Guidelines and Fundamental Considerations For Axle Balancing
Guidelines and Fundamental Considerations For Axle Balancing
Guidelines and Fundamental Considerations For Axle Balancing
Axle assembly balancing presents a unique set of challenges. The axle assembly
itself is a fairly complex device that typically has two components rotating at different
speeds, gear meshes and sometimes a clutch or locking mechanism. These
guidelines are set forth as an aid to understanding the essentials and as a
description of some less obvious considerations important to the balancing process.
As complexity increases and balance tolerances drop, it becomes increasingly
important to maintain a grasp on how the fundamentals of two-plane balance
measurement apply to rigid rotor balancing.
It is recommended that the reader review the section on terminology and definitions
as they are essential for a proper understanding of the topics presented. The terms
static unbalance and force unbalance are used interchangeably in this paper.
The Axle Assembly or Final Drive Unit transmits driving torque from the prop shaft to
the axles. It includes an input drive flange, pinion shaft, pinion gear, ring gear and a
differential unit.
Differential Carrier
Slinger
Drive Flange
Ring Gear
Pinion Bearings
Pinion Gear
The differential unit transfers torque to the right and left wheels and allows for
unequal wheel speed during cornering. There are various styles of differential
locking devices and clutches that limit wheel slip. These are not shown in the figure
above and are not discussed in detail in this paper.
Balancing an axle assembly typically means balancing the pinion shaft as well as
those components that are attached to it and turn at the same speed as the prop
shaft. The rotor is normally a rigid assembly that includes the pinion gear, pinion
shaft and flange. Balance corrections are almost exclusively single-plane
corrections. As such, only the static unbalance can be fully corrected. This is
consistent with most axle assembly balance specifications that only limit the
maximum static unbalance.
The axle assembly may have a clutch assembly that regulates torque transmission
from the flange to the pinion gear. It should be recognized that such a clutch splits
the pinion shaft into two separate shafts and adds a stack of clutch discs that are
guided and rotate with one or the other shaft section. The clutch discs are normally
loosely guided and can be a source of significant balance uncertainty. The two
shafts are synchronous when the clutch is engaged but are not angularly registered
to one another. It is important to understand the allocation of the balance tolerance
among these components. Balance criteria may be established separately for each
of these components or may be combined into a single assembly level tolerance. If
the tolerance is shared with a clutch mechanism, multiple measures are required to
properly characterize the combined unbalance of all components. Balance correction
at the axle assembly level is normally limited to the flange and those portions of the
rotor and clutch that are rigidly attached to the flange. Since only the flange is
available for balance correction the clutch and the pinion gear end of the pinion shaft
must be balanced prior to assembly. See appendix E for additional considerations.
Soft Suspension Balancing. The oldest balancing methods involve measuring the
lateral motion of a rotor as it spins in a soft suspension. This method of soft
balancing directly measures the motion of the part as it is free to rotate about its
central principal axis. The motion is measured at one or two planes. Depending on
the sensors employed the fundamental measurement may be displacement, velocity
or acceleration. Through integration or differentiation, each of these can be derived
from the others. Rotor motion is dependent on the mass properties of the rotor and,
to a lesser extent, the suspension. Unbalance alters the rotors mass properties by
shifting the mass center and/or principal inertia axes.
Hard bearing systems consist of a rigid base that supports a measuring platform
through two force measuring elements. The force measuring elements are usually
ceramic or quartz piezoelectric elements. The platform suspension will normally
have other support features to restrict motion to a single-plane. The stiffness of
these supports should be subordinate to that of the sensing elements so that the
largest percentage of unbalance force passes through the two force sensors. The
actual displacement of the platform is very small since it is only due to the
compression of the piezoelectric elements. Displacements are often only a few
nanometers resulting in adequate suspension stiffness over the intended running
speed.
The resonance requirement leads to rigid structures that must resist the unbalance
forces with minimum flex. Hard bearing balancers have evolved to be very heavy
duty, robust machines. The balance platform itself must also be stiff and lightweight
and is ideally suited for complex work holding solutions. Axle assemblies often
require unusual work holding that can be readily adapted to the measurement
platform.
Hard bearing balancers normally employ two sensing elements and as such are two-
plane balance measuring systems. As noted before, soft suspension machines may
have a suspension that mechanically rejects couple motion. This is not necessary
with hard suspension machines. It is difficult to design a suspension sufficiently stiff
in rotation to get complete couple rejection. Without a means to mechanically filter
the forces and/or motion associated with couple unbalance, any single sensor
measuring system runs the risk of recording the response at a nodal point. The
relationship between force unbalance and sensor output can be non-linear and
unreliable near such a node. On the other hand, couple rejection is easily performed
mathematically from the two-plane data. A properly calibrated two-plane system can
reject more than 99% of the couple unbalance. Isolating force unbalance to a degree
of accuracy much greater than what could be accomplished mechanically.
Balancing an axle assembly typically means balancing the pinion shaft and those
components that are attached to it. The most common configuration is a one piece
pinion shaft and gear with a drive flange.
When assembled, the pinion shaft and drive flange form a simple rigid rotor. The
total unbalance of the rotor includes static and couple unbalance in the pinion shaft,
the flange itself and the unbalances associated with mounting the drive flange and
prop shaft.
The one piece pinion gear and shaft is normally almost fully machined and has
relatively small residual unbalance. The axis of rotation is set by the pinion bearings.
They are piloted on the pinion shaft and if the rotor was only composed of the pinion
shaft, the rotor would be well balanced. However, the drive flange must be added.
The drive flange itself is often balanced as a component and will also have relatively
small residual unbalance. While the components may be well balanced,
imperfections in the assembly and alignment of these two balanced components
lead to an unbalanced rotor.
Bending of the pinion shaft has also been identified as a source of unbalance in
cases where the bearing preload technique does not provide for a well centered
axial load. Pinion shaft bending has the potential to unbalance the pinion shaft itself
as well as exaggerate mislocation of the flange.
The flange is normally fitted to the shaft with a spline so that engine torque can be
transmitted to the axle. The flanges position is established by eccentricity of the
spline as well as the squareness of the shoulder on which it rests. These geometric
errors give rise to two significant sources of unbalance.
2) Misalignment of the prop shaft and its associated end weight with respect to the
axis of rotation.
The unbalance associated with both the flange and the prop shaft end fitting is equal
to a weight times an eccentricity (ignoring couple unbalance due to rotation of inertia
axes). The effects of both can be combined and expressed as the total weight times
an eccentricity.
For the most part, axle assembly balancing is a correction of these two components.
End fitting unbalance is addressed at some length in following sections. While the
end fitting can be a significant source of unbalance, it may or may not be included in
the axle assembly balance tolerance. Whether it is or not, it must be recognized that
any prop shaft or drive mechanism that attaches a weight to the flange will be
contributing to the system unbalance.
A balance specification sets a limit for the amount of residual unbalance on a part or
assembly in order to achieve a combination of these goals. The third reason is a
prime motivation for axle balancing i.e., reducing the vibration perceived by the final
customer. Since the goal is to reduce perceived vibration the balance specification
should be targeted to this goal.
Three choices in specifying the maximum allowed unbalance are: 1) specify a force
unbalance, 2) specify force and couple unbalance and 3) specify two-plane
unbalance with plane locations
2) Specifying Force and Couple Unbalance. Limits on both force and couple
unbalance completely specify the allowed unbalance in any rigid rotor. In this case
the couple specification has units of [ML] and requires a location for the mass
center to complete the specification. For the purpose of establishing correlation
between unbalance and the perception of vibration, both force and couple unbalance
levels should be considered. This method is generally the most useful expression of
unbalance and can be used during the design stage to reduce both force and couple
unbalance to acceptable levels. Subsequent manufacturing variations generally re-
introduce unbalance, making balance correction necessary.
As in any case that involves a two-plane balance specification and only a single
correction plane, the acceptable residual unbalance will be some combination of
force and couple. Except for special cases, it is not possible to reduce both force
and couple unbalance to zero with the single correction. Some special cases are
A compound rotor that is constructed from two well balanced components will often
exhibit a state of quasi-static unbalance. The same state of unbalance will occur if a
balanced rotor is unbalanced with a single weight placed some distance away from
the composite mass center.
For a well balanced pinion shaft with a well balanced flange, mounting eccentricity
will produce a quasi-static state of unbalance.
The combination of static and couple unbalance can be corrected with a single
balance correction if the correction is made at the proper location. While the rotor
has both static and couple unbalance, the entire unbalance is attributed to
eccentricity of a single weight or mass. If the couple unbalance created by flange tilt
is ignored, the best correction location is a plane that contains the mass center of
the flange. Correction in any other plane will leave additional couple unbalance.
This leads to a discussion of the correction technique and a closer look at single-
plane and two-plane balance correction.
Single-plane balance correction works best when applied to rotors that are disk like.
Unbalance in a disk shaped rotor can often be almost entirely attributed to
eccentricity of the rotor itself with respect to the axis of rotation. A balance correction
made on the disk is often made very close to the plane that contains the mass
center and consequently is very effective in drawing the mass center back to the
axis of rotation without introducing significant couple. There is little need for a
second correction plane as the levels of couple unbalance are low. Rotors that are
not disk-like are prone to higher levels of couple unbalance. It is more likely that
correction planes on longer, cylindrical rotors will be some distance from the rotor
mass center and each correction will naturally and simultaneously alter both the
static and couple unbalance.
In the case of the pinion shaft, only one correction plane is allowed for a rotor that is
more like a cylinder than a disk. The location of the balance correction plane is such
that its influence on static and couple unbalance cannot be ignored. For this
discussion the right plane is the correction plane and is located on the flange.
Correction weight will be added or removed in the right plane. When referring to the
flange, reference is being made to the combined weight and unbalance of the flange
with an attached end weight. Since there is only one correction plane, the location of
the other plane, the left plane, is somewhat arbitrary. Furthermore there is the
dilemma of which correction is the best correction: 1) a force correction, 2) a right
plane correction or 3) some combination of force and couple corrections.
a) Left Plane Location. Consider a rotor with left and right corrections
(unbalance) already defined.
The couple correction is out of phase with the original left plane correction
and in phase with the original right plane correction. The magnitude of the
couple correction must equal the resultant moment
M = C(nd)
or
C = M/(nd),
For large values of n e.g., the left correction to the extreme left, the magnitude
of the couple correction approaches that of the original left plane correction.
The new left plane correction approaches zero while the new right plane
correction approaches R+L. This is nearly equivalent to making the entire
force correction in the right plane.
It is assumed here that the unbalance weight is small with respect to the rotor
weight. Unless the mass center lies in the right plane, such a rotor will exhibit
a quasi-static state of unbalance. If the mass center lies in the right plane,
couple unbalance will be zero and the rotor is a case of simple static
unbalance. The static unbalance is shared between right and left planes in
inverse proportion to their distance from the mass center.
a
LStatic = m r
D
RStatic = mr
(D a )
D
a
LCouple = m r
D
a
RCouple = mr
D
L=0
R = mr
The right plane correction contains the entire balance correction and is
entirely independent of both its location with respect to the mass center and
the distance between planes.
Again it is assumed here that the unbalance weight is small with respect to
the rotor weight.
c
LStatic = m r
D
RStatic = mr
(D c )
D
a
LCouple = m r
D
a
RCouple = mr
D
ca
L = m r
D
Dc+a
R = mr
D
The left plane correction represents the residual unbalance after making the
right plane correction. Note that the magnitude of the residual unbalance
depends on the ratio of b to D . For b = 0 or for D much larger than b , the
residual unbalance is negligible and both expressions reduce to those
presented in the previous case where the unbalance weight is in the right
plane. Otherwise for b 0 , the left plane correction is non-zero and is
b
proportional to the ratio of .
D
Geometric errors in the location and orientation of the pinion flange datums turn into
unbalance with mounting of the prop shaft to these datums. It is possible to calculate
this unbalance by determining how far off center the prop shaft end weight is placed
as a result of these geometric flaws. This analysis is useful in evaluating
repeatability as well as unbalance bias.
It is first necessary to know the mass properties of the end weight. If the end weight
is not the prop shaft itself, it should be representative of the prop shaft. A balance
bias will exist between measurements with tooling of different weights.
A portion of the prop shaft weight acts at the pivot point. This weight is designated
nW2 where W2 is the total weight of the prop shaft itself (less end fittings). There
are two geometric errors associated with the pinion flange that create unbalance: 1)
eccentricity of the pilot diameter and 2) non-perpendicularity of the flange.
1) Pilot Eccentricity. This is eccentricity with respect to the pinion shaft centerline
of rotation.
The variable n represents the percentage of the total weight of the prop shaft, W2,
associated with the end that is attached to the flange.
2) Non-Perpendicularity. This is a case where the pinion flange face is not exactly
perpendicular with the centerline of rotation.
1 2
For the same mounting errors, a balance correction error or bias is introduced when
tooling used to measure unbalance does not match the production component
weights and geometry.
Assume that the tooling and the production components center and square on the
same features:
0 = centering error
= squaring error.
The individual quantities are calculated as outlined above. Again, these are all vector
quantities.
See Appendix C for a method of combining W1 and W2 into a single equivalent end
weight.
See Appendix D for further description of the prop shaft end weight.
Vehicle Vibration
For a dynamic analysis, right and left plane unbalance values can be translated into
equivalent unbalance in any other two planes. This technique can be used to
estimate forcing functions at support points by translating unbalance to the points of
interest. Axle assembly mount stiffness should be considered in the analysis as a
resilient mount may provide some degree of isolation in the normal speed range.
Some difficulty may be encountered distributing the unbalance in cases where the
axle assembly is over-constrained where there are multiple paths for the force
and/or couple unbalance forces to be reacted. Assumptions and results should be
validated in vehicle tests.
More subjective, empirical testing is common for NVH evaluations. Various levels of
force and couple unbalance can be applied to an axle assembly which is then
evaluated in a test vehicle. Driver evaluations are compiled to establish limits for
vibration and unbalance. This technique is widely used for HVAC and drive train
components.
Balancing terminology used in this paper is consistent with that used in ISO 1925,
Balancing Vocabulary. A few of the more prominent terms are defined below.
AXIS OF ROTATION: The instantaneous line about which a body rotates. In the
case of a rigid rotor supported by rigid bearings, the axis of rotation is the shaft axis.
CENTRAL PRINCIPAL AXIS: The principal inertia axis that is most nearly
coincident with the shaft axis. The mass center lies on the central principal axis. This
is also known as the balance axis or the mass axis.
COUPLE UNBALANCE: The condition of unbalance for which the central principal
axis intersects the shaft axis at the mass center. By definition, the central principal
axis and the shaft axis are not parallel. The units of couple unbalance are [ML2] or
[WL2].
Since dynamic unbalance may be a combination of static and couple unbalance and
static and couple unbalance have different units, there are no unique units for
SHAFT AXIS: The straight line joining the journal centers. In the case of a rigid
rotor supported by rigid bearings, the shaft axis is the axis of rotation.
STATIC UNBALANCE: The condition of unbalance for which the central principal
axis is only displaced parallel to the shaft axis. Static unbalance is also known as
Force Unbalance. The units of static unbalance are [ML] or [WL].
Two-plane Calibration
The calibration technique employed by BTI is an application of the method for two-
plane influence coefficient balancing developed by Thearle (1934). The balance
platform has two sensors that give the balancer the ability to measure static
unbalance, couple unbalance or any combination of the two. The influence
coefficient method is well documented and only a brief description is offered here.
First, two calibration planes are defined and calibration weights are selected.
A and B are the unbalance signals and L and R will be known unbalances. By
placing weights in one or the other calibration planes, a known unbalance is applied.
Three measurement spins are required: weight in the left plane, weight in the right
plane and no weights. The measured signals and known unbalances are all vector
quantities that can be expressed as amount and angle or, for rotation around the Z
axis, as vectors in the X-Y plane
A = Axi + Ayj
B = Bxi + Byj
L = Lxi + Lyj
R = Rxi + Ryj .
Here i and j correspond to unit vectors along the X and Y axes respectively. The X
and Y components correspond in turn to the cosine and sine components of the
dynamic signal.
LX AX
= [ ]
R X B X
LY AY
= [ ]
RY BY
F=L+R
C = (L R).
Uc = L x d.
This is a cross product and results in a vector that lies normal to the couple
correction vectors, C. It properly reflects the axis about which the couple moment
acts.
The couple corrections can actually be made in any pair of planes that have the
same separation distance as the right and left planes. Or more generally, can be
made in any pair of planes with an adjustment for the distance between planes. If L
L' = (d/d')L
R' = (d/d')R.
Corrections in these two planes can be translated into any other two planes using
simple static analysis and constructing an equivalent force system. The corrections
can also be expressed as force and couple unbalance with respect to any single-
plane or point on the axis of rotation using the same technique. The following are all
equivalent force systems and illustrate moving a force correction from point O to
point P.
The force correction is equal to the true static unbalance in the rotor and as such
can be made in any plane. The couple correction will vary depending on where the
force correction is made. If the force correction is made in the plane that contains the
rotors mass center, the couple correction will reflect the true couple unbalance in
the part and is related to the tilt of the inertia axes.
An end weight has two discrete weights that are of interest in the balancing
application. It is often useful to combine these two discrete weights into one
equivalent weight. This is a mass center problem. As before, the end weight can be
described in terms of weights W1 and nW2 at distances X1 and X2 from the
mounting face.
The equivalent end weight will have weight W at a distance L from the mounting face
as shown on the right.
For a linear offset, the location of the end weight is unimportant. However it should
be clear that the total weight must match the sum of the two discrete weights
W = W1 + nW2.
2 1
If the unbalance of the single weight is to be equivalent to the total unbalance of the
two, then
and
L = (W1X1 + nW2X2)/( W1 + nW2).
A fully assembled prop shaft generally has 3 components joined at 2 pivot points.
The prop shaft is joined to the pinion flange and drives the differential through the
pinion flange.
The total weight of the prop shaft is the sum of these 3 individual components,
WTOTAL = W1 + W2 + W3.
The 2 components on the left hand side, the end fitting and the prop shaft, can have
a significant contribution to the system unbalance when mounted to the pinion flange
on the differential. The mounting is seldom perfect and will usually have radial runout
and face runout. These runouts may result in a centerline offset, , of the pilot
diameter and a flange mounting face that deviates from perpendicular by a small
angle, . Both are measured with respect to the centerline of rotation of the pinion
shaft.
These errors may be independent of each other. They position the mass center off
of the centerline of rotation of the pinion shaft resulting in unbalance.
If the mass properties of the prop shaft are understood, it is possible to include a
balance correction to the pinion and reduce the system unbalance. To calculate the
correction it is necessary to know the weights of the prop shaft and the pinion side
end fitting. The axial locations of the respective mass centers is also required, along
with the length of the prop shaft and the location of the pinion side pivot point.
If the prop shaft is symmetric or nearly symmetric, there is little error in assuming
that the mass center is at the midpoint between the pivots. Each pivot point will then
support half the weight of the prop shaft.
If the difference between X1 and X2 is small or if X1 and W1 are small with respect
to X2 and nW2, the end weights can be estimated by measuring the end weights of
the prop shaft when simply supported at the pivot points. Only the weight of the
pinion end is of interest.
1) Balance Measurement Process. For an axle assembly with a clutch, the balance
process is commonly split into two individual balance measurements with a
clocking process in between the two measurements. This adds the requirement
that the balancer be able to command a torque and hold the output shafts during the
clocking. Balance measurements are taken with the clutch at full torque so that the
pinion and flange do not slip with respect to each other. Between the two
measurements the holding torque is reduced and the input is turned 180 degrees
while the output is held. The second measurement is then taken at full torque and
the flange unbalance data is derived as the vector average of the two readings. The
clutch unbalance is derived (for reference) as half the vector difference of the two
readings.
It is also common to have a pre-clock step before the first measurement to set the
clutch discs in position. The clutch is composed of many friction discs that are
loosely guided on either the flange end or the pinion end of the shaft. The discs have
some freedom to move laterally and the ability to alter the unbalance of both ends of
the shaft. A pre-clock step often improves repeatability. This is accomplished by
holding the output, applying partial torque, and rotating the input one or more turns.
2) Balance Variation from the Clutch. Even with a pre-clocking step, the balance
variations from the clutch may be high with respect to tolerances. Any lateral motion
of the clutch discs or other loose clutch components can be a source of significant
balance uncertainty. Appropriate tolerance specifications are required as are
adequate quality management practices.
The addition of a joint, typically a spline joint, between the clutch and the pinion shaft
introduces another source of unbalance and unbalance variation. Various bearing
configurations lead to the pinion end centering the flange end or vice versa. This
interface needs to be understood with regard to axle assembly unbalance. Runout in
this joint leads to unbalance that may be associated with the pinion end of the shaft
The following examples are drawn from actual test data. Two parts are compared;
sample part 2 has the higher levels of initial unbalance. The axle assembly is
generic in nature, with a typically short pinion shaft and no clutch or slip limiting
devices. The right plane is the correction plane and is on the flange. Three
correction techniques are evaluated: 1) force correction, 2) right plane correction
with left plane at the axle centerline and 3) right plane correction with left plane 125
mm to the left of the axle centerline.
77.0 gmm 62 10,558.3 gmm 173 130.5 gmm 178 17,824.7 gmm 274
77.0 gmm 62 42.2 gmm 263 130.5 gmm 178 71.3 gmm 4
15.2 gmm 328 79.4 gmm 73 9.4 gmm 51 136.4 gmm 181
Note that force unbalance is the same for both expressions of force unbalance but
differs for couple unbalance. The difference is the plane separation or couple arm.
By dividing the couple unbalance moment [gmm] by the distance between planes
(250 mm) the couple force or unbalance is calculated. Note that there is also an
angle difference in the couple expressions. When expressed in [gmm], the angle is
to the axis about which the moment acts. When expressed in [gmm], the angle
corresponds to a couple force or unbalance in the left plane. The left plane couple
force is accompanied by an equal and opposite force in the right plane. These
couple forces both act in a direction normal to the moment vector.
0.0 gmm 0 77.0 gmm 242 0.0 gmm 0 130.5 gmm 358
0.0 gmm 0 3,790.9 gmm 238 0.0 gmm 0 2,339.3 gmm 321
15.2 gmm 328 15.2 gmm 148 9.4 gmm 51 9.4 gmm 231
Note that force unbalance is reduced to zero and that the couple unbalance is also
reduced. The fact that the couple unbalance is substantially reduced is an indication
that the unbalance is largely due to a single unbalance located near the right
correction plane. The small initial unbalance in the left plane is unchanged. The
amount of this residual determines the final couple unbalance. Also note that left and
right unbalances are equal and opposite. This reflects the fact that only couple
unbalance has been left in the part.
15.2 gmm 328 1,895.4 gmm 238 9.1 gmm 51 1,169.7 gmm 321
15.2 gmm 328 7.6 gmm 328 9.4 gmm 51 4.7 gmm 51
Right plane unbalance is reduced to zero as it should be for a perfect correction. The
residual unbalance in the left plane is the same as that left after the force correction
in the previous example. Residual unbalance is a combination of force and couple
unbalance.
77.0 gmm 62 10,558.3 gmm 173 130.5 gmm 178 17,824.7 gmm 274
77.0 gmm 62 28.2 gmm 263 130.5 gmm 178 47.5 gmm 4
10.1 gmm 328 78.3 gmm 69 6.2 gmm 51 134.4 gmm 180
Note that initial unbalance is not dependent on plane location when expressed as
force and couple unbalance with couple in [gmm]. The change in couple when
expressed in [gmm] is due to the larger distance between correction planes. Since
the distance between couple unbalance forces is increased, the couple unbalance
decreases accordingly.
10.1 gmm 328 1,263.6 gmm 238 6.2 gmm 51 779.8 gmm 321
10.1 gmm 328 5.1 gmm 328 6.2 gmm 51 3.1 gmm 51
Again, the right plane unbalance is reduced to zero as it should be for a perfect
correction. The residual force unbalance is slightly lower than that of example 2. This
is due to the greater distance between planes. As the left plane moves to the left, the
right plane correction becomes more like a force correction. The new left plane
unbalance reflects all of the unbalance remaining in the rotor. The left plane
unbalance could be used to predict the forces imposed on or transmitted through the
mounting point in the left plane.