by Lance Schall
Torque is twisting force, whether we are talking about engine performance or tightening nuts and bolts. Torque is measured in foot-pounds. With 50 pounds applied to a one foot long wrench handle, we get 50 lb x 1 ft = 50 ft-lb. Or one pound on a 50 foot long wrench gives the same twisting force.
Torque is only interesting because of its relationship to the really important data; preload. Preload is the tension force in the bolt, a measure of how hard the bolted pieces are clamped together. Preload is expressed in pounds and correct preload is essential to reliable connections.
Even a sturdy steel bolt behaves like a short stout spring and preload can be determined by measuring the stretching of the bolt. As the bolt only stretches a little teeny bit, accurate length measuring equipment is needed. Access to both ends of the bolt is also required. If money is no object, the center of the bolt can be drilled for measurement access. These conditions are usually not practical so automotive folks estimate preload from the torque.
Unfortunately, the torque applied to a fastener (engineerís word for bolt) is absorbed overcoming the friction in the threads and the friction between the bolt head and part. The torque required to obtain a desired bolt stretch varies significantly with these sources of friction. The coefficient of friction depends on the surface finish of the threads, size of bolt head friction area, the accuracy of their machining, and quality of fit to mating threads as well as any lubrication present.
A designerís rule-of-thumb is that from a given population of typical dry bolts, torqued to exactly the same value, the preload (bolt stretching force) is expected to vary by 25 percent (lubricated bolts 15 percent). Still, torque and bolt preload are closely related. This allows the design engineer to provide a torque setting that he or she knows will result in the desired preload. It is an inexact science however, and the designer will always try to accommodate as much safety factor as possible.
The most precise torque wrenches in common use are bending beam style. These use a flat spring steel cross section instead of the less expensive tubular handle. Torque indication is usually by a dial. A quality bending beam torque wrench is accurate to 1 or 2 percent. Next on the list are clicker style wrenches, popular for durability and reasonable accuracy. Also, you do not need to be watching a dial during use which is an advantage if you are working in cramped space. Premium tool manufacturers sell many of these wrenches. Typical accuracy is 2 to 5 percent. SK, Mac, Snap-On are sources for really nice pieces at about 300 US Dollars for a 1/2 inch drive wrench. Craftsman has decent clicker style wrenches for less than USD 100. At the bottom of the market are USD 20 handle bending style wrenches with pointer indicators. Craftsman, Snap-On and others make electronic strain gauge torque wrenches, if you have plenty of money. This is a case of getting what you pay for.
Continuous reading wrenches are better than clickers for setting bearing preloads or other running torques (Shop Manual page M-26, M-39, N-8).
Any wrench is only as good as its calibration. I calibrate my Craftsman clickers often and have been able to maintain an accuracy better than 2 percent. I will take a cheap ancient wrench that was calibrated yesterday over a expensive unit that has been laying on the floor of the shop for several months. Are your mechanicís tools in good condition, clean, and are they treated with respect?
Like most precision measuring instruments, torque wrenches are more accurate when used comfortably within design limits. Torquing at the detection limit or near the peak capacity of the wrench will introduce error. I would knock the top and particularly the bottom 20 percent of the scale off as being unreliable. ANSI Standard allows 4 percent error between 20 and 100 percent of full scale. Below 20 percent full scale, allowable error can be 30 percent or more. My solution was to get the small, medium and large Craftsman wrenches for the price of one SK wrench. Remember to store clicker wrenches with the handle unscrewed to prevent the load sensing spring from taking a set and upsetting the accuracy of your wrench. All beam benders and clickers can work either clockwise or counter-clockwise.
Any socket size or combination of adapters or extensions may also be used. I have seen cautions regarding adapters in torque wrench instructions. This is likely to be a reference to crowfoot wrenches, the open ended wrench adapter for torquing brake lines and the like where sockets can not be used. Improper use of crowfoot adapters can detectably alter the proper torque reading. The proper position for a crowfoot is 90 degrees from the wrench shaft.
U-joints can also cause large errors (proportional to their deflection). As long as the centerline of the torqued fastener extends perpendicularly through the business end of the wrench, you are OK.
Remember to double check each fastener torque. The twisting strain on the bolt effects the pre-load as the fastener is tightened. This twisting deformation of the fastener disappears when the wrench is removed. The second tightening involves less twisting of the bolt and gives a more precise torque reading.
As we all know, a fastener can sometimes be very difficult to remove. It is because of the possibility of exceeding the maximum limit of the wrench that manufacturers recommend against using the tool to remove bolts. It goes without saying, that a torque wrench is a measuring instrument. If your job requires a breaker bar, hammering or an impact wrench, use a tool suited to the purpose.
Mechanical design texts (Mechanical Engineering Design, Shigley, McGraw Hill) give wonderful equations for calculating preload from torques. These equations are rich with phi, theta, mu but offer little enlightenment. The Machineryís Handbook gives us a real appreciation for the many thread standards. What we really need to know is what happens when we put anti-seize on a bolt. The intuitive feeling that a lubricated bolt will be tighter (higher preload) at a given torque is true, but separating this phenomenon from the other causes of preload variation is not easy. The lubricating effect of the liquid will not cause a properly torqued nut to loosen but it can cause you to over-tighten a bolt. It turns out that SAE J1701M gives exactly what we are after.
You can generally use anti-seize on a bolt as you see fit by applying a simple rule of thumb: With anti-seize, the torque needed to achieve the same clamping force is 2/3 to 3/4 that of the dry bolt. A 100 ft-lb torque specification becomes 70 ft-lbs with anti-seize. I would avoid coating the head seating area of a fastener with lubricating compounds. If you really slather the whole works, you could double the design preload or more.
For critical applications this effect must be considered. Luckily, there is enough safety factor in automotive applications so they are fairly forgiving. In order to meet SAE J1102, a wheel stud must have a minimum tensile strength of 120,000 psi. Without calculating it, the stress in a Miata wheel stud is probably half that.
Anti-seize is a paste formed with grease, dry lubricants like molybdenum disulfide or graphite, and soft metal powder such as aluminum, copper, and nickel alloys. At high temperatures, the grease evaporates, leaving solid lubricant and soft metal. The soft metal particles deform and smear over the mating surface of the threads to fill in interstitial space. This prevents contact between iron atoms under extreme pressure. Sharing electrons cold welds the fastener, causing the nut and bolt to become as one. As a side benefit, anti-seize also protects against rust, oxidation, and galvanic reaction.
The Mazda Miata Shop Manual calls for the use of various thread locking compounds and sealants. I always try to use the genuine Mazda stuff. I figure if my car isnít worth a tube of Mazda thread locking compound, it must not be worth much. However, Permatex makes a wide variety of fine substitutes too. In addition to the Shop Manual requirements, I recommend the use of high temperature anti-seize on spark plugs, wheel lug bolts, and brake caliper parts.
Chemical thread locking compounds such as Loctite also protect against corrosion. Remember, they are anaerobic, curing in the absence of air. So donít amateurishly rush to install treated threads. The Loctite will wait for you. The Loctite glue sets, filling the space between mating parts. Thread locking compounds do not have as much of a lubricating effect as anti-seize so torque settings do not change as significantly. However, for successful locking, parts must be clean. Brake cleaning solvents do not generally leave a residue and are ideal. Locking compound on the crankshaft pulley bolt is a good idea.
There are also several ways to mechanically lock threaded fasteners. There are many designs for mechanical locks. One very popular style for bolted accessories are nylon insert locking (Nylok) nuts. These are a good choice, providing low cost, ease of assembly, and good functionality. The additional run-on torque for Nyloks is usually insignificant in terms of torque for proper preload. Presumably, the accessory comes with torque specifications where appropriate. Contrary to some expertís advice, the nuts are acceptable for repeated re-assembly. I discourage the reuse of the original Nyloks only if the nylon locking ring is noticeably worn. Keep in mind that high temperatures will melt the nylon; use split washers for the exhaust manifold.
Crimped bolts lock securely, but can damage studs when removed. The catalytic converter bolts are an example.
For some important applications, the locking of the fastener is separated from its torque by using cotter pins or bending tab washers. On the Miata, wheel bearing retaining nuts and tie rod ends are so locked.
Use Mazda hardware for items like flywheels, brake calipers, seatbelt attachment and the like. Piston rod big-end fasteners and differential ring gear bolts are extremely high stress parts. Always replace these with new (OEM or better) parts if they are disturbed.
I have seen many supposedly qualified mechanics ignore torque specifications, not bothering to look up the correct figure or significantly exceeding it on the assumption that more is better. The best plan is to aim for the center of the specification using the best equipment you have available. Sounds simple, but it is hard to resist giving that wrench an extra tug for good luck.