1

A NEW SHAFT ALIGNMENT TECHNIQUE

B.C. Howes

Beta Machinery Analysis Ltd., Calgary, Alberta, Canada, T3C 0J7

ABSTRACT by a "“laser alignment” iconoclast. There are

many reasons why dial indicators are a viable
A new technique for shaft alignment is easy to option to measure alignment. The capital cost
apply for verification of alignment and can save outlay for a laser system versus a set of dial
users $10,000 to $20,000 in equipment costs. It is indicators and magnetic bases is the obvious first
called the Reverse-Face Alignment technique. reason for looking at efficient alignment methods
Examples are given with photographs for different that do not involve laser systems. In the author’s
styles of couplings such as gear, elastomeric and experience, the Reverse-Face Method can allow
flex-pack types. Alignment acceptability is faster attachment and faster data collection versus
determined immediately without complex using a laser system. The format of the data from
geometrical calculations as necessary with other the Reverse-Face Method is immediately
methods. The alignment measurement equipment comparable with a guideline so that no
not only is inexpensive, but can be installed complicated calculations are required to determine
quickly for quasi-hot alignment checks. Remote if the dial readings are indicative of acceptable
readout and computer connection are possible, alignment or not. Compared to the Reverse Dial
but these complications are not perceived as Method, the installation of the dials is easier and
being a benefit in most cases. Pitfalls are faster for the Reverse-Face Method, and the
discussed for this and other methods for interpretation of the results is easier.
comparison purposes.
The state of shaft alignment is traditionally
1. INTRODUCTION described by including a measurement of parallel
offset between the shafts. However, the angularity
Shaft alignment can be measured and corrected in at flex-planes instead of parallel and angular offset
many ways. The goal is to cost-effectively and of the shafts to define alignment quality is a
efficiently get a machine aligned and back running. fundamental point in this paper.
The capital cost of the tools should be weighed
against the total time required to do the alignment. 2. ALIGNMENT METHOD SELECTION
Much of the time required to do an alignment is ISSUES
taken up by tasks not directly related to the
measurement of misalignment. Therefore, The following list contains some of the issues that
acceleration of a portion of the alignment task help to determine what will be the alignment
through large capital expenditure may be method of choice:
marginally beneficial.
• the time required to prepare to do the
In some cases, a tool to quickly check the state of alignment check
alignment is all that is required. What is the • the time required to actually do the alignment
alignment and is the alignment within guideline? check
In other cases, the alignment may be expected • the time required to change the alignment
(and found) to be good, but for trending purposes • the time to re-check the alignment after
it is desirable to document the current alignment correction
on an ongoing basis. Slow changes in alignment • the cost of hardware and labour
can indicate changes in foundations that require • the resulting payout
correction in the long term.

These days, the trend in some quarters seems to

be to think that using a laser alignment system is
the only way to do alignment. This paper is written
 2

3. METHODS DISCUSSED float-compensated A readings (eg: A*1 and A*2)

by the formula [(A-B)/2].
The methods that will be referred to in this paper There are two possible sign conventions for the
are: dial readings. If the dial indicator base is mounted
• Reverse-Face on one side of the flex-plane and the dial is
• Reverse-Rim (more familiarly called Reverse pointing away from the magnetic base as it
Dial Method) touches the other side of the flex-plane, then a
• Rim-Face, and positive change in the dial reading indicates the
• Laser coupling halves are closing. This is the standard
mounting convention. On the other hand, if the
A good reference in the discussion of alignment is dial is turned around to point back toward the base
a book by John Piotrowski called “Shaft Alignment as it touches the other side of the flex-plane, then
Handbook” (Ref. 1). Piotrowski’s allowable the sign convention is reversed: a positive change
misalignment guidelines will be discussed with this in the dial reading indicates the coupling halves
author'’ interpretation of how to apply the are opening. This is a potential source of
guidelines. Comments by Piotrowski, combined confusion for the uninitiated. Photograph 1 shows
with a reference from another paper, about gear a dial indicator in the standard orientation. (Dial
couplings are included. There is no mention of the indicators show a positive reading when the
Reverse-Face Method in Piotrowski’s book. plunger is pushed in.)

4. THE GOAL OF ALIGNMENT The diameter of the circle described by the dials
as the shaft is rotated is the basic dimension
The goal of the alignment process is to make the required for calculating flex-plane angularity. The
angularity at each flex plane of the coupling lengths between flex-planes and the distances
sufficiently small with the machine in operation. from flex-planes to feet on the machine-to-be-
This statement assumes a spool piece coupling. If moved, are also required before alignment
the coupling has a single flex-plane, then the corrections can be calculated.
offset between the centerlines of the shafts at the
flex-plane must be made sufficiently small, as well Turning the shaft to the usual 4 positions of the
as the angularity. Note the use of “sufficiently clock can be done by eye (often there are bolt
small” as opposed to “minimized”. In the case of patterns that help), or an inclinometer can be
gear couplings or u-joints, it is not desirable to attached to the shaft for more precise guidance as
eliminate angularity totally, as discussed below. to shaft angular position.

5. DESCRIPTION OF THE REVERSE-FACE Data quality is checked by comparing the sum of

METHOD the vertical dial readings with the sum of the
horizontal readings. Ideally, the sums should be
The dials indicators should be mounted in pairs, equal. In practice, small differences will be seen.
180 degrees apart, at each power transmission Repeat the collection procedure if the differences
point or flex-plane. The reason for this pairing is are “large”. Large is to be considered relative to
to compensate for axial float that can, and usually the dial readings. If the misalignment is large,
does, occur during the rotation of the shafts. then a larger difference between horizontal and
(There are those who use one face dial and vertical dial sums can be accepted. The final
attempt to force the shafts into the same axial sums for normal machines are usually no more
position for each reading. I do not recommend than 1 or 2 thou different. Some alignment
this approach.) methods offer the option of not making a full turn
on the dials or laser equivalent. This short-cut
Refer to Photograph 1 for an example of how to loses the data check discussed above.
mount a dial in the face direction across a flex-
plane. For discussion purposes, a viewpoint for the
machines is required. View the unit from the
The dials should be labelled distinctively (eg: dial driven machine, looking toward the driver. Left
A1 and dial B1, dial A2 and dial B2). Start with and right sides of the unit are determined this way.
dials A at 12:00 o’clock and B at 6:00. Record the The flex-planes can be referred to as the near and
readings in pairs for dials A at 12:00, 3:00, 6:00, far, near being closest to the driven machine.
9:00 and 12:00. Convert the A&B readings into
 3

Usually, the driver will be the machine to be the flex-planes, and are noted for problems due to
moved. misuse by users, in the author’s experience. On
The allowable angularity should be determined at the other hand, the speed of data collection and
the start of the job if speed is of the essence. The calculation of moves and shims is excellent. In
graph in Figure 1, is recommended for this some cases, the brackets are too bulky to swing a
purpose. The maximum shaft speed is required full circle. Some systems get around this problem
as well as the type of coupling. Gear couplings by calculating the missing readings from several
should be left below the bottom line. Convert the intermediate readings at known angles.
alignment guideline expressed in mils per inch into
mils TIR by multiplying the guideline by the The Reverse Dial or Reverse-Rim Method
diameter swept by the dials. The result is the requires complicated brackets plus correction for
largest difference that should be seen between the bracket sag. The installation of the brackets takes
vertical or the horizontal dial readings [the axial longer than mounting a set of face dials. The
float compensated readings of course.] resulting dial readings must be converted via
(Note: 1 mil = 1 thou = .001 inch) complex calculations to angularity numbers. The
cost of the brackets is greater than the cost of face
The measured flex-plane angularities in the dial equipment. The speed of collection of data is
vertical and horizontal planes are calculated by certainly no faster than the Reverse-Face Method.
dividing the difference between the dial readings
by the dial swept diameter. A sample spreadsheet The Rim-Face Method has the same
is included as an appendix. Alternatively, disadvantages as the Reverse-Rim Method for
compare the dial reading differences with the spool piece couplings. However, since it is
number calculated from the guideline angularity. insufficient for alignment determination in general
to measure an angle across a single flex-plane
The ultimate check of angularity is done by taking coupling [in other words, the Reverse-Face
the square root of the sum of the squares of the Method does not work in this instance], one of the
horizontal and vertical angularities at each flex- Rim-Face Method, or the other methods above, is
plane. This is the correct way to determine required when there is no spool piece. Rim-Face
angularity, but in most cases, will make only a is a logical method to complement the Reverse-
small difference. Face Method since no special brackets are
required to use Rim-Face for a single flex-plane
In all alignment work, the issue of hot versus cold coupling. All that is required is three magnetic
alignment must be addressed. I have found that in bases and dials. In addition, for couplings like that
most cases, the Reverse-Face Method allows me shown in Photograph 3, the Rim-Face Method is
to mount the magnetic bases and collect the dial the only logical method. Reverse-Rim or Laser
data within 5 minutes. In my experience, thermal can be used, but the brackets would be a sight to
changes of consequence occur after about 10 behold.
minutes. If hot alignment is more critical, [for
example, if pipe loads or shaft torque influence are 7. LIMITATIONS OF THE REVERSE-FACE
suspected to influence alignment], there are other METHOD
methods such as Vernier Alignment or Essinger
Bars that can be used. No method that requires Care must be taken when using the Reverse-Face
the shutdown of the machine to check alignment is Method on gear couplings or elastomeric
suitable, in the limit, for critical hot alignment couplings that can develop radial clearance due to
checks. wear. Refer to Photograph 2 for an example of
these two types of coupling on one machine. The
The calculation of the horizontal moves and radial clearance in the couplings must be
vertical shim changes is based on simple measured. (This is not a hardship, since the
geometry. A spreadsheet is shown in the clearance must be measured to monitor the wear
appendices that we use to do this calculation. in the coupling.) Then, the difference in the radial
clearance between ends, must be used to correct
6. REVERSE-FACE METHOD VERSUS the angularities calculated normally, A calculation
OTHER METHODS of the radial clearance-induced angularity at the
coupling flex planes can be done using simple
Laser Systems cost significantly more than a set geometry [angle = difference in the radial
of dial indicators, lack a calculated angularity at clearances/distance between flex-planes].
 4

The other limitation, as discussed above, is the middle of the coupling spool piece. However,
single flex-plane coupling, which requires the Rim- consider alignment limits based on parallel offset
Face Method. alone with no shaft angularity, and then shaft
angularity with no parallel offset. Either limit can
8. OPPORTUNITIES FOR ENHANCEMENTS be derived from flex-plane angularity
OF THE REVERSE-FACED METHOD considerations. In real alignments, the as-left
alignment will have a combination of offset and
If it is considered to improve the speed of data angularity. If both shaft offset and angularity limits
collection, or to automate the method, there are were reached at the same time, the resulting flex-
many things that could be done, as summarized plane angularity would be twice guideline.
below:
• special dial indicators with integral magnetic In other words, the guideline based on shaft offset
bases for specific coupling designs and angularity would have to be twice as strict to
• remote digital readout of dial indicators be equivalent in all cases to the flex-plane
• readings at 45 degree orientations, converted angularity guideline in Piotrowski’s chart.
to horizontal and vertical
• calculation on a spreadsheet with manual data 10. CONCLUSIONS
entry
• direct input to the computer using digital dial The author has found the Reverse-Face method to
indicators be a fast, accurate and inexpensive method of
• addition of a digital inclinometer doing alignment measurements. It is hoped that
• calculation of the angularities and corrections the reader will find the method to be useful, too.
based on less than 360 degree rotation of
11. REFERENCES
shafts
• calculation of shims at many planes
1) Piotrowski, John; “Shaft Alignment
• combination of cold and hot alignment data to
Handbook”; Marcel Dekker, Inc., New
calculate desired cold angularity
York and Base)
• error analysis calculation using tolerances on
dial readings. 2) Crease, A.B.; “Design Principles and
Lubrication of Gear Couplings”; Paper
The author has generated a spreadsheet to B1, International Conference on
calculate dial readings at orthogonal points [3:00, Flexible Couplings for High Powers and
6:00, 9:00] from dial readings at other angular Speeds, June, 1977.
locations. Again, this is a simple geometry
exercise. BIOGRAPHY:
9. GUIDELINES FOR ALIGNMENT AND Brian Howes is Chief Engineer for Beta Machinery
RELATED ISSUES Analysis Ltd., Calgary. His previous experience
includes: research and development in the area of
The original impetus to use the Reverse-Face pulsations and vibrations of reciprocating
Method was to make it easier to determine if a compressor piping systems, 28 years of
particular alignment was acceptable. Piotrowski’s troubleshooting problems, in many countries using
graph [see Figure 1] provides a strong argument a wide range of equipment including turbines,
for using angularity at each flex plane to determine centrifugal and plunger pumps, centrifugal, screw
acceptability. and reciprocating compressors, pulp refiners,
Gear couplings are a special case. The relative paper machines, ball mills, furnaces and piping
tooth velocity should be calculated to determine if systems. He has a Master of Science in Solid
the oil film will break down due to misalignment Mechanics from the University of Calgary, and is a
(maximum of 5 in/sec pk (Ref. 2)). On the other member of the Board of Directors of the Canadian
hand, minimum angularity is also required to Machinery Vibration Association (CMVA).
ensure that the oil will get between the teeth.

Many manufacturers and most Laser Alignment

systems use guidelines based on the offset and
angularity between the shaft centrelines at the
 5

Picture 1: Flex-pack coupling with a face dial mounted, showing interfering piping underneath coupling
[102795]

Picture 2: Gear coupling on the left, and Elastomeric coupling on the right [102847]
 6

Picture 3: A LoRez Coupling – single flex-plane – use Rim-Face Method

Figure 1: Alignment tolerance Chart (with polynomial curve fit equations) after Piotrowski
 7

APPENDICES

1. Spreadsheet to document Reverse-Face dial readings and calculate angularities.

2. Spreadsheet to calculate hot alignment angularities given the cold alignment and Essinger Bar data.
3. Spreadsheet to calculate shims and moves for many feet from Reverse-Face data

Appendix 1

Alignment Report
[Reverse Face Method]
Owner:
Location:
Unit: DeLaval HV-8
Coupling: Make: ? Speed 600 max
[rpm]
Model: ? Target angularity limit[thou/inch] 1.5 to 0.8 when hot
Date: August18/99 Machine temperature during readings: cold
Analyst: B. Howes Target cold alignment: compressor 10 Thou
higher than the engine
Dimension Diameter of dial circle: 26 inches
s:
Coupling shim pack spacing: 13 Inches
As left angularities: Vertical Horizontal Total
@Fixed machine shim pack: -0.83 -0.36 0.90 thou/inch
@Moved machine shim pack: 0.88 0.25 0.92 thou/inch
Guideline for angularity is a function of speed. Check book by Piotrowski for graph [page233].
Predicted hot angularities: Vertical Horizontal Total
@Fixed machine shim pack: -0.06 -0.36 0.36 thou/inch
@Moved machine shim pack: 0.12 0.25 0.28 thou/inch
Note: {The predicted hot alignment angularities are based on the assumption that the engine will rise
10 thou more than the compressor between cold stopped and hot running}
{The shaft motion within the bearing clearances is not included in these calculations.}

Dials on fixed machine side Dials on “moved machine” side

Dial “a”@ Dial “a” Dial “b” (a-b)/2 Dial “a”@ Dial “a” Dial “b” (a-b)/2
12:00 0 0 0 12:00 0 0 0
3:00 -3 7 -5 0 3:00 3.5 -14 8.75 0
6:00 -25 18 -21.5 6:00 16 -30 23
9:00 -21 7.5 -14.25 -9.25 9:00 12.5 -18 15.25 6.5
12:00 0 -1 0.5 12:00 0 0 0

sum of vert -21.5 sum of vert 23

sum of hor -19.25 sum of hor 24
closure 0.5 closure 0
fmy -21.5 -ve = open at bottom mmy 23 +ve = closed at bottom
fmx -9.25 -ve = open on left mmx 6.5 +ve = closed on left

sum of vertical should be equal to sum of horizontal

closure [equivalent dial reading at second 12:00] should be zero
equivalent dial reading [compensated for float] is (a-b)/2
fmy, fmx, mmy, mmx are the face results at 6:00 and 9:00 for the 2 planes
 8

Appendix 2

Essinger Bar readings to angularities at the power transmission points of a spool piece coupling
Refer to “Shaft Alignment Handbook”by Piotrowski for more details and
A graph on allowable alignment angularity Brian Howes, December, 1999

Unit: Boiler Feedwater #1 Location:

Date: Nov29/99 to Dec 2/99 Condition: As found
Dimensions [inches] up/right
Motor end L1 L2 L3 L4 L5 Pump end
77 10 7 9 81 down/left
Essinger Bar CHANGES FROM COLD TO HOT [thou]
E1 E2 E3 E4
Vertical 0.4 2.1 y1, y2 ,y3, y4 1.8 -0.7 +ve = up
Horizontal 0.3 0.6 X1, x2, x3, x4 -5.7 -0.7 +ve = right
MOB MIB PIB POB
Calculated position of shaft centerline at the flex planes after cold to hot change
Motor shaft PUMP SHAFT
V1 @ F1 2.32 V2 @ F2 2.08
Vertical V3 @ F2 2.48 V4 @ F1 2.29 +ve = up
Horizontal H1 @ F1 0.64 H2 @ F2 -6.26 +ve = right
H3 @ F2 0.67 H4 @ F1 -6.69

Calculated angles at the shim packs [flex points, power transmission p oints]
Motor end Pump end
Vertical Alpha1 -0.06 Alpha2 0.00 angle units are thou/inch
Horizontal Beta1 -0.99 Beta2 1.05 1 thou/inch = 1 milliradian
Total Gamma1 0.99 Gamma2 1.05
Guideline 0 0.5 Speed 3600 [rpm]
minimum maximum

Now assume the motor is set 6 thou lower than the pump,
Cold
Calculated position of shaft centerline after the change
From cold to hot at the flex planes
Motor shaft PUMP SHAFT
V1 @ F1 -3.68 V2 @ F2 2.08
Vertical V3 @ F2 -3.52 V4 @ F1 2.29 +ve = up
Horizontal H1 @ F1 0.64 H2 @ F2 -6.26 +ve = right
H3 @ F2 0.67 H4 @ F1 -6.69

Calculated angles at the shim packs [flex points, power transmission points]
Motor end Pump end
Vertical Alpha1 0.80 Alpha2 -0.85 angle units are thou/inch
Horizontal Beta1 -0.99 Beta2 1.05 1 thou/inch = 1 milliradian
Total Gamma1 1.27 Gamma2 1.35
Guideline 0 0.5 Thou/inch at Speed = 3600 [rpm]
minimum Maximum
 9

Appendix 3

Reverse-Face Alignment Calculation Sheet

[For unit where driver is the fixed machine]

Date: Time: Customer: Unit:

Reverse-Face Dial Readings

fixed machine=driver machine to be moved
a@ a b (a-b)/2 A@ a b (a-b)/
12 0 0 0 12 0 0 0
9 9
6 6
3 3
12 12
Dial Data Quality Check
sum of vert 0 sum of vert 0
sum of hor 0 sum of hor 0
closure 0 [from 12 to 12] closure 0
[The sums in the vertical and the horizontal should be equal]
[Closure of the dial readings at 12:00 should be zero]
Fmy 0 fmy 0
fmx 0 fmx 0
fmy fmx
Fixed fm* 0 0
Move fm* 0 0

DIA 26 SHIMS MOVES

CPLG 13 [thou] [thou] [fmy is the bottom less the top
L1 38 0 0 Dial reading after axial float compensation]
L2 73 0 0 [fmx is the right less the left
L3 53 0 0 Dial reading after axial float compensation]
L4 88 0 0
L5 33 0 0
L6 49 0 0
L7 74 0 0
L8 98 0 0
Lengths Ln are from the flex-plane closest to the fixed machine
“Fixed fm*” and “Move fm*” are the values of fmy or fmx at the flex-planes
on the fixed machine and the machine to be moved sides respectively.
Lengths Ln and Diameter DIA in inches CPLG, inches between flex-planes
Sign Conventions
looking from driven machine to driver
positive y is up
positive x is right
reference vertical readings to 0 @ 12 o’clock
reference horizontal readings to 0 @ 9 o’clock
Positive shim value means add shim