"Virgin Bolt Phenomenon"

***written by Superbolt C.E.O. Rolf Steinbock***

Superbolt Inc., is a manufacturer of high power multi-jackbolt tensioners. Multi-jackbolt tensioners (Fig. 1) depend on torquing jackbolts to create tension in a bolt or stud. In order to obtain an accurate relationship between the torque induced into the jackbolts and the tension achieved in the bolting system Superbolt Inc. performed many experiments. The experiments were designed to establish the friction factor for various lubrication compounds and the repeatability of the torque tension relationship in consecutive tightenings.

For most experiments bolts were machined from pre-heat treated A193-B7 material. Typically bolts were machined to a length of just under 12" with a small spherical button on each end so that the length could be measured with a 12" micrometer. The bolts were inserted into a heavy wall sleeve which was held in a vice. Opposite from the bolt head a hardened washer and a multi-jackbolt tensioner were installed (FIG. 2).

The jackbolts of the tensioner were torqued in 10 lb-ft steps to the maximum torque of 60 lb-ft. The length of the bolt was measured at each step. By using the applicable grip-strength, the bolt diameter and the elongation, the tension in the bolt at each step can be calculated according to Hook's Law.

Over a period of several years it was noticed that results obtained were sometimes erratic; this at first was contributed to a varying friction factor between the jackbolt and the nut body. Many different lubricants were tried, but results of the torque-tension relationship remained sometimes puzzling. Most of the time the torque tension relationship was quite consistent, but once in a while there was more elongation in the test bolt than there should be. It was noticed that the unexpected extra elongation occurred only on new bolts and not on previously used bolts.

It was decided to make a series of tests on a new bolt and to carefully record all the tests. A 1 1/2" bolt was machined from pre-heat treated A193-B7 bar stock. The bolt was loaded 5 times with a mechanical tensioner. The tensioner applied was a standard Superbolt MT-150-8/w multi-jackbolt tensioner, lubricated with Superbolt graphite lubricant (JL-G). The results of the tests are plotted in FIG. 3.

As can be seen from the graph, the first time the bolt is loaded, the torque-elongation curve is not linear. It takes less torque in the beginning of the curve than it does in the end. This bend in the curve could be interpreted as a non-linear friction factor; in other words the lubrication is more slippery at low loads than at high loads. This idea is soon disproven when the bolt is unloaded. The bolt does not return to its original length. Any subsequent loading starts at the new base length; the new base length is the length the bolt returned to after the first loading. If loaded the second time, the torque tension curve follows essentially (but not exactly) a straight line, but it turns exactly to the new base length without any further permanent elongation. The third, fourth and fifth loading follows (with small variations) along the same pass as the second loading. The line for the second to fifth loading is not exactly straight. It is a little steeper at the high end than in the beginning. From other tests this can be explained by an increase in friction factor. For most jackbolt lubricants tested, the friction factor increases slightly as the pressure on the jackbolts increases.

The phenomenon described above has been observed on other sizes of bolts and on bolts made from other materials. One test involved a 4.5" stud made from Carpenter Custom 450 stainless steel.

It seems that the above described phenomenon primarily occurs on bolts and studs machined from pre-heat treated materials. In tests made by Superbolt Inc. the non-linear elongation did not show on some commercial bolts that had been heat treated after manufacture. Since the bend in the curve occurs at the low end of the load curve, it has nothing to do with yielding after a material reaches its yield strength. In the above tests the material was not even near the point of yielding.

This writer is not aware of any studies that have been made of the above described phenomenon, and since it only occurs on new bolts, it will be referred to as the "Virgin Bolt Phenomenon".

The writer has made some speculations about the cause of the "Virgin Bolt Phenomenon". The writer believes that when bars of material are quenched from a high temperature, the outside of the bar strengthens first and shrinks at the same time; thus putting the core of the bar under severe pressure. The grains of the material in the core are pushed against each other and stay that way until there is a load reversal. When the outer part of the bar is removed, the pressure on the internal grains is removed, but the grains stay packed against each other. When a tensile load is applied on the remaining core material, the grains relax and reorient themselves to a tensile position. The writer believes that the reorientation phase is a plastic process. This factor explains why the modulus of elasticity normally used for steel does not apply during that phase. As the above tests show, after the grains have been reoriented to a tensile position during the initial tightening, they behave in a perfectly predictable manner and follow Hook's Law.

The above described "Virgin Bolt Phenomenon" can have some very serious consequences. If one would have taken the elongation measured during the first tightening to calculate the stress in the bolt according to Hook's Law, there would have been a substantial error. According to Hook's Law, it takes and elongation of .0154" on the 10.25" long bolt to produce 45,000 psi of bolt stress. On the virgin bolt, it took only 28 lb-ft to produce an elongation of .0154". After the bolt had been tightened one or more times, it took 37.5 lb-ft to obtain the same elongation. The error in the first tightening was 34%. The conclusion is, that one can never rely on extension measurements on machined virgin bolts, and it does not matter how accurate the measuring instrument is. If one wants to rely on measured elongation, it is absolutely necessary to tighten the bolt or stud to the required load once, to unload it again, and then to establish the new base length from which the elongation is measured.

The writer has personally conducted many tests involving machined bolts and studs, and he has observed the "Virgin Bolt Phenomenon" many times. A thorough scientific study of the phenomenon is urgently needed to establish under which conditions it occurs, what causes it, and how it should be dealt with. Of course a study like this is far beyond the means of a commercial enterprise like Superbolt Inc.

The writer hopes that by pointing out the "Virgin Bolt Phenomenon" he may prevent some people from making serious mistakes. He also hopes that some interested and well-funded entity will undertake a serious research program into the phenomenon.