Gauge R&R for Destructive Test

[jdm2008]

I am looking to complete a Gauge R&R, on a test that can only be completed once per part(it is not destructive per say, for the purposes of the test it is).
The point of the test is to measure a charactersitic on a device that we suspect may be varying.
Tradional suggestions may be to to use a subgroup and approximate the tradional Gauge R&R method by assuming by making a subgroup of parts identically.
However the whole point of the test is to determine whether the characteristic is consistent from device, and the process may not be capable of making an identical devices in a so this would defeat the purpose.
How would you do Gauge R&R or test validation on this test?

 

[John Broomfield]

jdm,

You may find this Minitsb blog entry and its discussion answers your question:

http://blog.minitab.com/blog/michelle-paret/a-simple-guide-to-gage-randr-for-destructive-testing

John

 

[Miner]

Also see my blog entry on non-replicable R&R studies.

Do you suspect that the device itself may vary from test to test?

 

[Bev D]

Is the device itself a measurement device?

I know you are probably trying to maintain some type of proprietary information but the more information you can give us the more useful - and more specific - advice we can give you...

 

[Bev D]

for example: if the characteristic of the device itself is a function (a dymanic event) rather than a dimension or property (a relatively static thing) and the device can do this function more than once, then we can devise a test that holds the test conditions constant.

 

[jdm2008]

The device being evaluated is not a measurement device. It's something that is activated(and can only be activated once, while keeping the same properties.
The measurement device I will build, will take a measurement on the device being evaluated(the finished device being evaluated and currently being sold).

 

[Miner]

Once activated, for how long does the device retain its properties?

A) Properties are demonstrated for a very short time (e.g., AED)
B) Properties are demonstrated for an extended period of time but immediately begin decaying (e.g., battery)
C) Properties are demonstrated for an extended period of time at a consistent level.

 

[gstewart]

Rest assured that if you go to a commercial laboratory to get something tensile tested, they don't do a new MSA based on your item.
The car companies a sort of losing touch with reality with a lot of their requirements.

We have developed a whole Engineering system over the last couple of years to predict whether things will work or not. That's why we can do stuff like build bridges without having to build 50 other bridges and knock them down to see if they are strong enough.
But science obviously isn't good enough if we have something vitally important like a cup holder in a car.