Torque is the way that threaded fasteners are tightened. The twisting motion pulls the external and internal members (we'll use bolt and nut in this example) together tighter and tighter until at some point they break. Torque is measured in terms of load times distance. Pounds inches (or feet) are used in inch systems, Newton meter (Nm) for metric operations. While torque has been used as long as there have been bolts and nuts, the process is still unclear in many people's minds. What affects it? How much is enough? How do I measure it and with what tools?

The purpose of tightening a bolt and nut is to produce a stretch between the two parts. The amount of force pulling the two parts together is called the clamp load (or pull up load or tension). To keep the joint from loosening, the amount of force pulling the two parts together must be greater than any outside force that tries to pull them apart. This stretch must resist everyday vibrations, as well as random impacts, such as hitting a curb or pothole. If this is done, the joint will last indefinitely. While this sounds simple, the question is, how do you determine the right clamp load, and how do you confirm it?

In addition to clamp load (needed for making the joint tight), there are other torque values used in fastening. Prevailing torque is the resistance caused when installing or removing a fastener. Usually the parts are free spinning, but often a locking feature, such as a plastic patch, adhesive or a deformed section, are added to the part for additional security against loosening. This torque must not be so great as to decrease the torque going into clamp load. To determine the locking device's effectiveness, a test is usually run to determine the limits of its strength. The torque necessary to start the parts moving, relative to each other (called Break Away torque) is the first value measured. Too low a value indicates the locking feature is not doing its job. The torque value measured in the next few turns (Prevailing Off torque) indicates how well the locking feature will resist loosening if it goes beyond the breakloose step. Many specifications call for as many as five removals (some military specs call for as many as 15 removals). This is for assurance that the part, if reused, will still function within acceptable limits.

Unfortunately there are many things in the joint and its installation that can affect the torque and its resultant clamp load. Obviously, friction and those things that affect it (oils, finishes, surface roughness), as well as hardness, dimensional fit, tool speed and type of tooling need to be considered. As could be expected with many variables, the tension produced by a specific torque will not be the same for every bolt or nut tightened. Plotting the results of torque vs. tension will not produce a single line but rather a band of readings. This is what is known as "torque scatter," the plus and minus values around the calculated average.

Why is scatter important? Let's look at a typical torque tension band (see chart at right). The average value looks fine as a single line. However, the factors that we mentioned produce a scatter band that is very wide (A'–A). So wide in fact that at the torque value chosen (X) which was to generate the correct clamp load (Y) the scatter band could now give us so much loading that it could exceed the tensile strength of the metal of either the bolt or nut at one side (Point A) or it could cause the part to become loose due to too little clamp load at the other end (Point A'). In this example, the engineer had determined that any load below line B will cause the part to vibrate loose.

The correct method to determine the right torque is to run actual tests in a laboratory using production parts and production tooling. Except for all but the most critical of joints, this is neither practical nor inexpensive. Torque tables, derived from hundreds of joints with various conditions of finish, are often used. While not accurate to the ninth degree, they offer a good ballpark guess at where the torque value should be. When a torque problem occurs the first thing is to check the value being used. If it is near the value given in the table, then the problem is probably due to another cause.

When looking at some other causes, you should consider how much the tool affects the clamp up. All installation tooling has variation, the amount being controlled by the type, speed of operation, conditions of assembly, etc. For this reason, all specifications are given as a range (X-Y or a value ± %). Ordinary air powered tooling is accurate to ± 15-25%. Electric tooling is a bit more accurate (more in the range of ± 12-20%) but the cost of tooling and energy charges makes assembly with DC nut runners unprofitable in some shops. Tools are adjusted by increasing or decreasing air (electric) pressure until the "gun" stalls at the torque set on test stand. This type of tool is used for joints that do not require any great accuracy.

Impact guns are about the worst, with accuracy being anyone's guess (probably in the range of ± 30-40%). Most major OEMs do not allow the use of impact guns in their assembly plants.

The next step in accuracy is torque-control tooling. These tools monitor the torque via an adjustment (set screw, etc.) that is turned to meet an adjustment factor determined by the tooling department. Accuracy is in the area of ±5-15%. The tools offer better control of the torque but at a higher cost.

Torque monitoring tools are as accurate as torque control tools, except that they have the capacity to record the actual torque values delivered on the line. They are used in areas where records of the actual values are needed (critical joints such as steering and suspension systems) for legal or other reasons. Many times this kind of tooling will be tied in with a "Green/Red light" system for further control in the assembly operation.

The most expensive and accurate tightening system is torque control system tooling. This torque operation is a complete system with gun, electronic transducer readouts, controllers, and data bank recorders that measure the torque as the joint is being tightened and feeds the information back into the control panel. The system will estimate where the proper turn off point is (as set by the specification) and stop the tightening process within ± 3-5% of the set value. This type of torque control tooling is the most expensive and is used for critical attachments such as wheel lugs. The operator simply pulls the "trigger" and holds on until the tool turns off (usually with a green light to indicate "Good Joint").

Hand torqueing, often used for inspection after the assembly, is not very good either. The best of operators has been estimated to be about 8 percent off when fresh on the job. When coupled with most hand wrenches (±2-3%) the amount of variation is quite a problem.

Tool speed contributes to inaccurate torque answers. The coefficient of friction is different between a bolt run at 15 RPM in a laboratory test stand and one run at 600 RPM on the assembly line. It is a common mistake for engineers to base their assembly torque values upon laboratory tests run with hand tools and slow laboratory speeds. The use of a new type of tooling, the Hydropulse gun, has caused some concerns with speed of run down. The tool consists of a very rapid (up to 8,000 RPM) rundown that changes to a pulsed slower impact-type tightening after a certain set point of torque is reached. The very high rundown, especially in situations of frictional resistance, can cause meltdowns as well as jamming and seizure of overheated parts.

Other factors that can change the tension produced by a specific torque are those that change the surface condition of the mating surfaces. One is the hardness of the parts and mating surfaces. If too hard, the surface is smoother, and smoother means less friction. As mentioned above, less friction in one area increases the loading in tension. Likewise, surface roughness is the same case. Rougher is more frictional to parts sliding (threading) over each other. The material of the joint should be an obvious factor but is often overlooked. Aluminum and soft metal are more frictional. Enough said? Dimensional differences (loose fit, tight fit) will obviously affect the way the parts go together.

Finish is a major contributor to torque problems. On one hand, the thickness contributes to tighter fits by filling up what little interspace there is between male and female threads. On the other hand, many of the new finishes today (used for higher corrosion resistance) are multi-element chemicals that may contain lubricants, Teflon-like compounds and torque modifiers for decreased driving torques and ergonomic considerations on the plants. This would work to decrease friction and raise tension loads. Conversely, some finishes show increased friction when used against each other.

Lubrication, both wanted and unspecified, causes torque tension problems. Many parts are tested without lubes and then cause concern when the assembly plants cannot meet values. Also, the line operators often think that the parts "go in faster and easier" when they are dipped into a can of oil/grease on line. Good, except this is not the value that was used to calculate the specification on.

When a torque problem is addressed the questions to be asked are:

• What are the torque values specified?
• What are the conditions of the assembly? This is the mating parts, [their grade, thread tolerance, hardness]. What finish are they both? Many times the presenter thinks that it is a fault of a part (bolt/nut, etc.) when it is almost always due to an interaction of factors.
• What tools and speeds are being used?
• How close to the limits of the material of the parts is the torque specification? The old fastener saying that "If tight is good, tighter is better" has been our albatross for years. If the joint appears to be loose let's just increase the torque specification.

The correct and exact torque value is best determined by laboratory testing. In the case of a critical joint, it is probably safer to suggest that a laboratory and fastener expert examine the problem than risk making a judgment call.