Determining a Good Variable for Small Scale Weld Integrity Capability Evaluation

[bobdoering]

We are welding a thin tube onto another thin tube of about .125" diameter. The goal is to have a leak proof weld. We would like to develop a meaningful specification to evaluate this weld. Here is the issue:

1. Leak testing is not a predictive measure. Of the continuum of welds that are perfect to asymptotically bad, the leak value (assuming measurement error elimination- c'mon, stop laughing) the leak value is from 0 to 0 - which is very bad resolution for predicting you are about to make a leaking weld!

2. Leak test data, even when there is a leak, is non-normal.

3. Visual criteria have the limitation of visual criteria, and to determine what characteristics are acceptable or not are still tied to the limitation of the leak test.

4. A measure of weld penetration by cross section is limited to the one cross section you take if cut longitudinally. What about all of the infinite other cross-sections around the diameter? If cut across the weld, it is so small smear, etc., may obliterate the void, and the void may be a difficult are to measure if you could clearly identify it. That...and it is only good if there is a void, what about conditions close to, but not yet making a void? Ideally, we would like something with predictability and resolution to avoid failure.

Any ideas?

[Coury Ferguson]

Can anyone help here? This is not my specialty.

[Phil Fields]

In Reply to Parent Post by bobdoering

We are welding a thin tube onto another thin tube of about .125" diameter. The goal is to have a leak proof weld. We would like to develop a meaningful specification to evaluate this weld. Here is the issue:

1. Leak testing is not a predictive measure. Of the continuum of welds that are perfect to asymptotically bad, the leak value (assuming measurement error elimination- c'mon, stop laughing) the leak value is from 0 to 0 - which is very bad resolution for predicting you are about to make a leaking weld!

2. Leak test data, even when there is a leak, is non-normal.

3. Visual criteria have the limitation of visual criteria, and to determine what characteristics are acceptable or not are still tied to the limitation of the leak test.

4. A measure of weld penetration by cross section is limited to the one cross section you take if cut longitudinally. What about all of the infinite other cross-sections around the diameter? If cut across the weld, it is so small smear, etc., may obliterate the void, and the void may be a difficult are to measure if you could clearly identify it. That...and it is only good if there is a void, what about conditions close to, but not yet making a void? Ideally, we would like something with predictability and resolution to avoid failure.

Any ideas?

Hello, what ype of of welding are you doing? Is this a manaul or automated operation? How are you leak testing?

Thank you,
Phil

[bobdoering]

In Reply to Parent Post by Phil Fields

Hello, what ype of of welding are you doing? Is this a manaul or automated operation? How are you leak testing?

Thank you,
Phil

Automated rotational laser welding, followed by automated leak testing - pressure decay type leak test.

[Phil Fields]

In a previous job we performed Orbital (TIG) welding for high purity gas lines (OEM equipment for Data Storage/Semiconductor). We performed a vacuum leak test, with a simple go-nogo criteria.

Phil

[prototyper]

There are non destructive test methods to validate welding, such as radiography, die penetration and magnetic crack detection.
Whatever method you use to check your welding is going to depend on how critical the welding is, what specifications you have to satisfy and how much you want to spend.

I have a product which is Tig welded stainless steel pipework and is covered by the European Pressure Directive. This part has to be pressure tested, x-rayed and checked with die penetrant to satisfy the specifications and legislation.

For other products, such as automotive chassis components, a section and etch is good enough.

If you supply more detail on the application I may be able to steer you in the right direction.

[bobdoering]

The issue is that the size of the weld - .125" inconel tubing - may make radiography, die penetration and magnetic crack detection difficult. As far as criticality, the weld quality is the basis of the leak test passing - and the customer has determined in their mind that it is critical.

Again, the real problem is not finding cracks and leaks, it is trying to find a variable measurable that can be used when the process is about to make leaks, and make an adjustment to avoid it (SPC?). We are also implementing sectioning and etching, but it is destructive and has other issues...like tracking a 3D crack in a 2D section, and so forth.

The holy grail is a predictable variable measurement that is non-destructive. Didn't say it was around, but sure wanted to check to be sure! The reason why this is important is that the customer wants us to put together the specification for the weld - and I would prefer a 'holy grail' type of test , rather than a pass/fail test. Attribute is inspecting in quality, and I would love to avoid that!

[prototyper]

Sometimes the only approach is as per the NADCAP guidelines for welding.
Ensure your equipment is properly maintained, keep tight control on all of the weld parameters, in your case the most critical is probably the amount of heat and the speed of rotation.

Gap condition and preparation of the tube ends will also be key to consistant welding.

Experiment to determine the optimum parameters for the process then have them set in stone!