AGMA 912- A04

AMERICAN GEAR MANUFACTURERS ASSOCIATION

Mechanisms of Gear Tooth Failures

AGMA 912- A04

AGMA INFORMATION SHEET

(This Information Sheet is NOT an AGMA Standard)
 American Mechanisms of Gear Tooth Failures
AGMA 912--A04
Gear
Manufacturers CAUTION NOTICE: AGMA technical publications are subject to constant improvement,
revision or withdrawal as dictated by experience. Any person who refers to any AGMA
Association
technical publication should be sure that the publication is the latest available from the As-
sociation on the subject matter.
[Tables or other self--supporting sections may be referenced. Citations should read: See
AGMA 912--A04, Mechanisms of Gear Tooth Failures, published by the American Gear
Manufacturers Association, 500 Montgomery Street, Suite 350, Alexandria, Virginia
22314, http://www.agma.org.]
Approved October 23, 2004
ABSTRACT
This information sheet describes many of the ways in which gear teeth can fail and recommends methods for
reducing gear failures. It provides basic guidance for those attempting to analyze gear failures. It should be
used in conjunction with ANSI/AGMA 1010--E95 in which the gear tooth failure modes are defined. They are
described in detail to help investigators understand failures and investigate remedies. This information sheet
does not discuss the details of disciplines such as dynamics, material science, corrosion or tribology. It is
hoped that the material presented will facilitate communication in the investigation of gear operating problems.
Published by

American Gear Manufacturers Association

500 Montgomery Street, Suite 350, Alexandria, Virginia 22314
Copyright  2004 by American Gear Manufacturers Association
All rights reserved.
No part of this publication may be reproduced in any form, in an electronic
retrieval system or otherwise, without prior written permission of the publisher.
Printed in the United States of America
ISBN: 1--55589--838--6

ii
AMERICAN GEAR MANUFACTURERS ASSOCIATION AGMA 912--A04

Contents
Page
Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
1 Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2 Normative references . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
3 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
4 Wear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
5 Scuffing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
6 Plastic deformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
7 Contact fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
8 Cracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
9 Fracture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
10 Bending fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Tables
1 Fracture appearance classifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

 AGMA 2004 ---- All rights reserved iii

AGMA 912--A04 AMERICAN GEAR MANUFACTURERS ASSOCIATION

Foreword
[The foreword, footnotes and annexes, if any, in this document are provided for
informational purposes only and are not to be construed as a part of AGMA Information
Sheet 912--A04, Mechanisms of Gear Tooth Failures.]

AGMA Standard 110.01 was first published in October 1943 as means to document the
appearance of gear teeth when they wear or fail. The study of gear tooth wear and failure
has been hampered by the inability of two observers to describe the same phenomenon in
terms that are adequate to assure uniform interpretation. AGMA Standard 110.02 became a
national standard, B6.12, in 1954. A revised standard with photographs, AGMA 110.03,
was published in 1960. The last version, AGMA 110.04, was published in 1979 and
reaffirmed by the members in 1989, with improved photographs and additional material.

ANSI/AGMA 1010--E95, approved December 1995, is a revision of AGMA 110.04. It

provides a common language to describe gear wear and failure, and serves as a guide to
uniformity and consistency in the use of that language. It describes the appearance of gear
tooth failure modes and discusses their mechanisms, with the sole intent of facilitating
identification of gear wear and failure. Since there may be many different causes for each
type of gear tooth wear or failure mode, it does not standardize cause, nor prescribe
remedies.

AGMA 912--A04 was developed to compliment ANSI/AGMA 1010--E95 with some

information on probable cause and recommendations for remedies. Gear design and
failure analysis are both art and science. To design gears, the gear engineer needs
analytical tools, plus practical field experience. Gear failures can be a part of this
experience. They can provide valuable information and their correct analysis can help find
the correct remedy to reduce future problems.

The first draft of AGMA 912--A04 was developed in October, 1995. It was approved by the
AGMA membership on October 23, 2004.

Suggestions for improvement of this document will be welcome. They should be sent to the
American Gear Manufacturers Association, 500 Montgomery Street, Suite 350, Alexandria,
Virginia 22314.

iv  AGMA 2004 ---- All rights reserved

AMERICAN GEAR MANUFACTURERS ASSOCIATION AGMA 912--A04

PERSONNEL of the AGMA Nomenclature Committee

Chairman: Dwight Smith . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cole Manufacturing Systems, Inc.

ACTIVE MEMBERS

M. Chaplin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contour Hardening, Inc.

R. Errichello . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . GEARTECH
T. Miller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CST -- Cincinnati
G.W. Nagorny . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nagorny & Associates
J. Rinaldo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Atlas Copco Compressors, Inc.
O. LaBath . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gear Consulting Services of Cincinnati, LLC

ASSOCIATE MEMBERS

A.S. Cohen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Engranes y Maquinaria Arco, S.A.

R. Green . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . R7 Group
H. Hagiwara . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nippon Gear Company, Ltd.
I. Laskin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Consultant
E. Lawson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M&M Precision Systems Corporation
D.A. McCarroll . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ZF Industries
D.R. McVittie . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gear Engineers, Inc.
L.J. Smith . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Invincible Gear Company
R.E. Smith . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . R.E. Smith & Company, Inc.
D. Woodley . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Texaco Lubricants Company

 AGMA 2004 ---- All rights reserved v

AGMA 912--A04 AMERICAN GEAR MANUFACTURERS ASSOCIATION

(This page is intentionally left blank.)

vi  AGMA 2004 ---- All rights reserved

AMERICAN GEAR MANUFACTURERS ASSOCIATION
American Gear Manufacturers
Association --
Mechanisms of Gear
Tooth Failures
1 Scope
This information sheet describes many of the ways in
which gear teeth can fail and recommends methods
for reducing gear failures. It provides basic guidance
for those attempting to analyze gear failures. The
information sheet should be used in conjunction with
ANSI/AGMA 1010--E95 in which the gear tooth
failure modes are defined. Similar definitions can
also be found in ISO 10825. They are described in
detail to help investigators understand failures and
investigate remedies.
The information presented in this document applies
to spur and helical gears. However, with some
exceptions the information also applies to bevel,
worm and hypoid gears. Discussion of material
properties is primarily restricted to steel.
1.1 System investigations
Gear system dynamic problems are beyond the
scope of this information sheet. However, it is
important to recognize that many gear failures are
influenced by problems with the gear system, such
as high loads caused by vibration. When investigat-
ing gear failures, it is necessary to consider that the
cause may stem from a problem with the system
rather than the gears.
1.2 Analysis by specialists
It is not the intent of this information sheet to discuss
the details of disciplines such as dynamics, material
science, corrosion or tribology. It is hoped that the
material presented will facilitate communication in
the investigation of gear problems.
 AGMA 2004 ---- All rights reserved
AGMA 912--A04
2 Normative references
The following standards contain provisions which
are referenced in the text of this information sheet.
At the time of publication, the editions indicated were
valid. All standards are subject to revision, and
parties to agreements based on this document are
encouraged to investigate the possibility of applying
the most recent editions of the standards indicated.
ANSI/AGMA 1010--E95, Appearance of Gear Teeth
-- Terminology of Wear and Failure
ISO 10825:1995, Gears -- Wear and damage to
gear teeth -- Terminology
3 Analysis
3.1 Failure experience
Gear design is both an art and a science. To design
better gears, the gear engineer needs good analyti-
cal tools plus practical field experience. Gear
failures are a part of this experience because they
provide valuable information about the multitude of
failure modes that can occur. Gear failures should
be analyzed to identify the failure mode, and attempt
to determine the cause of the failure. Failure
analysis can help to find the correct remedy to
reduce future problems.
3.2 Quantitative analysis
Gear “failure” is frequently subjective. For example,
a person observing gear teeth that have a bright,
mirror finish may think that the gears have “run--in”
nicely. However, another observer may believe that
the gears are wearing by polishing. Whether the
gears should be considered usable or not depends
on how much wear is tolerable. The gears might be
unusable if the wear causes excessive noise or
vibration. But the word “excessive” in itself is
subjective, and some measure of gear accuracy,
noise or vibration can be used to resolve whether the
gears are usable. Some failures are more obvious,
such as when several gear teeth fracture and the
transmission of power ceases. In these cases the
gears have failed. However, there may not be
agreement on the cause of the failure (failure mode).
To find the basic cause or causes of a failure, one
1
AGMA 912--A04
must discern the difference between primary and
secondary failure modes. Bending fatigue may be
the ultimate failure mode. However, it is often a
consequence of some other mode of failure, such as
scuffing or macropitting. Because multiple failure
modes can occur concurrently, the primary mode of
failure often can only be observed in its early stages
before it is masked by secondary, competing failure
modes.
Failure modes vary in significance. For example,
contact fatigue is often less serious than bending
fatigue. This is because contact fatigue usually
progresses relatively slowly, starting with a few pits
which increase in size and number. As the teeth
deteriorate, the gears may generate noise or vibra-
tion which warns of an impending failure. In contrast,
bending fatigue breaks a tooth with little warning.
It is often helpful to monitor the operating gear
system by measuring temperature, noise and vibra-
tion, analyzing the lubricant for contamination, or by
visual inspection of the gear teeth. These actions
may help to warn of failure before it occurs.
3.3 How to analyze gear failures
3.3.1 Failure conditions
When gears fail, there may be incentive to quickly
repair or replace failed components and return the
gear system to service. However, because gear
failures provide valuable data that may help prevent
future failures, a systematic inspection procedure
should be followed before repair or replacement
begins.
The failure investigation should be carefully planned
to preserve evidence. The specific approach can
vary depending on when and where the inspection is
made, the nature of the failure, and time constraints.
3.3.1.1 When and where
Ideally, the site visit and failed components should
be inspected as soon after failure as possible. If an
early inspection is not possible, someone at the site
must preserve the evidence based on specific
instructions.
3.3.1.2 Nature of failure
The failure conditions can determine when and how
to conduct an analysis. It is best to shutdown a failing
gear unit as soon as possible to limit damage. To
preserve evidence, carefully plan the failure inves-
2  AGMA 2004 ---- All rights reserved
AMERICAN GEAR MANUFACTURERS ASSOCIATION
tigation including shutdown, in--situ inspections,
gear unit removal, transport, storage, and disassem-
bly. However, if the gears are damaged but still
functional, the company may decide to continue
operation and monitor damage progression. In this
case, the gear system should be monitored under
experienced supervision. For critical applications,
examine the gears with magnetic particle or dye
penetrant inspection to ensure there are no cracks
before operation is continued. In all applications,
check for damage by visual inspection and by
measuring temperature, sound, and vibration.
Collect samples of lubricant for analysis, drain and
flush the reservoir, and replace the lubricant.
Examine the oil filter for wear debris and
contaminants, and inspect magnetic plugs for wear
debris.
3.3.1.3 Time constraints
In some situations, the high cost of shutdown limits
time available for inspection. Such cases call for
careful planning. For example, dividing tasks
between two or more analysts reduces time
required.
3.3.2 Prepare for inspection
Before visiting the failure site, interview on--site
personnel and explain what is needed to inspect the
gear unit including personnel, equipment, and
working conditions.
Request a skilled technician to disassemble the
equipment. However, make sure that no work is
done on the gear unit until it can be observed. This
means no disassembly, cleaning, or draining of the
oil. Otherwise, a well--meaning technician could
inadvertently destroy evidence. Emphasize that
failure investigation is different from a gear unit
rebuild, and the disassembly must be carefully
controlled.
Verify that gear unit drawings, disassembly tools,
and adequate facilities are available. Inform the site
supervisor that privacy is required to conduct the
investigation and access is needed to all available
information.
Obtain as much background information as pos-
sible, including manufacturer’s specifications, ser-
vice history, load data, and lubricant analyses. Send
a questionnaire to the site personnel to help expedite
information gathering.
AMERICAN GEAR MANUFACTURERS ASSOCIATION
3.3.3 Inspect in--situ
Before starting the inspection, review background
information and service history with the contact
person. Then interview those involved in design,
installation, startup, operation, maintenance, and
failure of the gear unit. Encourage them to tell
everything they know about the gear unit even if they
feel it is not important.
3.3.3.1 External examination
Before removing and disassembling the gear unit,
thoroughly inspect its exterior. Use an inspection
form to record important data that would otherwise
be lost once disassembly begins. For example, the
condition of seals and keyways must be recorded
before disassembly. Otherwise, it may be impossi-
ble to determine when these parts were damaged.
Before cleaning the exterior of the gear housing,
inspect for signs of overheating, corrosion,
contamination, oil leaks, and damage.
3.3.3.2 Gear tooth contact patterns
Clean the inspection port cover and the immediate
area around it. Remove the cover being careful not
to contaminate the gear unit. Observe the condition
of gears, shafts, and bearings.
The way gear teeth contact indicates how they are
aligned. Record tooth contact patterns under loaded
or unloaded conditions. No--load patterns are not as
reliable as loaded patterns for detecting misalign-
ment because marking compound is relatively thick
and no--load tests do not include misalignment
caused by load, speed, or temperature. Therefore,
follow no--load tests with loaded tests whenever
possible.
See ISO/TR 10064--4:1998, clause 9 for information
regarding contact pattern tests.
3.3.3.3 No--load contact patterns
For no--load tests, paint the teeth of one gear with
soft marking compound and roll the teeth through
mesh so compound transfers to the unpainted gear.
Turn the pinion by hand while applying a light load to
the gear shaft by hand or brake. Lift transferred
patterns from the gear with clear tape and mount the
tapes on white paper to form a permanent record.
 AGMA 2004 ---- All rights reserved
AGMA 912--A04
3.3.3.4 Loaded contact patterns
For loaded tests, paint several teeth on one or both
gears with machinist’s layout lacquer. Thoroughly
clean teeth with solvent and acetone, and paint with
a thin coat of lacquer. Run the gears under load for
sufficient time to wear off the lacquer and establish
the contact patterns. Photograph patterns to obtain a
permanent record.
Record loaded contact patterns under several loads,
for example, 25%, 50%, 75%, and 100% load.
Inspect patterns after running about one hour at
each load to monitor how patterns change with load.
Ideally, the patterns should not change much with
load. Optimum contact patterns cover nearly 100%
of the active face of gear teeth under full load, except
at extremes of teeth along tips, roots, and ends,
where contact is lighter as evidenced by traces of
lacquer.
3.3.3.5 Endplay and backlash
Inspect endplay, radial movement of the input and
output shafts, and gear backlash.
3.3.4 Remove gear unit
3.3.4.1 Mounting alignment
Measure alignment of shaft couplings before remov-
ing the gear unit. Note the condition and loosening
torque of all fasteners including coupling and mount-
ing bolts. Check for possible twist of the gear
housing by measuring any movement of the mount-
ing feet as mounting bolts are loosened. Install four
dial indicators, one at each corner of the gear unit.
Each indicator should record the same vertical
movement if there is no twist. If not, calculate the
twist from relative movements.
3.3.5 Transport gear unit
Fretting corrosion is a common problem that may
occur during shipping. Ship the gear unit on an
air--ride truck, and support the gear unit on vibration
isolators to help avoid fretting corrosion. If possible,
ship the gear unit with oil. To minimize contamina-
tion, remove the breather and seal the opening, seal
labyrinth seals with silicone rubber, and cover the
gear unit with a tarpaulin.
3.3.6 Store gear unit
It is best to inspect the gear unit as soon as
possible. However, if the gear unit must be stored,
store it indoors in a dry, temperature controlled
environment.
3
AGMA 912--A04
3.3.7 Disassemble gear unit
Explain the objectives to the technician who will be
doing the work. Review the gear unit assembly
drawings with the technician, checking for potential
disassembly problems. Verify the work will be done
in a clean, well--lighted area, protected from the
elements, and all necessary tools are available. If
working conditions are not suitable, find an alternate
location for gear unit disassembly.
NOTE: Unless the technician is familiar with the proce-
dure, it is wise to remind him that disassembly must be
done slowly and carefully (technicians are usually
trained to work quickly).
After the external examination, thoroughly clean the
exterior of the gear unit to avoid contaminating the
gear unit when opening it. Measure all tapered roller
bearing endplays before disassembling the gear
unit, since excessive endplay can be the cause of
gear misalignment. Disassemble the gear unit and
inspect all components, both failed and undamaged.
3.3.8 Inspect components
3.3.8.1 Inspect before cleaning
Mark relative positions of all components before
removing them. Do not throw away or clean any
parts until they are examined thoroughly. If there are
broken components, do not touch fracture surfaces
or fit broken pieces together. If fractures cannot be
examined immediately, coat them with oil and store
the parts so fracture surfaces are not damaged.
Examine functional surfaces of gear teeth and
bearings and record their condition. Before cleaning
the parts, look for signs of corrosion, contamination,
and overheating.
3.3.8.2 Inspect after cleaning
After the initial inspection, wash the components
with solvents and re--examine them. This examina-
tion should be as thorough as possible because it is
often the most important phase of the investigation
and may yield valuable clues. A low power
magnifying glass and 30X pocket microscope are
helpful tools for this examination.
It is important to inspect bearings because they often
provide clues as to the cause of gear failure. For
example:
-- bearing wear can cause excessive radial
clearance or endplay that misaligns gears;
4  AGMA 2004 ---- All rights reserved
AMERICAN GEAR MANUFACTURERS ASSOCIATION
-- bearing damage may indicate corrosion,
contamination, electrical discharge, or lack of
lubrication;
-- plastic deformation between rollers and
raceways may indicate overloads;
-- gear failure often follows bearing failure.
3.3.8.3 Document observations
Identify and mark each component (including gear
teeth and bearings), so that it is clearly identified by
written descriptions, sketches, and photographs. It
is especially important to mark all bearings, including
inboard and outboard sides, so their location and
position in the gear unit is identified.
Describe components consistently. For example,
always start with the same part of a bearing and
progress through the parts in the same sequence.
This helps to avoid overlooking any evidence.
Describe important observations in writing using
sketches and photographs where needed. The
following guidelines are to help maximize chances
for obtaining meaningful evidence:
-- Concentrate on collecting evidence, not on
determining cause of failure. Regardless of how
obvious the cause may appear, do not form
conclusions until all evidence is considered.
-- Document what you see. List all observations
even if some seem insignificant or if you don’t
recognize the failure mode. Remember there is
a reason for everything, and it may become
important later when considering all the evidence.
-- Document what is not observed. This is
helpful to eliminate certain failure modes and
causes. For example, if there is no scuffing, it can
be concluded that gear tooth contact tempera-
tures were less than the scuffing temperature of
the lubricant.
-- Search the bottom of the gear unit. Often this
is where the best preserved evidence is found,
such as when a tooth fractures and falls free
without secondary damage.
-- Prepare for the inspection. Plan work careful-
ly to obtain as much evidence as possible.
-- Control the investigation. Watch every step of
the disassembly. Don’t let the technician get
ahead of the inspection. Disassembly should
stop while inspecting and documenting the condi-
tion of a component, then proceed to the next
component.
AMERICAN GEAR MANUFACTURERS ASSOCIATION
-- Insist on privacy. Do not be distracted. If
asked about conclusions, answer that they can-
not be formed until the investigation is complete.
3.3.8.4 Gather gear geometry
The load capacity of the gears should be calculated.
For this purpose, obtain the following geometry data,
from the gears and housing or drawings:
-- number of teeth;
-- outside diameter;
-- face width;
-- gear housing center distance;
-- whole depth of teeth;
-- tooth thickness (both span and topland
thickness).
3.3.8.5 Specimens for laboratory tests
During inspection, hypotheses regarding the cause
of failure will begin to formulate. With these
hypotheses, select specimens for laboratory testing.
Take broken parts for laboratory evaluation or, if this
is not possible, preserve them for later analysis.
After completing the inspection, be sure all parts are
coated with oil and stored properly so that corrosion
or damage will not occur.
Oil samples can be very helpful. However, an
effective analysis depends on how well the sample
represents the operating lubricant. To take samples
from the gear unit drain valve, first discard stagnant
oil from the valve. Then take a sample at the start,
middle, and end of the drain to avoid stratification. To
sample from the storage drum or reservoir, draw
samples from the top, middle, and near the bottom.
These samples can uncover problems such as
excessive water in the oil due to improper storage.
Ask if there are new unused components. These are
helpful to compare with failed parts. Similarly,
compare a sample of fresh lubricant to used
lubricant.
3.3.8.6 Obtain all items
Before leaving the site, make sure that everything
needed including completed inspection forms, writ-
ten descriptions and sketches, photos, and test
specimens are obtained.
It is best to devote two days minimum for the failure
inspection. This affords time after the first day’s
inspection to collect thoughts and analyze collected
 AGMA 2004 ---- All rights reserved
AGMA 912--A04
data. Often the first day’s inspection discloses a
need for other data, which can be gathered on the
second day.
3.3.9 Determine failure mode
When several failure modes are present, the primary
mode needs to be identified. Other modes may be
consequences of the primary mode. These may or
may not have contributed to the failure. There may
also be evidence of other independent problems that
did not contribute to the failure.
The classes of gear failure modes to be discussed
are:
-- wear, see clause 4;
-- scuffing, see clause 5;
-- plastic deformation, see clause 6;
-- Hertzian (contact) fatigue, see clause 7;
-- cracking, see clause 8;
-- fracture, see clause 9;
-- bending fatigue, see clause 10.
An understanding of these modes will assist in
identifying the cause of failure.
3.3.10 Calculations and tests
In many cases, failed parts and inspection data do
not yield enough information to determine the cause
of failure. When this happens, gear design calcula-
tions and laboratory tests may be needed to develop
and confirm a hypothesis for the probable cause.
3.3.10.1 Gear design calculations
Gear geometry data aids in estimating tooth contact
stress, bending stress, lubricant film thickness, and
gear tooth contact temperature based on trans-
mitted loads. Calculate values according to ap-
propriate rating method standards such as
ANSI/AGMA 2001--C95. Compare calculated val-
ues with allowable values to help determine risks of
micropitting, macropitting, bending fatigue, and
scuffing.
3.3.10.2 Laboratory examination and tests
Microscopic examination may confirm the failure
mode or find the origin of a fatigue crack. Light
microscopes and scanning electron microscopes
(SEM) are useful for this purpose. A SEM with
energy dispersive X--ray is especially useful for
identifying corrosion, contamination, or inclusions.
If the primary failure mode is likely to be influenced
by gear geometry or metallurgical properties, check
5
AGMA 912--A04
for any geometric or metallurgical defects that may
have contributed to the failure. For example, if tooth
contact patterns indicate misalignment or interfer-
ence, inspect the gear for accuracy on gear inspec-
tion machines. Conversely, where contact patterns
indicate good alignment and loads are within rated
gear capacity, check teeth for metallurgical defects.
Conduct nondestructive tests before any destructive
tests. These nondestructive tests, which aid in
detecting material or manufacturing defects and
provide rating information, include:
-- surface hardness and roughness;
-- magnetic particle or dye penetrant inspection
for cracks;
-- acid etch inspection for surface temper;
-- gear tooth accuracy inspection.
Then conduct destructive tests to evaluate material
and heat treatment. These tests include:
-- microhardness survey;
-- microstructural determination using acid
etches;
-- determination of grain size;
-- determination of nonmetallic inclusions;
-- SEM microscopy to study fracture surfaces.
3.3.11 Form and test conclusions
When all calculations and tests are completed, one
or more hypotheses for the probable cause of failure
should be formed, then determine if the evidence
supports or disproves the hypotheses. Evaluate all
evidence that was gathered including:
-- documentary evidence and service history;
-- statements from witnesses;
-- written descriptions, sketches, and photos;
-- gear geometry and contact patterns;
-- gear design calculations;
-- laboratory data for materials and lubricant.
Results of this evaluation may make it necessary to
modify or abandon initial hypotheses, or pursue new
lines of investigation.
Finally, after thoroughly testing the hypotheses
against the evidence, reach a conclusion about the
most probable cause of primary failure. In addition,
identify secondary factors that may have contributed
to the failure.
6  AGMA 2004 ---- All rights reserved
AMERICAN GEAR MANUFACTURERS ASSOCIATION
3.3.12 Report results
The failure analysis report should describe all
relevant facts found during analysis, inspections and
tests, weighing of evidence, conclusions, and rec-
ommendations. Present data succinctly, preferably
in tables or figures.
Good photos are especially helpful for portraying
failure characteristics. If possible, include recom-
mendations for repairing equipment, or making
changes in equipment design or operation to prevent
future failures.
3.4 Modes of failure
ANSI/AGMA 1010--E95 provides nomenclature for
modes of gear failure. The gear failure modes are
discussed and detailed.
This information sheet provides additional informa-
tion on gear tooth failures, causes and remedies.
Also see references in clause 2 and the bibliography
for additional information on gear failure modes and
lubrication related failures.
4 Wear
4.1 Adhesion
Adhesive wear is classified as “mild” if it is confined
to the oxide layers on the gear tooth surfaces. If,
however, the oxide layers are disrupted and bare
metal is exposed, the transition to severe adhesive
wear (scuffing) may occur. Scuffing is discussed in
clause 5. For the present, it is assumed that scuffing
has been avoided.
When new gear units are first operated the contact
between the gear teeth may not be optimum
because of unavoidable manufacturing inaccura-
cies. If the tribological conditions are favorable, mild
adhesive wear occurs during running--in and sub-
sides with time, resulting in a satisfactory lifetime for
the gears. The wear that occurs during running--in is
beneficial if it creates smooth tooth surfaces (in-
creasing the specific film thickness) and increases
the area of contact by removing minor imperfections
through local wear. It is recommended that new
gearsets be run--in by operating for at least the first
10 hours at one--half load.
The amount of wear that is considered tolerable
depends on the expected lifetime for the gears and
requirements for the control of noise and vibration.
AMERICAN GEAR MANUFACTURERS ASSOCIATION
The wear is considered excessive when the tooth
profiles wear to the extent that high dynamic loads
are encountered or the tooth thickness is reduced to
the extent that bending fatigue becomes possible.
Some gear units operate under ideal conditions with
smooth tooth surfaces, high pitchline speed, and
high lubricant film thickness. It has been observed,
for example, that turbine gears that operated almost
continuously at 150 m/s pitchline speed still had the
original machining marks on their teeth even after
operating for 20 years. Most gears however, operate
between the boundary and full--film lubrication
regimes, under elastohydrodynamic (EHD) condi-
tions. In the EHD regime, provided that the proper
type and viscosity of lubricant is used, the wear rate
usually reduces during running--in and adhesive
wear virtually ceases once running--in is completed.
If the lubricant is properly maintained (kept cool,
clean and dry) the gearset should not suffer an
adhesive wear failure.
Many gears, because of practical limits on lubricant
viscosity, speed and temperature, must operate
under boundary--lubricated conditions where some
wear is inevitable. Highly--loaded, slow speed (less
than 0.5 m/s pitchline velocity), boundary--lubricated
gears are especially prone to excessive wear. Tests
with slow--speed gears [1] have shown that nitrided
gears have good wear resistance while carburized
and through--hardened gears have similar, lower
wear resistance. Reference [1] concluded that
lubricant viscosity has a large influence on slow--
speed, adhesive wear. It found that high viscosity
lubricants reduce the wear rate significantly. It also
found that some very aggressive additives that
contain sulphur--phosphorous extreme pressure
additives can be detrimental with very slow--speed
(less than 0.05 m/s) gears, giving higher wear rates
than expected.
Methods for reducing adhesive wear
-- Use smooth tooth surfaces;
-- Run--in new gearsets by operating the first 10
hours at one--half load;
-- Use high speeds if possible. Highly--loaded,
slow--speed gears are boundary lubricated and
especially prone to excessive wear;
-- For very slow--speed gear (less than 0.05
m/s), use lubricants with no sulphur--phospho-
rous additives or those additives that have proven
to be less aggressive to the tooth surfaces;
 AGMA 2004 ---- All rights reserved
AGMA 912--A04
-- Use an adequate amount of cool, clean and
dry lubricant of the highest viscosity permissible
for the operating conditions;
-- Use nitrided gears if they have adequate ca-
pacity.
4.2 Abrasion
Abrasive wear on gear teeth is usually caused by
contamination of the lubricant by hard, sharp--edged
particles. Contamination enters gear units by being
built--in, internally--generated, ingested through
breathers and seals, or inadvertently added during
maintenance.
Sand, machining chips, grinding dust, weld splatter
or other debris may find their way into new gear units.
To remove built--in contamination, it is generally
worthwhile to drain and flush the gearbox lubricant
after the first 50 hours of operation, refill with the
recommended lubricant, and install a new oil filter.
Internally--generated particles are usually wear
debris from gears, bearings or other components
due to Hertzian (contact) fatigue, macropitting, or
adhesive and abrasive wear. The wear particles can
be abrasive because they become work hardened
when they are trapped between the gear teeth.
Internally--generated wear debris can be minimized
by using accurate, surface--hardened gear teeth
(with high macropitting resistance), smooth tooth
surfaces and clean high viscosity lubricants.
Magnetic plugs may be used to capture ferrous
particles that are present at startup, or are generated
during operation. Periodic inspection of the magnet-
ic plug may be used to monitor the development of
ferrous particles during operation. Magnetic wear
chip detectors with alarms are also available.
The lubrication system should be carefully main-
tained and monitored to ensure that the gears
receive an adequate amount of cool, clean and dry
lubricant. For circulating--oil systems, fine filtration
helps to remove contamination. Filters as fine as 3
micrometers have been used to significantly in-
crease gear life, where the pressure loss in the filter
can be tolerated. The lubricant may have to be
changed or processed to remove water and maintain
additive levels. For oil--bath gear units, the lubricant
should be changed frequently because it is the only
way to remove contamination. In many cases the
lubricant should be changed at least every 2500
operating hours or six months, whichever occurs
first. For critical gear units a regular program of
lubricant monitoring can be used to show when
maintenance is required. The lubricant monitoring
7
AGMA 912--A04
may include such items as spectrographic and
ferrographic analysis of contamination along with
analysis of acidity, viscosity, and water content.
Used filter elements may be examined for wear
debris and contaminants.
Kidney--loop type systems may also be used to clean
oil. Electrostatic agglomeration systems may be
used to reduce the amount of very fine particles that
normally would pass through the filters. Other
systems may be used to remove water from the oil.
Breather vents are used on gear units to vent internal
pressure which occurs when air enters through seals
or when the air within the gearbox expands and
contracts during normal heating and cooling. The
breather vent should be located in a clean, non--
pressurized area and it should have a filter to prevent
ingression of airborne contaminants. In especially
harsh environments, the gearbox can sometimes be
completely sealed, and the pressure variation can be
accommodated by an expansion chamber with a
flexible diaphragm.
All maintenance procedures which involve opening
any part of the gear unit or lubrication system should
be carefully performed in a clean environment to
prevent contamination of the gear unit.
Abrasive wear due to foreign contaminants such as
sand or internally--generated wear debris is called
three body abrasion. Two body abrasion occurs
when hard particles or asperities on one gear tooth
abrade the opposing tooth surface. Unless the tooth
surfaces of a surface--hardened gear are smoothly
finished, they may act like files if the mating gear is
appreciably softer. This is the reason that a worm is
polished after grinding before it is run with a bronze
worm gear.
Methods for reducing abrasive wear
-- Flush unit thoroughly before initial operation;
-- Remove built--in contamination from new
gear units by draining and flushing the lubricant
after the first 50 hours of operation. Refill with
clean recommended lubricant and install a new
filter;
-- Minimize internally--generated wear debris
by using smooth tooth surfaces and high viscosity
lubricants;
-- Minimize ingested contamination by main-
taining oil--tight seals and using filtered breather
vents located in clean, non--pressurized areas;
8  AGMA 2004 ---- All rights reserved
AMERICAN GEAR MANUFACTURERS ASSOCIATION
-- Minimize contamination that is added during
maintenance by using good housekeeping
procedures;
-- For circulating--oil systems, use fine filtration;
-- Use an agglomeration system to remove very
fine particles;
-- Change or process the lubricant to remove
water;
-- For oil--bath systems, change the lubricant at
least every 2500 hours or every six months, or as
recommended by the manufacturer;
-- Monitor the lubricant with spectrographic and
ferrographic analysis together with analysis of
acidity, viscosity and water content.
4.3 Polishing
The gear teeth may polish to a bright, mirror--like
finish if the anti--scuff additives in the lubricant are
too chemically aggressive, or a fine abrasive is
present. Although the polished gear teeth may look
good, polishing wear can be undesirable if it reduces
gear accuracy by wearing the tooth profiles away
from their ideal form. Anti--scuff additives such as
sulfur and phosphorous are used in lubricants to
prevent scuffing (they will be covered when scuffing
is discussed). Ideally, the additives should react only
at temperatures where there is a danger of welding.
If the rate of reaction is too high, and there is a
continuous removal of the surface films caused by
very fine abrasives in the lubricant, polishing wear
may become excessive.
Polishing wear can be prevented by using less
chemically active additives and clean oil. The
anti--scuff additives should be appropriate for the
service conditions. The use of any dispersed materi-
al, such as some anti--scuff additives, should be
monitored since it may precipitate or be filtered out.
The abrasives in the lubricant should be removed by
using fine filtration or frequent oil changes.
Methods for reducing polishing wear
-- Use a less chemically aggressive additive
system;
-- Remove abrasives from the lubricant by
using fine filtration or frequent oil changes.
4.4 Corrosion
Corrosion is the chemical or electrochemical reac-
tion between the surface of the gear and its
environment. Corrosion usually leaves a stained,
rusty appearance and can be accompanied by rough
AMERICAN GEAR MANUFACTURERS ASSOCIATION
irregular pits or depressions. Identification of metal
corrosion products is an indication of corrosion. For
example, the identification of --Fe2O3 H2O by X--ray
diffraction on pitted steel is evidence of rusting.
Corrosion commonly attacks the tooth surface and it
may proceed intergranularly by preferentially attack-
ing the grain boundaries of the gear surfaces.
Etch pits from corrosion on the active flanks of gear
teeth cause stress concentrations which may initiate
macropitting fatigue cracks. Etch pits on the root
fillets of gear teeth may promote bending fatigue
cracks.
Water reduces fatigue life by causing hydrogen
embrittlement which accelerates fatigue crack
growth.
The particles of rust are hard and they can cause
abrasive wear of the gear teeth.
Corrosion is often caused by contaminants in the
lubricant such as acid or water. Overly reactive,
anti--scuff additives can also cause corrosion espe-
cially at high temperatures. Corrosive wear caused
by contamination or formation of acids in the
lubricant can be minimized by monitoring the lubri-
cant acidity, viscosity and water content and by
changing the lubricant when required.
Methods for reducing corrosion
A gear lubricant should be changed if the neutraliza-
tion number increases 0.5 units over the baseline
value of the unused product, the water content is
greater than 0.1%, or the viscosity increases or
decreases to the next ISO viscosity grade.
Gear units not properly protected during storage can
become corroded. If the gear unit must be stored,
special precautions are usually required to prevent
rusting of the components. Condensation occurs
when humid air is cooled below its dew point and the
air--water mixture releases water, which collects in
the form of droplets on exposed surfaces. It may
occur where there are frequent, wide temperature
changes. For long term storage, it is best to
completely fill the gear unit with oil and plug the
breather vent. This minimizes the air space above
the oil level and minimizes the amount of condensa-
tion. Where this is not practical, all exposed metal
parts, both inside and outside, should be sprayed
with a heavy duty rust preventative. If stored
outdoors, the gear unit should be raised off the
ground and completely enclosed by a protective
covering such as a tarpaulin. The gears should be
 AGMA 2004 ---- All rights reserved
AGMA 912--A04
rotated frequently to distribute oil to the gears and
bearings.
4.5 Fretting corrosion
Fretting occurs between contacting surfaces that are
pressed together and subjected to cyclic, relative
motion of extremely small amplitude. It occurs most
often in joints that are bolted, keyed or press--fitted,
in bearings, splines or couplings. It can also occur on
gear teeth under specific conditions where the gears
are not rotating and are subjected to vibration such
as during shipping.
Under fretting conditions, the lubricant is squeezed
from between the surfaces and the motion of the
surfaces is too small to replenish the lubricant. The
natural, oxide films that normally protect the surfaces
are disrupted, permitting metal--to--metal contact
and causing adhesion of the surface asperities. The
relative motion breaks the welded asperities and
generates extremely small wear particles which
oxidize to form iron--oxide powder (Fe2O3), which
has the fineness and reddish--brown color of cocoa.
The wear debris is hard and abrasive, and is the
same composition as jewelers rouge. Fretting
corrosion tends to be self--aggravating because the
wear debris builds a dam which prevents fresh
lubricant from reaching the contact area.
Fretting corrosion is sometimes responsible for
initiating fatigue cracks, which, if they are in high
stress areas, may propagate to failure.
Methods for reducing fretting corrosion
-- Ship the gear unit on an air--ride truck;
-- Support the gear unit on vibration isolators;
-- Ship the gear unit filled with oil.
4.6 Cavitation
Cavitation has been known to occur in the lubricant
film between mating gear teeth. Cavitation is
characterized by the formation of vapor filled
bubbles at the interface between a solid and a liquid,
generally in an area of low pressure. When the
bubbles travel into a region of high pressure they
collapse as they change state from gas to liquid. The
implosion of the bubbles transmits localized forces to
the surface which cause fracture of the surface
asperities. To the unaided eye, a surface damaged
by cavitation may appear to be rough and clean as if
it were sandblasted. The microscopic craters
caused by cavitation are deep, rough, clean and
have a honeycomb appearance.
9
AGMA 912--A04
4.7 Electrical discharge damage
Gear teeth may be damaged if electric current is
allowed to pass through the gear mesh. Electrical
discharge damage is caused by electric arc dis-
charge across the oil film between the active flanks
of the mating gear teeth. The electric current may
originate from many sources, including:
-- electric motors;
-- electric clutches or instrumentation;
-- accumulation of static charge and subse-
quent discharge;
-- during electric welding on or near the gear
unit if the path to ground is not properly made
around the gears rather than through them.
An electric arc may produce temperatures high
enough to locally melt the gear tooth surface. To the
unaided eye, a surface damaged by electrical
discharge appears as an arc burn similar to a spot
weld. On a microscopic level, small hemispherical
craters can be observed. The edges of the crater are
smooth and they may be surrounded by burned or
fused metal in the form of rounded particles that were
once molten. An etched metallurgical section taken
transversely through the craters may reveal austeni-
tized and rehardened areas in white, bordered by
tempered areas in black.
The damage to the gear teeth is proportional to the
number and size of the points of arcing. Depending
on its extent, electrical discharge damage can be
destructive to the gear teeth. If arc burns are found
on the gears, all associated bearings should be
examined for similar damage.
Methods for reducing electrical discharge
damage
Electric discharge damage can be prevented by
providing adequate electrical insulation or grounding
and by ensuring that proper welding procedures are
enforced.
5 Scuffing
Scuffing is damage caused by localized welding
between sliding surfaces. It is accompanied by
transfer of metal from one surface to another due to
welding and tearing. It may occur in any sliding and
rolling contact where the oil film is not thick enough to
prevent metal--to--metal contact. It is characterized
10
AMERICAN GEAR MANUFACTURERS ASSOCIATION
by a microscopically rough, matte, torn surface.
Surface analysis that shows transfer of metal from
one surface to the other is evidence of scuffing.
Scuffing can occur in gear teeth when they operate in
the boundary lubrication regime. If the lubricant film
is insufficient to prevent significant metal--to--metal
contact, the oxide layers that normally protect the
gear tooth surfaces may be broken through, and the
bare metal surfaces may weld together. The sliding
that occurs between gear teeth results in tearing of
the welded junctions, metal transfer and damage.
In contrast to macropitting and bending fatigue,
which only occur after a period of running time,
scuffing may occur immediately upon start--up. In
fact, gears are most vulnerable to scuffing when they
are new and their tooth surfaces have not yet been
smoothed by running--in. It is recommended that
new gears be run--in under one--half load to reduce
the surface roughness of the teeth before the full
load is applied. The gear teeth can be coated with
iron manganese phosphate or plated with copper or
silver to reduce the risk of scuffing during the critical
running--in period. Also, the use of an anti--scuff
additive, for example, SP hypoid oil, can help
prevent scuffing and promote polishing during run--
in, but oil should be changed to the operational oil
after run--in.
The basic mechanism of scuffing is not clearly
understood, but there is general agreement that it is
caused by frictional heating generated by the
combination of high sliding velocity and intense
surface pressure. Critical temperature theory [2] is
often used for predicting scuffing. It states that
scuffing will occur in gear teeth that are sliding under
boundary--lubricated conditions, when the maxi-
mum contact temperature of the gear teeth reaches
a critical magnitude. For mineral oils without anti--
scuff additives, each combination of oil and gear
tooth material has a critical scuffing temperature
which is constant regardless of the operating condi-
tions [3]. The critical scuffing temperature may be
constant for synthetic lubricants and lubricants with
anti--scuff additives, and should be determined from
tests which closely simulate the operating conditions
of the gears.
Most anti--scuff additives are sulfur--phosphorus
compounds which form boundary lubricating films by
chemically reacting with the metal surfaces of the
gear teeth at local points of high temperature.
Anti--scuff films help prevent scuffing by forming
solid films on the gear tooth surfaces and inhibiting
 AGMA 2004 ---- All rights reserved
AMERICAN GEAR MANUFACTURERS ASSOCIATION
true metal--to--metal contact. The films of iron sulfide
and iron phosphate have high melting points,
allowing them to remain as solids on the gear tooth
surfaces even at high contact temperatures. The
rate of reaction of the anti--scuff additives is greatest
where the gear tooth contact temperatures are
highest. Because of the sliding action of the gear
teeth, the surface films are repeatedly scrapped off
and reformed. In effect, scuffing is prevented by
substituting mild corrosion in its place. Anti--scuff
additives may promote micropitting. Some anti--
scuff additives may be too chemically active (see
4.3). This may necessitate a change to less
aggressive additives, such as potassium borate,
because it deposits a boundary film without reacting
to the metal.
For mineral oils without anti--scuff additives, the
critical scuffing temperature increases with increas-
ing viscosity, and ranges from 150° to 300°C.
According to [3], the critical temperature is the total
contact temperature, Tc, which consists of the sum of
the gear bulk temperature, Tb, and the flash temper-
ature, Tf:
T c = T b + Tf (1)
The bulk temperature is the equilibrium temperature
of the surface of the gear teeth before they enter the
meshing zone. The flash temperature is the local
and instantaneous temperature rise that occurs on
the gear teeth due to the frictional heating as they
pass through the meshing zone.
Anything that reduces the total contact temperature
will lessen the risk of scuffing. The lubricant
performs the important function of removing heat
from the gear teeth. A heat exchanger can be used
with a circulating oil system to cool the lubricant
before it is sprayed at the gears. Higher viscosity
lubricants or smoother tooth surfaces help by
increasing the specific film thickness, which in turn
reduces the frictional heat, and therefore the flash
temperature.
Scuffing resistance may be increased by optimizing
the gear geometry such that the gear teeth are as
small as possible, consistent with bending strength
requirements, to reduce the temperature rise
caused by sliding. The amount of sliding is
proportional to the distance from the pitch point and
is zero when the gear teeth contact at the pitch point,
 AGMA 2004 ---- All rights reserved
AGMA 912--A04
and largest at the ends of the path of action. Profile
shift can be used to balance and minimize the
temperature rise that occurs in the addendum and
dedendum of the gear teeth. The temperature rise
may also be reduced by modifying the tooth profiles
with slight tip and/or root relief to ease the load at the
start and end of the engagement path where the
sliding velocities are the greatest. Also, the gear
teeth should be accurate and held rigidly in good
alignment to minimize the tooth loading and
temperature rise.
The gear materials should be chosen with their
scuffing resistance in mind. Nitrided steels such as
Nitralloy 135M are generally found to have the
highest resistance to scuffing, while some stainless
steels may scuff even under near--zero loads. The
thin oxide layer on these stainless steels is hard and
brittle and it breaks up easily under sliding loads,
exposing the bare metal, thus promoting scuffing.
Anodized aluminum also has a low scuffing resist-
ance. Hardness alone does not seem to be a reliable
indication of scuffing resistance.
Methods for reducing the risk of scuffing
-- Use smooth tooth surfaces produced by care-
ful grinding or honing;
-- Run in new gearsets by operating for the first
10 hours at one--half load;
-- Protect the gear teeth during the critical run--
in period by use of a special lubricant, coating
(such as iron manganese phosphate), or by
plating (such as copper or silver);
-- Use lubricants of adequate viscosity for the
operating conditions;
-- Use lubricants that contain anti--scuff addi-
tives such as sulfur, phosphorous, or dispersions
of potassium borate, PTFE, and others;
-- Cool the gear teeth by supplying an adequate
amount of cool lubricant. For circulating--oil
systems, use a heat exchanger to cool the
lubricant;
-- Optimize the gear tooth geometry by using
small teeth, profile shift and profile modification;
-- Use accurate gear teeth, with uniform load
distribution during operating;
-- Use nitrided steels for maximum scuffing
resistance.
11
AGMA 912--A04
6 Plastic deformation
Plastic deformation is permanent deformation that
occurs when the stress exceeds the yield strength of
the material. It may occur at the surface or subsur-
face of the active flanks of the gear teeth due to high
contact stress, or at the root fillets due to high
bending stress.
6.1 Indentation
The active flanks of gear teeth can be damaged by
indentations caused by foreign material which be-
comes trapped between the teeth. Depending on
the number and size of the indentations, the damage
may or may not initiate failure. If plastic deformation
associated with the indentations causes raised
areas on the tooth surface, it creates stress con-
centrations which may lead to subsequent Hertzian
fatigue. For gear teeth subjected to contact stresses
greater than 1.8 times the tensile yield strength of the
material, local, subsurface yielding may occur. The
subsurface plastic deformation causes grooves
(brinelling) on the surfaces of the active flanks of the
teeth corresponding to the lines of contact between
the mating gear teeth.
6.2 Cold flow
Cold flow is plastic deformation that occurs at a
temperature lower than the recrystallization
temperature.
6.3 Hot flow
Hot flow is plastic deformation that occurs at a
temperature higher than the recrystallization
temperature.
6.4 Rolling
Plastic deformation may occur on the active flanks of
gear teeth caused by high contact stresses and the
rolling and sliding action of the gear mesh. Often the
surface material is displaced from the pitch line of the
driving gear teeth toward both the roots and tips
forming burrs. The surface material of the driven
gear is displaced towards the pitchline forming a
ridge. A corresponding groove is formed along the
pitchline of the driving gear.
6.5 Rippling
Rippling is periodic, wave--like undulations of the
surfaces of the active flanks of gear teeth. The peaks
or ridges of the undulations run perpendicular to the
direction of sliding. The ridges are wavy along the
12
AMERICAN GEAR MANUFACTURERS ASSOCIATION
length of the tooth, creating a fish scale appearance.
Rippling is caused by plastic deformation at the
surface or subsurface. It usually occurs under high
contact stress and boundary--lubricated conditions.
6.6 Ridging
Ridging is the formation of pronounced ridges and
grooves on the active flanks of gear teeth. It
frequently occurs on the teeth of slow--speed,
heavily loaded worm or hypoid gearsets.
6.7 Root fillet yielding
Gear teeth may be permanently bent if the bending
stress in the root fillets exceeds the tensile yield
strength of the material. The bending deflection at
initial yielding is small and there is a margin of safety
before gross yielding causes significant gear tooth
spacing error. If the teeth have sufficient ductility, ini-
tial yielding at the root fillets redistributes the stress
and lowers the stress concentration. Hence, root fil-
let yielding may only result in rougher running and a
higher noise level. However, if the yielding causes
significant spacing errors between loaded teeth that
are permanently bent and unloaded teeth that are
not, subsequent rotation of the gears usually results
in destructive interference between the pinion and
gear teeth.
6.8 Tip--to--root interference
Plastic deformation and abrasive wear may occur at
the tips of the teeth and at the roots of the teeth of the
mating gear due to tip--to--root interference. The in-
terference can be caused by geometric errors in the
profiles such as excessive form diameter, spacing
errors, deflection under load, or a center distance
that is too short.
7 Contact fatigue
7.1 Macropitting
Macropitting is a fatigue phenomenon which occurs
when a shear related fatigue crack initiates either at
the surface of the active flank of the gear tooth or at a
small depth below the surface. The crack usually
propagates for a short distance in a direction roughly
parallel to the tooth surface before turning or
branching to the surface. When cracks grow to the
extent that they separate a piece of the surface
material, a pit is formed. If several pits grow together
to form a larger pit, it is often referred to as a “spall”.
There is no endurance limit for contact fatigue, and
 AGMA 2004 ---- All rights reserved
AMERICAN GEAR MANUFACTURERS ASSOCIATION
macropitting occurs even at low stresses if the gears
are operated long enough. Macropitting often
initiates at non--metallic inclusions in the gear
material. Because there is no endurance limit, gear
teeth must be designed for a suitable, finite lifetime.
To prolong the macropitting life of a gearset, the
designer must keep the contact stress low, material
strength high, material relatively free of inclusions,
and the lubricant specific film thickness high. There
are several geometric variables such as diameter,
face width, number of teeth, and pressure angle that
may be optimized to lower the contact stress.
Material alloys and heat treatment are selected to
obtain hard tooth surfaces with high strength, such
as carburizing or nitriding. Maximum macropitting
resistance is obtained with carburized gear teeth
because they have hard surfaces, and carburizing
induces beneficial compressive residual stresses
which effectively lower the shear stresses. High
lubricant specific film thickness is obtained by using
smooth tooth surfaces and an adequate supply of
cool, clean and dry lubricant that has high viscosity
and a high pressure--viscosity coefficient.
Methods for reducing the risk of macropitting
-- Reduce contact stresses by reducing loads or
optimizing gear geometry;
-- Use clean steel, properly heat treated to high
surface hardness;
-- Use smooth tooth surfaces;
-- Use an adequate amount of cool, clean and
dry lubricant of adequate viscosity;
-- Adequate surface hardness and case depth
after final processing.
7.2 Micropitting
On relatively soft gear tooth surfaces, such as those
of through hardened gears, Hertzian fatigue forms
large pits with dimensions on the order of
millimeters. With surface hardened gears, such as
carburized, nitrided, induction hardened or flame
hardened, pits may occur on a much smaller scale,
typically only 10 micrometers deep. To the naked
eye, the areas where micropitting has occurred
appear frosted, and “frosting” is a popular term for
micropitting. Researchers [4] have referred to the
failure mode as “grey staining” because the
light--scattering properties of micropitting gives the
gear teeth a grey appearance. Under the
microscope it appears that micropitting propagates
 AGMA 2004 ---- All rights reserved
AGMA 912--A04
by the same fatigue process as macropitting, except
the pits are extremely small.
Many times micropitting is not destructive to the gear
tooth surface. It sometimes occurs only in patches,
and may arrest after the tribological conditions have
improved by running--in. The micropits may actually
be removed by polishing wear during running--in, in
which case the micropitting is said to “heal”. Howev-
er, there have been examples where micropitting
has escalated into full scale macropitting, leading to
the destruction of the gear teeth.
The lubricant’s specific film thickness is an important
parameter that influences micropitting. Damage
seems to occur most readily on gear teeth with rough
surfaces, especially when they are lubricated with a
low viscosity lubricant. Gears finished to a mirror--
like finish have eliminated micropitting. Slow--speed
gears are prone to micropitting because their film
thickness is low. Hence, to prevent micropitting, the
specific film thickness should be maximized by using
smooth gear tooth surfaces, high viscosity lubri-
cants, and if possible high speeds. ANSI/AGMA
9005--E02 gives recommendations for viscosity as a
function of pitchline velocity.
Methods for reducing the risk of micropitting
-- Use smooth tooth surfaces or coatings;
-- Use an adequate amount of cool, clean and
dry lubricant of the highest viscosity possible;
-- Use high speeds if possible;
-- Use carburized steel with proper carbon
content in the surface layers;
-- Reduce load, modify profiles.
7.3 Subcase fatigue
Subcase fatigue may occur in case (surface) hard-
ened gears such as those that are carburized,
nitrided or induction hardened. The origin of the
fatigue crack is below the surface of the gear tooth,
frequently in the transition zone between the case
and core where the cyclic shear stresses exceed the
shear fatigue strength. The crack typically runs
parallel to the surface of the gear tooth before
branching to the surface. The branched cracks may
appear at the surface as fine longitudinal cracks on
only a few teeth. If the surface cracks join together,
long shards of the tooth surface may break away.
The resulting crater is longitudinal with a relatively
flat bottom and sharp, perpendicular edges. Fatigue
beach marks may be evident on the crater bottom
formed by propagation of the main crack.
13
AGMA 912--A04
Subcase fatigue is influenced by contact stresses,
residual stresses and material fatigue strength. The
subsurface distribution of residual stresses and
fatigue strength depends on the surface hardness,
case depth and core hardness. There are optimum
values of case depth and core hardness which give
the proper balance of residual stresses and fatigue
strength to maximize resistance to subcase fatigue.
Inclusions may initiate fatigue cracks if they occur
near the case--core interface in areas of tensile
residual stress.
Overheating gear teeth during operation or
manufacturing, such as grind temper, may lower
case hardness, alter residual stresses, and reduce
resistance to subcase fatigue. See 8.3 for discussion
of grind temper.
Methods for reducing the risk of subcase fatigue
-- Reduce contact stresses by reducing loads or
optimizing gear geometry;
-- Use steel with adequate hardenability to
obtain optimum case and core properties;
-- Achieve optimum values of surface hard-
ness, case depth and core hardness to maximize
resistance to subcase fatigue;
-- Use analytical methods to ensure that sub-
surface stresses do not exceed subsurface
fatigue strengths;
-- Avoid overheating gear teeth during
operation or manufacturing.
8 Cracking
8.1 Hardening cracks
Cracking in heat treatment occurs because of
excessive localized stresses. These may be caused
by nonuniform heating or cooling, or by volume
changes due to phase transformation. Stress risers
will make the part more susceptible to cracking.
Hardening cracks are generally intergranular with
the crack running from the surface toward the center
of mass in a relatively straight line. Crack formation
may be related to some of the same factors which
cause intergranular fracture in overheated steels. If
cracking occurs prior to tempering, the fracture
surfaces will be discolored by oxidation when the
gear is exposed to the furnace atmosphere during
tempering.
14
AMERICAN GEAR MANUFACTURERS ASSOCIATION
Cracks resulting from heat treatment sometimes
appear immediately, but at other times may not
appear until the gears have operated for a period of
time.
8.1.1 Thermal stresses
Thermal stresses are caused by temperature differ-
ences between the interior and exterior of the gear,
and increase with the rate of temperature change.
Cracking can occur either during heating or cooling.
The cooling rate is influenced by the geometry of the
gear, the agitation of the quench, quench medium,
and temperature of the quenchant. The temperature
gradient is higher and the risk of cracking greater
with thicker sections, asymmetric gear blanks and
variable thickness rims and webs.
8.1.2 Stress concentration
Features such as sharp corners, the number,
location and size of holes, deep keyways, splines,
and abrupt changes in section thickness within a part
cause stress concentrations, which increase the risk
of cracking.
Surface and subsurface defects such as nonmetallic
inclusions, forging defects such as hydrogen flakes,
internal ruptures, seams, laps, and tears at the flash
line increase the risk of cracking.
8.1.3 Quench severity
Quenching conditions should be designed consider-
ing size and geometry of the gear, required
metallurgical properties, and hardenability of the
steel.
Quench severity and the risk of cracking are greater
with vigorously agitated, caustic, or brine quen-
chants and much less with quiescent, slow--oil
quenchants.
Hardening cracks may not occur while the gear is in
the quenching medium, but later if the gear is
allowed to stand after quenching without tempering.
8.1.4 Phase transformation
Transformation of austenite into martensite is al-
ways accompanied by expansion, and may result in
cracking. See [5].
8.1.5 Methods for reducing the risk of hardening
cracks
-- Design the gear blanks to be as symmetric as
possible and keep section thickness uniform;
-- Minimize abrupt change in cross section. Use
chamfers or radii on all edges, especially at the
 AGMA 2004 ---- All rights reserved
AMERICAN GEAR MANUFACTURERS ASSOCIATION
ends of the teeth and at the edges of the gear
tooth toplands;
-- Select steel type carefully;
-- Design the quenching method, including the
agitation, type of quenchant and temperature of
the quenchant, for the specific gear and
hardenability of the steel;
-- Temper the gear immediately after
quenching.
8.2 Steel grades
In general, the carbon content of steel should not
exceed the required level; otherwise, the risk of
cracking will increase. The suggested average
maximum carbon content for water, brine, and
caustic quenching are given below:
Induction hardening:
Complex shapes 0.40%
Simple shapes 0.60%
Furnace hardening:
Complex shapes 0.35%
Simple shapes 0.40%
Very simple shapes (such as bars) 0.50%
8.2.1 Part defects
Surface defect or weakness in the material may also
promote cracking, for example, deep surface seams
or nonmetallic stringers in both hot--rolled and
cold--finished bars. Other problems are inclusions
and stamp marks. Forging defects in small forgings,
such as seams, laps, flash line or shearing cracks as
well as in heavy forgings such as hydrogen flakes
and internal ruptures, aggravate cracking. Similarly,
some casting defects, for example, in water--cooled
castings, promote cracking.
8.2.2 Heat treating practice
Anneal alloy steels prior to hardening (or any other
high--temperature treatment, such as forging or
welding) to produce grain--refined microstructure
and relieve stresses. Improper heat treating practic-
es, such as nonuniform heating or cooling, contrib-
ute to cracking. Water hardening or air hardening
can cause cracking if the steel is not properly
processed. For example, the lack of tempering or
use of oil quenching with an air hardening steel can
lead to cracking. However, common practice in the
treatment of air hardening steels is to initially quench
in oil until “black” (about 538°C), followed by air
 AGMA 2004 ---- All rights reserved
AGMA 912--A04
cooling to 66°C prior to tempering. This practice
minimizes the formation of scale.
8.2.3 Tempering practice
The longer the time the steel is kept at a temperature
between room temperature and 100°C after the
complete transformation of martensite in the core,
the more likely the occurrence of quench cracking.
This arises from the volumetric expansion caused by
isothermal transformation of retained austenite into
martensite.
There are two tempering practices which lead to
cracking problems: tempering soon after quenching,
that is, before the steel parts have transformed to
martensite in hardening, and superficial surface
(skin) tempering, usually observed in heavy sections
(50 mm and thicker in plates and 75 mm and greater
in diameter in round bars).
It is the normal practice to temper immediately after
the quenching operation. In this case, some restraint
must be exercised, especially for large sections
(greater than 75 mm) in deep--hardening alloy
steels. The reason is that the core has not yet
completed transformation to martensite with expan-
sion while the surface projections, such as flanges,
begin to temper with shrinkage. This simultaneous
volume change produces radial cracks. This prob-
lem can become severe if rapid heating practice
(such as induction, flame, lead or molten salt bath) is
used for tempering.
8.3 Grinding cracks
Cracks may develop on the tooth surfaces of gears
that are finished by grinding. The cracks are usually
shallow and appear either as a series of parallel
cracks or in a crazed, wire--mesh pattern. Like
hardening cracks, they may not appear until the
gears have operated for a period of time. Cracks
may be caused by the grinding technique if the
grinding cut is too deep, grinding feed is too high,
grinding speed is too high, grinding wheel grit or
hardness is incorrect, or flow of coolant is insuffi-
cient. Grinding cracks may result from transforma-
tion of retained austenite to martensite in response
to the heat or stresses imposed by grinding. See [6].
Steels with hardenability provided by carbide--form-
ing elements such as chromium are prone to
grinding cracks. This is especially true for carburized
gears with a case that has high carbon content,
particularly if there are carbide networks.
15
AGMA 912--A04
Areas of the tooth surface where overheating has
occurred can be detected by etching the surface with
nital. See ANSI/AGMA 2007--C00. Barkhausen
(eddy--current) inspection may be used if properly
qualified for the specific part. Magnetic particle or
dye penetrant inspection can be used to detect
grinding cracks.
Methods for reducing the risk of grinding cracks
-- Control grinding technique to avoid local over
heating;
-- For carburized gears, control microstructure
to limit carbides;
-- Use nital etch to inspect ground surfaces for
tempering;
-- Use magnetic particle or dye penetrant in-
spection of ground surfaces to detect grinding
cracks.
8.4 Rim and web cracks
If the gear rim is thin, less than twice the gear tooth
whole depth, it is subjected to significant alternating
rim--bending stresses, which are additive to the
gear--tooth bending stress and may result in fatigue
cracks in the rim.
Web cracks can be caused by cyclic stresses due to
vibration when an excitation frequency is near a
natural frequency of the gear blank.
Stress concentrations due to defects such as
inclusions, notches in the root fillets, and details such
as keyways, splines, holes and sharp web--to--rim
fillets can cause cracks.
Magnetic particle or dye penetrant inspection should
be used to ensure that the gear tooth fillets, gear rim
and gear web are free of flaws.
Methods to reduce the risk of rim or web cracks
-- Use adequate rim thickness;
-- Design the gear blank such that its natural fre-
quencies do not coincide with the excitation fre-
quencies;
-- Pay attention to details that cause stress con-
centrations such as keyways, splines, holes and
web--to--rim fillets;
-- Use magnetic particle or dye penetrant in-
spection to ensure that the gear tooth fillets, gear
rim and gear web are free of flaws;
-- Control manufacturing to avoid notches in the
root fillets.
16
AMERICAN GEAR MANUFACTURERS ASSOCIATION
8.5 Case--core separation
Case--core separation occurs in surface hardened
gear teeth when internal cracks occur near the case
core boundary. The internal cracks may pop to the
surface of the teeth causing corners, edges or entire
tips of the teeth to separate. The damage may occur
immediately after heat treatment, during subsequent
handling, or after a short time in service.
Case--core separation is believed to be caused by
high residual tensile stresses at the case--core
interface when a case is very deep.
Because cracks follow the case--core interface, tips
of teeth have concave fracture surfaces, and re-
maining portions of teeth have convex fracture
surfaces. Chevron (beach) marks may be apparent
on fracture surfaces if the fracture was brittle. These
marks are helpful because they point to the failure
origin. Beach marks may be found on fracture
surfaces if cracks grew by fatigue. Inclusions pro-
mote case--core separation especially when they
occur near the interface.
When case--core separation is suspected as the
cause of failure, intact teeth should be sectioned to
determine if there are subsurface cracks near the
tips of the teeth.
On carburized gears, case depth at the tip can be
controlled by avoiding narrow toplands or masking
the toplands with copper plate to restrict carbon
penetration during carburizing.
Methods for reducing the risk of case--core
separation
-- Control case depth especially at the tips of the
gear teeth. On carburized gears, avoid narrow
toplands or mask toplands of the teeth to restrict
carbon penetration;
-- Temper gears immediately after quenching;
-- Use generous chamfers or radii on edges of
the gear teeth to avoid stress concentrations;
-- Control the alloy content, cleanliness of the
steel, and the core hardness. They all influence
the probability of case--core separation.
9 Fracture
When a gear tooth is overloaded because the local
load is too high, it may fail by fracturing. If it fractures,
the failure may be a ductile fracture preceded by
 AGMA 2004 ---- All rights reserved
AMERICAN GEAR MANUFACTURERS ASSOCIATION AGMA 912--A04

appreciable plastic deformation, a brittle fracture fracture. The critical stress intensity is a function of
with little prior plastic deformation, or a mixed--mode the material toughness.
fracture exhibiting both ductile and brittle
The toughness of a gear material depends on many
characteristics.
factors especially temperature, loading rate and
If fatigue cracks grow to a point where the remaining constraint (state of plane stress or plane strain) at the
tooth section can no longer support the load, a location of flaws. Many steels have a transition
fracture will occur. In this sense the remaining temperature where the fracture mode changes from
material is overloaded, however, the fracture is a ductile--to--brittle as temperature decreases. The
secondary failure mode that is caused by the primary transition temperature is influenced by the loading
mode of fatigue cracking. rate and constraint. The ductile--to--brittle transition
can be detected with the Charpy V--notch impact
Gear tooth fractures without prior fatigue cracking test. Some high strength, alloyed, quenched and
are infrequent, but may result from shock loads. The
tempered steels do not exhibit a transition tempera-
shock loads may be generated by the driving or
ture behavior. For low temperature service, the
driven equipment. They may also occur when
transition temperature is of primary importance, and
foreign objects enter the gear mesh, or when the
gear materials should be chosen which have
gear teeth are suddenly misaligned and jam together
after a bearing or shaft fails. transition temperatures below the service
temperature.
Fractures are classified as brittle or ductile depend-
The compliance of shafts and couplings in a drive
ing on their macroscopic and microscopic
system helps to cushion shock loads and reduce the
appearance (see table 1).
loading rate during impact. Gear drives with close--
coupled shafts and rigid couplings have less
compliance. If drive systems with low compliance
Table 1 -- Fracture appearance classifications
must be used in applications where overloads are
Characteristic Brittle Ductile expected, the gears should be large enough to
of fracture fracture fracture absorb the overloads with reasonable stress levels.
surface See [7].
light reflection bright gray (dark)
shiny dull The toughness of a material depends on its elemen-
tal composition, heat treatment and mechanical
texture crystalline silky
grainy matte processing. Many alloying elements that increase
rough smooth the hardenability of steel also decrease its tough-
coarse fine ness. Exceptions are nickel and molybdenum that
granular fibrous (stringy) increase hardenability while improving toughness.
orientation flat slant, or flat Tests on the impact fracture resistance of carburized
square angular, or steel have found the following, see [8]:
square
pattern radial ridges shear lips -- High--hardenability steels have greater im-
chevrons pact fracture resistance than low--hardenability
plastic negligible appreciable steels;
deformation -- High nickel content does not guarantee good
(necking or impact fracture resistance, but nickel and
distortion)
molybdenum in the right combination result in
microscopic cleavage dimples (shear) high impact fracture resistance;
features (facets)
-- High chromium and high manganese
contents tend to give low impact fracture resist-
9.1 Brittle fracture ance.
Toughness can be optimized by keeping the carbon,
Brittle fracture occurs when tensile stress exceeds a
phosphorus and sulfur content as low as possible.
critical stress intensity. Part shape, machining
marks, and material flaws may lead to stress Fracture initiates at flaws which cause stress con-
concentration, which usually plays a role in brittle centrations. The flaw may be a notch, crack, surface

 AGMA 2004 ---- All rights reserved 17

AGMA 912--A04
tear, surface or subsurface inclusion, or porosity.
The flaw size may be small initially, but it may initiate
a fatigue crack that can grow until a critical size is
reached, at which point the crack may extend in a
brittle fracture. The critical flaw size is not constant,
but depends on the geometry of the part, shape and
orientation of the flaw, applied stress, and the
fracture toughness of the material at the service
temperature and loading rate.
The root fillets of gear teeth are especially vulnerable
to fracture because this is the location where tooth
bending stresses are highest. Clean materials
increase fracture resistance.
The gear tooth geometry should be selected to
reduce the tensile bending stress in the root fillets.
The gear teeth may be cut with full--fillet tools to
obtain large root fillets with minimum stress con-
centrations. If the gears are to be finished by shaving
or grinding, protuberance tools should be used to
reduce the risk of notching the root fillets. Case
hardening by carburizing or nitriding can be benefi-
cial because these hardening processes may induce
compressive residual stresses which reduce the net
tensile bending stresses. Also, controlled shot
peening can be used to increase compressive
residual stresses.
Methods for reducing the risk of brittle fracture
-- Use materials with high cleanliness;
-- Use materials and heat treatments that give
high toughness, such as steel with sufficient
hardenability to obtain a microstructure of primari-
ly tempered martensite. Avoid embrittlement by
using steel in which the desired hardness will be
achieved without tempering in the range of 250 to
400 degrees C;
-- Do not use steels at service temperatures
below their transition temperature;
-- Reduce loading rates by using compliant
shafts and couplings;
-- Protect gears from impact loads by using load
limiting couplings;
-- Use steels with high nickel content. For car-
burized gears, nickel and molybdenum in the right
combination gives maximum toughness. Do not
use steels with high chromium and manganese
content. Keep the carbon, phosphorus and sulfur
content as low as possible;
-- Use fine grained steel;
18
AMERICAN GEAR MANUFACTURERS ASSOCIATION
-- Eliminate flaws, especially in the root fillets of
gear teeth. Use magnetic particle on dye pene-
trant inspection to detect flaws;
-- Reduce tensile bending stresses by
optimizing gear tooth geometry;
-- Use case hardening, or shot peening, or both
to increase compressive residual stresses.
9.2 Ductile fracture
Gear tooth failures that occur solely by ductile
fracture are relatively infrequent because most
fractures occur at a pre--existing flaw which tends to
promote brittle behavior. Factors that promote
ductile rather than brittle fracture are:
-- high material toughness;
-- high temperature;
-- slow loading rate;
-- no significant material flaws;
-- low tensile stress;
-- high shear stress.
Under these conditions gear teeth yield when the
bending stresses exceed the yield strength of the
material, and subsequently shear off with significant
plastic deformation before ductile fracture.
10 Bending fatigue
Although bending fatigue cracks may occur else-
where on gear teeth, they usually initiate in the root
fillets on the tensile side of the teeth. The geometry
of the root fillets may cause significant stress con-
centrations, which, combined with a high bending
moment, results in high bending stress.
Fatigue is a progressive failure consisting of three
distinct stages:
Stage 1 Crack initiation
Stage 2 Crack propagation
Stage 3 Fracture
Most of the fatigue life is occupied by stages 1 and 2
until the cracks grow to critical size where sudden
fracture occurs in stage 3. The fracture may be duc-
tile, brittle, or mixed--mode depending upon the
toughness of the material and the magnitude of the
applied stress (see discussion in clause 9).
During stage 1 the peak bending stress is less than
the yield strength of the material and no gross yield-
 AGMA 2004 ---- All rights reserved
AMERICAN GEAR MANUFACTURERS ASSOCIATION
ing of the gear teeth occurs. However, local plastic
deformation may occur in regions of stress con-
centrations or areas of structural discontinuities,
such as surface notches, grain boundaries or inclu-
sions. The cyclic, plastic deformation occurs on slip
planes that coincide with the direction of maximum
shear stress. The cyclic slip continues within these
grains, usually near the surface where stress is high-
est, until cracks are initiated. The cracks grow in the
planes of maximum shear stress and coalesce
across several grains until they form a major crack
front.
The stage 2 propagation phase begins when the
crack turns and grows across grain boundaries
(transgranular) in a direction approximately perpen-
dicular to the maximum tensile stress. During the
propagation phase, the plastic deformation is con-
fined to a small zone at the tip of the crack, and the
surfaces of the fatigue crack usually appear smooth
without signs of gross plastic deformation. Under the
scanning electron microscope, ripples may be seen
on a fatigue cracked surface, called fatigue stri-
ations. They are thought to be associated with alter-
nating blunting and sharpening of the crack tip, and
correspond to the advance of the crack during each
stress cycle. The orientation of the striations is at 90
degrees to the crack advance. If the crack propa-
gates intermittently, it may leave a pattern of macro-
scopically visible “beach marks”. These marks
correspond to various positions of the crack front
where the crack arrested, because the magnitude of
the stress changed.
Beach marks are helpful to the failure analyst be-
cause they aid in locating the origins of fatigue
cracks. The origin is usually on the concave side of
the curved beach marks and is often surrounded by
several, concentric beach marks. Beach marks may
not be present, especially if the fatigue crack grows
without interruption under cyclic loads that do not
vary in magnitude. The presence of beach marks is
a strong indication that the crack was due to fatigue,
but not absolute proof, because other failure modes
sometimes leave beach marks, and stress corrosion
under changing environment. If there are multiple
crack origins, each producing separate crack propa-
gation zones, ratchet marks may be formed. They
are caused when adjacent cracks, propagating on
different crystallographic planes, join together form-
ing a small step. Ratchet marks are often present on
 AGMA 2004 ---- All rights reserved
AGMA 912--A04
the fatigue crack surface of gear teeth because mul-
tiple fatigue crack origins may occur in the root fillet.
10.1 Low--cycle fatigue
Low--cycle fatigue is defined as fatigue where mac-
roscopic plastic strain occurs in every cycle, and the
number of cycles to failure is usually less than
10,000. It is an uncommon failure mode for gear
teeth except for instances where the gear teeth are
greatly overloaded. The surface conditions of a gear
tooth subjected to low--cycle fatigue and the material
cleanliness are less important than under high--cycle
fatigue loading because the cyclic, plastic deforma-
tion tends to relax both stress concentrations and
residual stresses. Cracks may initiate within the
gear teeth as well as on the surface, and a smaller
fraction of the life is spent in initiating rather than
propagating cracks. Low--cycle fatigue life can be
extended by maximizing ductility and toughness
(see 9.1 for discussion regarding factors that
promote toughness). Reference [9] recommends
the following methods to increase the toughness of
carburized gears:
-- Use steels which contain nickel as a major
(more than 1%) alloying element;
-- Quench to produce 15 to 30% retained
austenite in the case microstructure;
-- Temper an as--quenched case hardness of
58 HRC, or higher, down to 55 HRC, or lower
(avoid tempering temperatures of 250 to 400 de-
grees C because of embrittlement of the core).
Caution must be exercised when designing against
low--cycle fatigue because many of the recommen-
dations that improve low--cycle fatigue life decrease
the high--cycle fatigue life. It is better to avoid low--
cycle fatigue by reducing the local stress level.
10.2 High--cycle fatigue
High--cycle fatigue is defined as fatigue where the
cyclic stress is below the yield strength of the materi-
al. Most gear teeth fail by high--cycle fatigue rather
than low--cycle fatigue. Cracks usually initiate at the
surface of the gear tooth root fillets and a large frac-
tion of the life is spent initiating rather than propagat-
ing cracks. High--cycle fatigue life can be extended
by maximizing the ultimate tensile strength of the
material and ensuring that the microstructure of the
surface of the gear teeth is optimum. Reference [9]
recommends the following methods to increase the
high--cycle bending fatigue of carburized gears:
-- Eliminate bainite, pearlite, and network
carbides from the case microstructure;
19
AGMA 912--A04
-- Eliminate microcracks especially near the
surface of the root fillets;
-- Maximize residual compressive stress in the
case by using a steel with a lower possible carbon
content;
-- Eliminate defects on the surfaces of the root
fillets.
There are several geometric variables, such as di-
ameter, face width, number of teeth, pressure
angles, and addendum modification that may be op-
timized to lower the bending stress and increase the
bending fatigue life. The gear tooth geometry should
be designed to reduce the tensile bending stress in
the root fillets. The gear teeth should be cut with full--
fillet tools to obtain large radius root fillets with mini-
mum stress concentrations. If the gears are to be
finished by shaving or grinding, they should be
finished without notching the root fillets. See [10].
Case hardening by carburizing or nitriding can be
beneficial because these hardening processes may
induce compressive residual stresses which reduce
the net tensile bending stresses. Also, controlled
shot peening can be used to increase compressive
residual stresses. For carburized gears there are
optimum values of case hardness, case depth and
core hardness that give the best balance of residual
stresses and fatigue strength to maximize gear tooth
resistance to bending fatigue.
20
AMERICAN GEAR MANUFACTURERS ASSOCIATION
Methods for reducing risk of high--cycle bending
fatigue
-- Use cleaner steels, properly heat treated by
carburizing;
-- Use case hardening, or shot peening, or both
with proper process control to increase compres-
sive residual stresses;
-- For case hardened gears specify values of
case hardness, case depth and core hardness to
maximize resistance to bending fatigue;
-- Use steel with sufficient hardenability to
obtain a microstructure of primarily tempered
martensite in the gear tooth root fillets;
-- Avoid embrittlement by using a steel in which
the desired hardness will be achieved without
tempering in the range of 250 to 400 degrees C;
-- For carburized gears, make sure that the
microstructure of the case is essentially free of
bainite, pearlite, network carbides and especially
microcracks;
-- Use fine--grain steel;
-- Ensure that the surfaces of the root fillets are
relatively free from notches, tool marks, cracks,
nonmetallic inclusions, decarburizing, corrosion,
intergranular oxidation, or other potential stress
risers;
-- Use vacuum (low pressure) carburizing to
prevent decarburizing, intergranular oxidation,
and uneven case depth;
-- Reduce bending stresses by reducing loads
or optimizing gear geometry, especially the shape
of the root fillet.
 AGMA 2004 ---- All rights reserved
AMERICAN GEAR MANUFACTURERS ASSOCIATION
Bibliography
The following documents are either referenced in the text of AGMA 912--A04, Mechanisms of Gear Tooth
Failures, or indicated for additional information.
1. Littman, W.E., The Mechanism of Contact
Fatigue, Interdisciplinary Approach to the Lubrica-
tion of Concentrated Contacts, NASA SP--237,
1970, pp. 309--377.
2. Blok, H., Les Temperatures de Surface dans
Les Conditions de Graissage Sons Pression Ex-
treme, Second World Petroleum Congress, Paris,
June, 1937.
3. Blok, H., The Postulate About the Constancy of
Scoring Temperature, Interdisciplinary Approach to
the Lubrication of Concentrated Contacts, NASA
SP--237, 1970, pp. 153--248.
4. Ku, P.M., Gear Failure Modes -- Importance of
Lubrication and Mechanics, ASLE Trans., Vol. 19,
No. 3, 1975, pp. 239--249.
5. Shipley, E.E., Failure Analysis of Coarse--Pitch,
Hardened and Ground Gears, AGMA Paper No.
P229.26, 1982, pp. 1--24.
6. Tanaka, S., et al, Appreciable Increases in
Surface Durability of Gear Pairs with Mirror--Like
Finish, ASME Paper No. 84--DET--223, 1984, pp.
1--8.
7. Deformation in Fracture Mechanics of Engi-
neering Materials, Richard W. Hertzberg, John
Riley & Sons, New York, 1983.
8. Diesburg, D.E., and Smith, Y.E., Fracture
Resistance in Carburizing Steels, Metal Progress,
Parts I, II and III, May, June and July, 1979.
9. Kern, R.F., and Suess, M.E., Steel Selection -- A
Guide for Improving Performance and Profits, John
Wiley, 1979.
10. Sandberg, E., A Calculation Method for Subsur-
face Fatigue, Proc. of International Symposium on
Gearing and Power Transmissions, Vol. 1, Aug. 30
-- Sep. 3, 1981, Tokyo, pp. 429--434.
Adams, J.H., and Godfrey, D., Borate Gear Lubri-
cant--EP Film Analysis and Performance, Lubrica-
 AGMA 2004 ---- All rights reserved
AGMA 912--A04
tion Engineering, Vol. 37, No. 1, Jan. 1981, pp.
16--21.
AGMA 925--A03, Effect of Lubrication on Gear
Surface Distress
ANSI/AGMA 2001--C95, Fundamental Rating Fac-
tors and Calculation Methods for Involute Spur and
Helical Gear Teeth
ANSI/AGMA 2007--C00, Gears -- Surface Temper
Etch Inspection After Grinding
ANSI/AGMA 9005--E02, Industrial Gear Lubrication
Drago, R.J., Fundamentals of Gear Design, Butter-
worths, 1988.
Dudley, D.W., Gear Wear, Wear Control Handbook,
ASME, 1980, pp. 755--830.
Dudley, D.W., Handbook of Practical Gear Design,
McGraw--Hill, 1984.
Errichello, Robert L., Gear Failure Analysis Semi-
nar, Chapter 3, 2002, AGMA.
Godfrey, D., Recognition and Solution of Some
Common Wear Problems Related to Lubrication
and Hydraulic Fluids, Lubrication Engineering, Feb.
1987, pp. 111--114.
Hunt, J.B., Ryde--Weller, A.J., and Ashmead,
F.A.H., Cavitation Between Meshing Gear Teeth,
Wear, Vol. 71, 1981, pp. 65--78.
ISO/TR 10064--4:1998, Cylindrical gears -- Code of
inspection practice ---- Part 4: Recommendations
relative to surface texture and tooth contact pattern
checking
Kron, H.O., Gear Tooth Sub--Surface Stress Analy-
sis”, Unabridged Text of Lectures, Vol. 1, World
Congress on Gearing, Paris, France, June 22--24,
1977, pp. 185--202.
Lynwander, P., Gear Drive Systems, Marcel Dekker,
1983.
Metals Handbook, Failure Analysis and Prevention
(Failures of Gears), Vol. 10, 8th ed., pp. 507--524.
21
AGMA 912--A04
Milburn, A., Errichello, R., and Godfrey, D., Polish-
ing Wear, AGMA Paper No. 90 FTM 5, Oct., 1990.
Mudd, G.C., A Numerical Means of Predicting the
Fatigue Performance of Nitride--Hardened Gears,
Proc. Inst. Mech. Engrs., Vol. 184, Part 30, paper
12, 1969--70, pp. 95--104.
Parrish, G., The Influence of Microstructure on the
Properties of Case--Carburized Components, ASM,
1980.
Pedersen, R., and Rice, S.L., Case Crushing of
Carburized and Hardened Gears, Trans. SAE, Vol.
69, 1961, pp. 370--380.
Sharma, V.K., Walter, G.H., and Breen, D.H., An
Analytical Approach for Establishing Case Depth
22
AMERICAN GEAR MANUFACTURERS ASSOCIATION
Requirements in Carburized Gears, ASME paper
No. 77--DET--152, pp. 1--11.
Shipley, E.E., Gear Failures, Machine Design, Dec.
7, 1967, pp. 152--162.
Winter, H. and Weiss, T., Some Factors Influencing
the Pitting, Micropitting (Frosted Areas) and Slow
Speed Wear of Surface Hardened Gears, ASME
Paper No. 80--C2/DET--89, 1980, pp. 1--7.
Wulpi, D.J., How Components Fail, ASM, 1966.
Ueno, T., et.al., Surface Durability of Case--Carbu-
rized Gears -- On a Phenomenon of Grey -- Staining
of Tooth Surface, ASME Paper No. 80--C2/DET--27,
1980, pp 1--8.
 AGMA 2004 ---- All rights reserved
(This page is intentionally blank)
 PUBLISHED BY
AMERICAN GEAR MANUFACTURERS ASSOCIATION
1500 KING STREET, ALEXANDRIA, VIRGINIA 22314