How Much Do You Know About Compressive Stress Relaxation? CSR Part 3
Parts 1 and 2 of this series discussed the theory behind CSR testing and what to look for in a CSR result curve. This 3rd and final section will focus on how to use CSR data and apply it to real world applications and how to incorporate it into a material specification.
For the reasons discussed previously, it is important to view a full CSR curve, rather than a single data point, and to resist the urge to draw conclusions from incomplete data. For example, Figure 1 compares a FKM to an HNBR material. Because the fluorocarbon material has a larger viscoelastic loss within the first 24 hours of the test, it appears to be worse (less retained seal load) than the HNBR for most of the test duration. However, the slope of the HNBR curve is steeper than that of the fluorocarbon, and the curves of retained load force cross at about the 2,300 hour point. If these curves are extrapolated, the HNBR is predicted to reach the point of zero residual load force at 4,262 hours, whereas the fluorocarbon is not expected to reach the same point until 8,996 hours have elapsed. Had the HNBR material been selected for this application based solely on the higher percent retained load force observed at 1,008 hours, the end user would have achieved roughly half of the service life they could have enjoyed had they selected the FKM compound instead.
Figure 1: An HNBR and a fluorocarbon in engine oil.
4 Limitations of CSR data
Several caveats remain in attempting to use CSR curve data to predict real-world performance. First, a CSR procedure cannot mimic real-world application conditions. In fact, it is not intended to. CSR testing is meant to be an accelerated aging test, where specimens are exposed to constant temperatures hotter than anticipated in the actual application. While some continuous CSR test machines are capable of temperature cycling, no CSR test can realistically replicate the actual temperature conditions, pressure fluctuations, or vibration seen in a real-world application. Second, CSR testing cannot replicate actual fluid exposure conditions seen in typical seal applications. CSR testing completely immerses a rubber specimen in a small amount of fluid that is kept at a high temperature – typically hot enough and with enough air exposure to cause the fluid to degrade. In a real-world application, a seal is usually installed in a groove with fluid exposure on one side and air on the other. In many applications, the fluid is contained within a sealed system that minimizes oxidative degradation of the fluid. Third, there is as yet no definitive correlation between the onset of leakage and residual load force, regardless of whether that force is expressed in absolute terms (Newtons or pounds-force, for example) or as a percentage of the initial developed load force. Finally, because low temperatures are more challenging for a seal, cycling the temperature to include load force measurements at -40°C would be considered a worst-case scenario. Continuous measurement CSR devices are now commercially available with the capacity for running such a temperature cycle, but their use to date has been extremely limited due to the cost of the equipment.
Several factors should be considered before incorporating CSR testing into a material specification. CSR is a more expensive procedure than compression set, and it requires dedicated and expensive test equipment. As a result, it is not always appropriate to replace compression set testing in a material specification with CSR testing. Second, establishing limits for “what good looks like” should never be done arbitrarily. It is extremely difficult to convert a real world application service life into a number of test hours at a set temperature. Therefore, it is more practical to start with a material that is known to maintain a satisfactory service life in the application, test it for CSR at a relevant temperature, and establish meaningful limits based on the performance of that material. Finally, resist the temptation to get too exotic with specification limits. For example, it may seem valuable to impose specification limits related to the slope of the CSR curve or to the extrapolated “time to zero load force”. These can be valuable tools for gathering engineering insight, but they are far too complicated to incorporate as specification limits. CSR is a more complicated test than compression set, and it requires more “homework” to set proper limits and test conditions.
In conclusion, Compressive Stress Relaxation is a powerful tool for comparing the performance of two or more materials in a particular application, provided it is done on an apples-to-apples basis. It also serves as a reasonable means of (roughly) estimating the long-term service life of a seal material in a given application. There are too many assumptions for it to be used to guarantee a specific level of performance, and there remains much work to be done in correlating results to real-world observations, but it currently offers the best laboratory-scale means of evaluating a seal material for long-term use in a given application.