A computer controlled quenching simulator for observing phase transformation kinetics in metals by electrical resistance measurements

The commonly used method of dilatometry to study the martensitic transformation leads to inaccuracies in the measured transformation temperatures and kinetics due to temperature gradients within the sample. Samples studied using commercial dilatometers are often of insufficient size to allow mechanical testing. This thesis details work to design and construct an instrument that overcomes the limitations of commercial dilatometers for the study of the martensitic phase transformation in steel. An introduction to why the study of the martensitic phase transformation is important to the new generation of automotive press hardened steels is presented. A review of the techniques that have been used in the literature to study the martensitic transformation kinetics including metallography, dilatometry, resistivity, thermal, acoustic emission, X-ray, and neutron diffraction follows. A case is then presented as to why electrical resistivity measurements may be the best technique for the small scale metallurgical laboratory. Details of the construction of a machine that uses phase angle control of a 2000 A alternating current for the Joule (resistance) heating of samples suitable for thin sheet tensile testing are presented. The heating system of this machine has a predicted maximum heating rate of 50°C/s for 2 mm thick steel. The constructed machine is able to perform the heat treatments either in vacuum or an inert atmosphere, and uses argon quenching to achieve a cooling rate in excess of 40°C/s for a 2 mm thick sample. The heat treatment is fully computer controlled. The results show that the phase transformations in pure cobalt, pure titanium, and a eutectoid steel are clearly observable using electrical resistance measurements. The effect of the transformation kinetics on the measured transformation temperatures is summarized. A detailed analysis of the effect of the common mode electrical noise of the small measurement voltage is presented, along with suggestions about how to improve the accuracy of such measurements