The gradual loss of preload of steel fasteners set into zinc alloy pressure die castings and held at constant temperatures of 50°C or 80°C was continuously monitored over a period of 150 hours using a calibrated load cell. For the commercial alloys Alloy 3, Alloy 5 and ZA-8, the degree of load loss was found to obey a power-law kinetic relationship with time. Choice of alloy had a powerful effect on preload relaxation: Alloy 5 was markedly superior to Alloy 3 with ZA-8 best of all by a small margin.
The effects of changing the geometry of the screws and the methods of producing the threads in the alloy castings were determined for the three alloys in a series of detailed experiments on M6x1 screws torqued to give an initial preload of 6 kN (1349 lbf.). These tests showed that the worst resistance to load relaxation was obtained by using a self-tapping screw in a plain hole in the casting. Pre-threaded castings threaded by a roll-forming tap gave a significant improvement but the best results came from the use of a cutting action tap. An increase in thread engagement length from 10 mm to 15 mm gave another large increase in retained load, but increasing thread pitch in the range 0.75 to 1.25 mm was found to give only a minor improvement.
For a given screw size, decreasing the outer diameter of the boss on the casting in which it was set increases the compressive stress under the screw head, but simulation experiments showed that this had only a small negative effect on preload relaxation.