Sample sizes for Testing Capital Equipment

Does anyone have a good reference or information on how to select a sample size for testing capital equipment?

7 Replies
Trish Borzon
1129 Posts

Hi @Sangita Dave - thanks for joining us here! I'm going to tag a couple of experts here.

@Duke Okes , @Steven Prevette , @Grace Duffy - do you have any suggestions?

Here's a discussion on the topic already that you might find helpful

Duke Okes
192 Posts

That's a very broad question. Can you be more specific (e.g., examples)?

By “capital equipment” I assume you are referring to high cost, low volume production, “one of a kind” sorts of items. Many times you are basically forced into a 100 percent sample size by traditional sampling. When I was in the military I was educated on testing of significant acquisitions such as a warship. The primary approach there is to work with subsystems and sample those systems, along with performance testing of production lines from which the subsystems came from. One also has to plan in how the various subsystems work together. This usually requires some for of acceptance trial, such as “sea trials” following ship construction. Another approach is prediction of lifetime for the components based upon the engineering design of the components and the known wear and tear expected.

Again, this is guessing at the context of your meaning of “capital equipment”.

This may be a useful ASME link Low Volume Manufacturing - ASME

For example a biopsy device using electronics to control mechanics, indicators and vacuum.

I like the testing at the subsystem level idea. For a simple biopsy device, this may not be necessary; however, I can see this being beneficial for a biopsy system with several components (monitor, actual biopsy device, controls, etc.). Thank you for your suggestions!

Thank you Trish!!!

“Credible Reliability Test” planning (CRT) uses field reliability of previous generations of similar equipment, subsystems, or parts, because, although designs may change, processes, shipping, installations, training, environments, and customers don't change. It uses ASQ RD monograph https://sites.google.com/site/fieldreliability/credible-reliability-prediction/ to make a Bayes prior and uses that to optimize costs of units on test, test time, consumer risk, producer risk, and FUD (uncertainty)