembly.

Using an arbitrary key length usually results in machine vibration that is not the difference between a very rough running machine and a smooth one. Instead it results in the difference between a mediocre or "good" running machine and a very smooth running machine. For example, a "good" vibration level that was obtained without these precautions, could be an amplitude of 0.12 in/sec or 3.0 mm/sec. By using the suggested precautions, the orbit could be reduced to about 0.04 in/sec or about 1 mm/sec. In this situation numbers don't show how this could affect machine bearing or seal life as much as a diagram of their orbits (orbit of shaft centerline around the axis of rotation.)

Now visualize not the orbit itself, but the shaft's outside diameter acting on that bearing or seal. With the smaller orbit, would there be longer bearing or seal life? If so, would that extra life be worth the extra precaution of making sure that the proper key length is used?

The last possibility is to simply balance the assembled rotors with whatever key is convenient. But that balance will hold true only for those rotors using that specific key length. Another way is to balance rotors with their coupling half attached, indicating that on the maintenance record, so that a trim balance may be alerted if the coupling half is ever changed.

The above procedure describes what should be used for almost all rotors. However, there are many key designs that are not as simple as described. For example, there are keyways that are straight all along their depths but rounded at each end. They are not the usual straight ones with square ends, but instead have rounded ends. The key completely fills the keyway. The rotor to which it is assembled, such as a sheave or fan, may not have exactly the same geometry. Other keys have even more complicated shapes, as the portion in the shaft may not have the same volume as that in the attached rotor. To determine how to compensate for each part being balanced, don't try to work with just length, height and width (as you would for a simple key), but instead work with total volume. During the balancing operation, the shaft for example, will use the same volume of metal as at the final assembly. The attached rotor will also use the same volume as in the final assembly.

Notice that this could get quite messy, especially when the determination of key length is left to the balancing machine operator or technician performing the final assembly. The determination of key lengths during balancing and assembly should be decided by a more technically-minded person. Once the simple procedures are determined for each type of key used, they should be taught to those operating the balancing machine and to those who do the final assembly. Very soon, all will become simple and easy to accomplish routines.

A final note with a very subjective opinion on the part of the writer. Up until the year 1990, the half key procedure for balancing as described above was used in the Western Hemisphere, Japan and the United Kingdom. However, for Continental Europe and most of the rest of the world, another system was used. That system required that the "full key" of complete length and cross section be attached to the primary rotor for its balancing. The attached rotor (called a "fitment"), such as a sheave, fan, and coupling half, did not have its empty keyway filled for balance. This resulted in unbalance, making it necessary to balance each fitment.

The principle was that a properly, dynamically balanced primary rotor, including the whole key, assembled to a fitment dynamically balanced -- without a key -- would comprise a balanced total assembly. This was certainly correct. However, there was much argument between some European ISO Committee members and those who favored using the half key method. Those who favored the half key method reasoned that the fitment (especially a coupling) would, therefore, not require dynamic balancing. (See section "Unbalance in Couplings.") Upon a vote, the half key system was made an ISO international standard, effective in the year 1990. That meant that European primary rotors balanced before 1990 used a full key; but a replacement coupling balanced, for example, in 1992 was balanced with a half key. The combination produces a total unbalance due to wrong key weight that is greater than anything experienced before. The writer feels that the present standard is incorrect, but as it is now an international standard, it must be taken into consideration when attempting to obtain precision balance (or explaining why two balanced rotors can produce an unbalanced assembly). To make it worse, the Committee did not require rotors be marked in some simple way to indicate whether balanced with a half or full key. This difficulty is not present when using rotors made in the Western Hemisphere, United Kingdom and Japan as the half key method was used before and after 1990. However, it is a problem for all other rotors that were balanced in different periods, or when a pre-1990 European made rotor is assembled to a Western Hemisphere made rotor (regardless of the year it was balanced).


Textbook Index

This textbook contains only part of the information in our Practical Solutions seminar.

Link to Seminar Schedule. Order a print version of this entire textbook.

Home PageEmail Update