On Standard Screw Threads
(Part 3 of Nuts & Bolts)
I am still pursuing the question "Do modern metric nuts and bolts come loose spontaneously with greater facility than the nuts and bolts of my childhood; and, if that is the case, why are the modern nuts worse than the British Standard Whitworth nuts of my youth?".
When I raised the question 15 years ago (See ref. [1]) I baulked at the problem of calculating the effect of 'groove-angle' on (a) friction, and (b) the the force causing the nut to unwind. But I pointed my finger at the increased groove-angle as possible culprit for the present casecade of falling nuts. However, buoyed by my success 4 years ago in calcultaing the packing density of uniform spheres [2], I now venture to tackle the 3-dimensional mathematics of the bolt, and its helical thread [3].
I suppose a bolt could have a helical groove with a square section, and indeed there may be some special applications in which such threads are cut. However, the groove of the standard bolt of 19th and early 20th century British engineering had a symmetridal, v-shaped profile, where the angle beweet the upward facing surface and the downward facing surface was 55º. Now, in the 21st century, internatinally, that angle is 60º.
90º⊐ 55º (> 60º (>
(Image from boltscience.com)
Once tightening is complete, there will a tension in the bolt along its axis related (by Hook's law) to the elastic stretching of the bolt. That tension (F, in the sketches below) will be countered by pressure on the surface of the (upper) bolthead and the surface of the nut in contact with the lower plate contact area. But F will also be felt by the entire surface area of the bolt thread that is in contact with the mating surface of the nut, where it will generate friction. It is that friction, proportional to the surface area in contact and to the force pressing them together, that stops the nut from unwinding.
Let me consider first the square-sectioned groove (sketched in Fig. 2 below.) In the middle-top sketch I have tried to draw a square-bottomed grove on the right side of a right-handed bolt, trying to show that, as the thread comes towards us, it dips by a small amount below the horizontal; I have suggested 5º (A pitch of 1.27mm in circumference of 19.95mm gives a 'lead-angle' of 3.65º). In the middle-bottom sketch I have tried to draw the more normal v-bottomed groove, with a similar small dip of 5º.
(Fig. 2)
In the left-top I have suggested that the Hook's law tension (F) operates along the axis of the bolt. It is met, of course, by the nut. I have supposed that the friction that develops between bolt and nut is proportional to the area of the two mated surfaces, and to the force between them that acts normal to the mated surfaces. I have likewise supposed that the force that unwinds the nut operates in the plane of the nut.
From the sketch in the left-top, it would seem plausible that, in the case of the square-bottomed groove, the normal to the bearing surface is only 5º off vertical; leaving most of the force (85/90) remaining to generate friction (and a residual 5/90 of the force to unwind the nut). In the case of the v-bottomed groove it would seem that a large portion of the of the force F operates at the wrong angle to cause friction, 27.5/90 in the case of the BSW nut, 30/90 in the case of the modern ISOmetric nut. (I have not managed to see what that tangental force does; does it stretch the nut? )
Thinking along the above lines I have managed to convince myself that the v-bottomed groove does not maximise the friction between nut and bolt, because the tension in the bolt is not normal to the mated surfaces. And in this regards the modern bolts of the ISOmetric series with a 60º groove are worse than the British Standard Whitworth (BSW) with 55º groove. Incidentally the American UTS series also uses the 60º groove.
Perhaps a more significant difference between the BSW groove and the shallower ISOmetric groove is the smaller area of the mated surfaces in the latter. To show how this could be significant, I have sketched (Fig. 3) an equilateral triangle (sides 2, 2, 2; angles 60º, 60º, 60º), and a second isosceles triangle where one of the angles has been narrowed to 55º. You will see that the friction-generating surface of the latter is 8% larger.
(Fig. 3)
These two effects combined might explain my conviction that the nuts of today are more inclined to come loose than the nuts of the 'Whitworth' era.
(I have also shown that a small element of the elastic tension in the bolt is directed towards un-twisting (loosening) the nut. It is dependent on the ratio of the pitch to the circumference. But in that regard the nuts of the ISOmetric series are very similar to those of the BS Whitworth series, and are occasionally 'finer' (See my Nuts and Bolts1). I have not yet managed to clarify whether the tension in the bolt applied to the sloping side of the groove could have a resultant component that un-twists the nut, like the progress of a sailing boat with a beam wind. I would urge engineers like those at boltscience.com to investigate these questions. )
REFERENCES
[1] https://occidentis.blogspot.com/2010/06/nuts-and-bolts.html
[2] https://occidentis.blogspot.com/2021/09/excluded-volume-between-close-packed.html