Getting reliability test samples for $2.7 billion Rover

Adam Bahret
Adam Bahret

Check out Part 1 of this Article: A test chamber that is a National Historic Landmark

Reliability Testing needs a bunch of sample units of the product to be tested. Sample test units are stressed at nominal stress or accelerated stress to measure the current reliability (ALT), to remove hidden failures (HALT), or a combination of both (RG). 

Regardless of the type chosen, it needs a bunch of sample units of the designed product to test. What if the product designed is an expensive, one-off system that cannot be replicated?  NASA designed and launched the Perseverance Rover to Mars. It was a $2.7 billion product that required the invention of new technologies, the use of ultra-cleanroom, sensitive electronics, and a vast supplier network. 

How do you get a few spare samples for a project like this to conduct reliability testing? Would the management agree to spend more budget or delay the launch so we can perform testing? Of course, not. Most product development projects are constrained by time and budget and we have to find a workaround.

Here are some ideas: 

  1. Test backup spares: One-off products are precious samples. We cannot test it to take life out of it before launching. In similar situations, we can perform testing on the backup spares that we have. Results from testing the backup spare are used to modify the design or approve the release of the one-off product.
  2. Test a replica: Complex mechanisms like the ‘Sample Collection System’ used in Perseverance are built with extreme care of materials and cleanliness. It is assembled in a cleanroom with conditions equivalent to a medical facility. We cannot contaminate that system’s build with testing. Here, a replica of the mechanism is used to test it in a “dirty room” for verification and validation. The same idea applies to expensive mechanisms. If a diamond tip probe is used in a metrology system, an affordable replica of the mechanism can be used to conduct testing. The goal is to simulate the same stress experienced by the diamond tip probe and monitor system performance.
  3. Test subsystem/component: Testing three units of the entire Rover is impossible. But testing three units of the wheel motor is possible. Breakdown the system into its subsystems or components. Identify the critical sections and perform reliability testing on those alone. When testing is done, the results are put together to form a system-level analytical model.
  4. Test for targeted failure mode: In discovery testing, samples are consumed to identify the severe failure mode from many. To avoid this consumption, a targeted test that verifies the design against that failure mode can be employed. This technique works only if there’s sufficient research in identifying the severe failure mode – either from internal expertise, past reliability records, or research projects.

Development of the Perseverance rover used all the above ideas to varying degrees. These are efficient methods that we can translate to conduct reliability testing in other industries.

Share this post