Our auditor feels that we need to verify a leak tester. It is an old fashioned water tank type tester. The part is submerged in the water. A gas is run through the part. If bubbles come to the surface, the part is scrapped as a leaker. Would showing that the reference part does not "leak" under a standard pressure and standard amount of time do it? I am open to suggestions.

Hello!

We perform two different type leak tests. One of which is a pressure hold test. Given a tank/vessel, we pressurize and hold for a period of time. The trick is determining the criteria. For example, pressurizing to 10 PSI and saying leak criteria is <1PSI/hour, is much easier than pressurizing to 80 PSI, with a criteria of <1PSI/hour.

The second, and one that may be of interest to you, is the helium leak test. We fill the vessel with Helium (maybe in your case, run helium through the part), and use a calibrated helium leak detector to check for leaks.

Boy the detectors we have are sensitive! If there is Helium leaking, it will find it! They only require a few ppm to increase the alert of detected Helium.

Hi Rho Rho,

Are your products manufactured to any particular standards or requirements? If so, you may want to check if there are anything mentioned about 'pressure testing'.

I think what your auditor is not comfortable with is the sensitivity of you 'old fashion' method - but old fashion may not be bad for it's time tested!

Assuming nothing is mentioned about method of testing, it may be a good idea to test to twice the working pressure. For verification purpose, I suggest some modification to the method you have in mind.

First you need a suitably calibrated pressure guage (with suitable resolution also) at the compressed air side to monitor air pressure and any drop in pressure. Next you carry out the test in increments of maybe 20% of the final testing pressure. Maintain the pressure for the standard time which you had decided at every increment in pressure. Record the pressure reading, time, observation of any air bubbles coming out from your part and pressure reading at the end of the cycle (to know if there's any drop in pressure). Repeat until you reach the final testing pressure. Next, repeat the test on a 'rejected' sample. There should be a corresponding drop in pressure upon the appeareance of the first bubble and you may concluded that you tester is sensitive enough.

A word of caution - if there's a drop in pressure guage reading but no noticeable bubbles, it is usually an indication of leakage in some connection points. Soap solution maybe helpful to detect such areas.

Hope this helps.

Regards.