August_AMP_Digital

joint (Fig. 1c)

[3]

. As indicated, loading on the bolt is mini-

mal until the working load exceeds preload. Thus, for a

condition where cyclic loads are smaller than preload, lit-

tle or no cyclic damage occurs to the fastener. The concept

of higher preloads resulting in increased fatigue perform-

ance may be counterintuitive, which might otherwise sug-

gest that increasing the tightening force of a bolt may

increase propensity for failure. Insufficient preload has

been attributed to the cause of the majority of fastener fa-

tigue failures

[4]

. Therefore, understanding preload is criti-

cal in both the design and failure analysis of fasteners.

Pretension is often adjusted by applying torque to the

bolt, screw, or nut. Torque application increases tension in

the fastener and clamping force between components. It is

necessary to understand the difference between clamping

force and the amount of torque applied to the screw or

bolt. These two metrics generally correlate, but are differ-

entiated by frictional forces. Fatigue susceptibility can be

reduced by increasing clamping force and fastener tension,

which is most often controlled by applied torque. However,

factors such as material hardness and surface finish affect

friction

[3]

, and therefore affect the extent to which applied

torque is translated into increased fastener tension and

clamping force.

Guidance is provided with respect to the degree to

which fasteners should be tightened. Bolts in tension

should exert as much force on the joint as it and the joint

members will stand

[4]

. A rule of thumb is to load the bolt to

60% of the material’s yield strength

[5]

While tension and high clamping force are critical for

joints, applied forces (often cyclic) can significantly reduce

preload. Generally, increased preload reduces the propen-

sity for fastener loosening. To minimize susceptibility to

loosening, the highest preloads possible should be used—

up to, but not exceeding yield

[4]

. Loosening reduces pre-

load and the force required for joint separation, and can

greatly increase the propensity for fatigue failure and fret-

ting. However, increasing the preload also increases sus-

ceptibility to overload upon installation, or may increase

susceptibility to SCC.

Excessive preloads can also increase mean stress to lev-

els that may contribute to fatigue for joints subjected to rel-

atively low cyclic stress amplitudes. The effect of increased

mean stress on fatigue performance is often described by a

Goodman line, which plots stress amplitude as a function

of mean stress. A schematic Goodman line is shown in

Fig. 2, where increasing mean stress results in decreased

ability for the component to withstand applied alternating

stress. As shown by the Goodman line, excessive mean

stresses (near the fastener tensile stress) can reduce fatigue

performance. However, in most joints, high preloads pre-

clude significant cyclic stresses, as shown in Fig. 1.

Hex-head case study

A short case study highlights some common issues ob-

served in fastener failures. In this example, multiple broken

steel hex-head screws were submitted for failure analysis.

These screws had been part of equipment used in a silicon

wafer fabrication system. The screw service environment

was in a cleanroom, with exposure to no known corrosive

media. In addition to the failed screws, material specifica-

tions, drawings, and installation instructions and specifica-

tions were provided. Representative images from some of

the screws are presented here.

Laboratory analysis

Through the course of the investigation, it was estab-

lished that the screws in question were installed with a tor-

sional specification of 50 in.-lb. During service, screws

were subjected to cyclic bending loads. Screws had the fol-

lowing specifications:

• UNRF 10-32 thread-form

• 38-43 HRC

• Minimum yield and tensile strengths of 162 and

180 ksi

• Max recommended load: 3150 lb

• Recommended torque setting: 79 in.-lb

• Material composition in accordance with ASTM

A 574 specifies 10 different steel alloys, the majority

of which are 41xx-series steels.

Screws were analyzed to look for macroscopic indica-

tions of failure mode. As shown in Fig. 3, fracture occurred

in the threaded region. The fracture plane was generally

perpendicular to the screw length, with no significant plas-

tic deformation. Given that the screws were known to be

loaded in tension and bending, fracture perpendicular to

ADVANCED MATERIALS & PROCESSES •

AUGUST 2014

Fig. 2 —

Schematic plot shows the typical relationship between

alternating and mean stress.

0 20 40 60 80 100

Mean stress, ksi

Stress amplitude, ksi

Fatigue limit

or fatigue strength

Synthetic Haigh diagram

with Goodman line

Goodman line

Ultimate

strength

Alloy steel 90 ksi ultimate

tensile strength

Goodman line

represents uncorrected

estimate

at 50% failure

Criterion: Separation