**4. Discussion**

This is the first study that has experimentally shown that the grease performance as determined by weld load and friction coefficient can be influenced by the speed ramp up time or delay in motor speed—an unknown four-ball test parameter until now. It is fascinating to see that a tenth of second delay in motor attaining the mean speed of 1770 rpm, that is an increase in speed ramp up time from 0.15 s to 0.25 s, had increased the weld load and decreased the friction coefficient for both the greases. Grease weld load represents the inability of EP additives in grease to form a stress activated antiseizure tribofilms that resist sudden rise of local temperature on the steel balls. Seizure occurs when the local flash temperature, reaches the melting point of steel (approximately 1400 ◦C) resulting in welding of four steel balls [13,14]. A tenth of a second delay in attaining the mean speed means the local temperature is still below the melting point temperature of steel. Therefore, we observed an increase in weld load at higher speed ramp up time. It was possible to never reach the local melting point temperature of steel at a speed ramp up time of 0.95 s. This can be attributed to the timescale over which heat is built-up and dissipated that becomes more important than overall energy available at the contact, referred to as friction power intensity [15]. The frictional power loss (units of J/s·mm2) and heat dissipation rate (units of J/mm2) differ with the dissipation rate accounting for the timescale the contact is subject to the available frictional power. The frictional power loss and dissipation rate were calculated using the available data in the four-ball tester:


Frictional power loss per contact area (for all three balls) is [coefficient of friction × actual normal force × sliding velocity]/(total contact area), that results in 206 J/s·mm2.

In Figure 2A (inset) the ramp up time to reach the maximum speed was different, which implies that the available frictional power of 206 J/s·mm2, was dissipated at different rates for the three different speed ramp-up time.

Dissipated heat or heat flow at different ramp-up time can be calculated using the linear method (frictional power × ramp-up time) and the integral of the area within the curve depicted in Figure 2. Please refer to Table 1 for the heat dissipation rates. Both the methods indicate that there was a severe increase in loss of heat (5 times) as the speed ramp up time was increased from 0.15 s to 0.95 s. Thus, the available heat was not adequate for the local conditions to reach to the critical flash temperatures at higher ramp up time thereby leading to a higher weld load. Unfortunately, there are no tools to measure the local temperature rise in a four-ball tester that could have helped us to experimentally confirm this hypothesis.


**Table 1.** Calculated heat dissipation rates measured for the different speed ramp up time.

Antiseizure tribofilms formed on the steel surface depends on the type of additive composition in the greases. Grease Y that had poor resistance to seizure compared with grease X had demonstrated a better behavior and it was equal to grease X. This was again triggered by a tenth of a second delay in motor speed. It was interesting to observe that performance of antiseizure tribofilms in grease Y was highly exaggerated by increasing the speed ramp up time compared to grease X. This is an evidence that the tribofilms can react differently to the speed ramp up time. Although it is interesting to investigate the physicochemical nature of these tribofilms, the focus of this paper has been limited to demonstrating the changes in lubricants performance.

Lubricity of grease as determined by friction coefficient was also exaggerated by speed ramp up time. A tenth of a second delay in motor speed had showed a tremendous improvement in grease friction. This can be attributed to an increase in corrected load. At higher load the steel surface and tribofilms are subjected to extreme pressure conditions that can decrease their surface roughness, that could have resulted in decrease in the friction [16]. In line with this mechanism the grease Y that had higher corrected load compared to grease X also showed lower friction compared to grease X.
