It depends on whether you are trying to impart a strain amplitude or a stress amplitude into the device. Generally if you want to achieve a desired strain you control the displacement (deflection) applied to the device. Figure I shows a test frame with a multiple specimen test fixture. In that example, NiTi struts are being tested at specific strain amplitude and the system is set up provide an applied displacement. Displacement control lends itself to testing multiple specimens mounted in parallel because you can use a multiple specimen fixture to apply a common displacement across all of the samples. To vary the R Ratio from specimen to specimen you can also adjust the mean applied strain on various specimens while leaving the applied dynamic displacement amplitude fixed. This figure shows load cells mounted in series with the test specimens to measure the resulting applied load. They can also be used to detect specimen breakage. They are optional however and not required to perform the test.
Most people would assume the corollary to be true for applying load or stress to a specimen.... meaning that in an ideal world to control the stress you need to control the load. For a load control test the load cell is mandatory because the system needs to be able to measure the load in order to control it. The test system also can’t control the load without a dedicated actuator to apply the load and for this reason load control tests have typically been relegated to single specimen tests. Figure II shows a single specimen load control test setup. For tests lasting hundreds of million cycles most people don’t have the resources to devote a separate test system to each specimen so they settle for something more practical… the multiple specimen fixture with load cell monitoring.
If your specimen is relatively linear and compliant, you can use the multiple specimen fixture to apply a comparable load across all of the specimens. Take testing a piece of mesh for example. If the mesh is relatively compliant you can achieve a pretty consistent applied load from specimen to specimen by applying the same displacement to each. Lets say your specimen compliance is 2mm/N and you want to apply 0 to 3N to the specimen, you would need to apply a displacement of 0 to 6mm. With compliant specimens your ability to match loads between specimens depends on the variability of the compliance from specimen to specimen. If your compliance varies 5% from specimen to specimen then the applied load will vary 5%.
You can also use displacement control to apply loads across multiple stiff specimens by either mounting the specimens in series or by mounting the specimens in parallel and adding compliant members (aka springs)to the various specimen load trains. Figure III shows what the addition of a springs in series with the specimens looks like. Even though the specimens are very stiff, the compliance created by the springs softens the specimen load train so that a common known applied displacement generates a similar load across specimens. By matching spring stiffness factors across all specimens, the generation of similar load levels can be achieved.
At MDT we try to make testing more affordable for our customers by implementing cost-effective but scientifically sound solutions such as these to increase the sample count on tests on a daily basis.