Fillets

In reply to a previous article that discussed crankshaft fillets and their importance in terms of crankshaft reliability, a reader left a comment saying most crankshaft fillets are not true radii, but are in the form of an ellipse or other conic section. Although I’m not aware of anyone offering elliptical or conic fillets – or indeed anyone specifying them – it is certainly possible, with modern computer numerical control (CNC) manufacturing techniques, to specify exactly what form of fillet is desired.

To produce a fully ground elliptical or conic fillet requires either the use of a specially dressed wheel (‘dressing’ is the process of producing the desired form on a grinding wheel) or to have a grinding wheel with a suitably small corner fillet and to control the position of the wheel accurately during grinding. There is a separate cause for concern though if we are trying to produce an elliptical or conic fillet on a surface-hardened crankshaft, in that we need to accurately ‘offset’ the final profile during the manufacturing stage, meaning that there is an extra emphasis on precision. The use of specially dressed grinding wheels is nothing new in crankshaft manufacture, as ‘barrelled’ main bearing journals have this requirement.

If we are to produce any form of ellipse using conventional turning, the operation is pretty straightforward if using CNC machinery, which is not restricted to following simple radii. However, in recent years, as explained in a recent article on crankshaft manufacture, there has been a marked trend toward the use of CNC milling to rough and finish crankpins. Here, the options for producing the required conic section fillet are to use a cutter with a smaller corner radius and to generate the corner radius by careful control of the cutter, or to use a cutter with the desired corner profile ground onto it. Either method would work, but the latter is perhaps preferable in terms of reducing machining time.

So, if we can produce an elliptical or conic section fillet, would we want to do so deliberately? Well, there are some pointers in standard texts on the subject. “Peterson’s Stress Concentration Factors” refers to work from the 1940s which examined the stress concentration factors for a flat stepped bar in bending for elliptical fillets of varying ‘aspect ratios’ between 1 (that is, a simple radius) and higher ratios (ellipses). For this simple case there appears to be a clear advantage with the elliptical fillet, with the theoretical stress concentration factor, Kt, being lower for elliptical fillets, although the magnitude of this advantage is generally blunted somewhat as the various factors are applied to give a fatigue stress concentration factor. However, it should still present a quantifiable advantage.

Fortunately, living in this age of powerful computing, we are able to simulate more complex load cases and geometries quickly using FEA. A ‘quick' and dirty’ comparison using geometry approximating a simple, low-overlap crankshaft with no boring of the pins or main bearings showed that there was no advantage to be gained by using elliptical fillets, although for this particular geometry there was surprisingly little disadvantage owing to the geometries simulated. The case was one of bending, and the fillet ellipses simulated were a) semi-major axis = semi-minor axis = 3 mm (that is, a plain 3 mm radius); b) semi-major axis = 2 x semi-minor axis, semi-major axis = 3 mm, parallel to crankpin axis; and c) semi-major axis = 2 x semi-minor axis, semi-major axis = 6 mm perpendicular to the crankpin axis.

Although there exists the manufacturing techniques to make crankshafts with elliptical or conic fillets, from a very limited simulation there appears to be little incentive to make such a crankshaft.

Written by Wayne Ward