Sample size for IEC 60601-1-11 environmental tests


Dear Friends
similar to my last question - but for a different standard,
I am trying to understand what is the proper sample size for environmental tests when using the IEC 60601-1-11 standard.
where can I find a reference to a sample size for it?

Thank you all


Peter Selvey

Super Moderator
Sample size itself is not generally mentioned in type testing.

The basic requirement (as per the definition in the standard) is that a type test should be representative of production. In theory the number of samples required to be confident about compliance depends on the expected variability of the test result and the margin to the limit. For example a transformer temperature test run to stable temperatures result might be 72C, and the limit is 100C. In this case, a single test is fine. On the other hand if the result is 97C, and the limit 100C, would could not be sure about compliance in regular production.

In safety related type testing the situation is made complicated as many of the requirements in the standard are conservative, and the test itself may be based on multiple worst case assumptions (worst case load, highest ambient, longest time etc). In such a case, a single test may be fine by virtue that potential variability in production has already been taken into account when developing the test method and limits.

For IEC 60601-1-11, the main point is performance. Many performance specifications are also conservative, but usually there will be one or two which are critical and need close attention to making sure a type test is representative. For this we should fall back to basic principles. A home use thermometer for example, usually requires a precision resistors in the measurement circuit. If the manufacturer is cheap and uses 1% 250ppm resistors and other low quality parts, it is possible that the environmental tests pass but the results show up significant changes in measurement accuracy at 0C and 45C. if the result is close to the limit, we cannot be confident about regular production. More samples are required (or better design ...).

On the other hand if the manufacturer uses 0.1% 50ppm resistors and other quality parts, and calculates the error budget and confirms it is well within the limits at the temperature extremes, and tests at 0C and 45C show no detectable change, one sample may be enough.
Top Bottom