Circumferential Weld Start & Stop Location Overlap

[kcme2005]

I have a GTAW butt weld at the joint between two cylindrical SST peices (diameter approx 3/4"). Of 11 units tested, one failed a static pressure test with a crack forming in the weld. The location of the failure point is interesting to me as the crack is positioned immediately beyond the end point of the weld.
Our welders do have a slight "ramp up" and "ramp down" at both the beginning and end of the weld. My gut tells me that there simply was not enough overlap between the start location and stop location of the weld to get full penetration, and thus the weld was structurally weaker in that location.
My question for you all is this: Is there a standard distance that our welders should be overlapping from the start location to the stop location? 45°, 1 in, any known standard?
As always, any suggestion is greatly appreciated.

 

[Mr168]

Was this an orbital weld or was it done manually?

Complete penetration has nothing to do with overlap. If the welder did not adequately tie in his start and stop at the root, then that is a welder induced defect, not something that would need to be address procedurally. This poses even less of an issue if tack welds are being made and prepped appropriately, as your final tie in on the root itself will typically come into a fully feathered edge to assure it is entirely consumed.

As you travel around the circumference, you want to maintain peak amperage until you have tied into your start/tack enough so that you do not have a low spot in that layer's deposition, whether it be stopping or ramping down to early/abruptly, or leaving a crater at the end of the weld.

Do the welders continue to add filler metal as the amperage is ramping down? Doing so can create a lack of fusion/cold lap, which may lead to cracking.

How about overall deposition? If the root pass is deposited too thin or washed excessively, you may get cracking as the result of shrinkage forces. While I doubt this is the case with such a small diameter, it is possible if the root is performed autogenously.

Was the joint properly shielded? Rapidly forming oxides/sugaring can create inclusions that cause failure on tie-ins.

Rarely have I seen the actual amount of overlap have any bearing on weld quality as long as the layer deposited is of uniform thickness and is properly tied in.

 

[kcme2005]

Thanks for your response.
To answer your questions, this is a manual weld.
The parts are held together with threads and torqued to a set value, so there are no tack welds needed.
Also, each part is made with a square edge to mate cleanly, so no filler metal is added in the weld.
Thanks again for the info.

 

[Mr168]

Were the welds made using a rotating positioner, or were the parts fixed? Positioners tend to be beneficial in maintaining proper torch angle in situations like this.

If you plan on doing a large run of these parts, it may be worthwhile looking into renting some equipment from AMI or Magnatech for autogenous tube welds. They are absolutely ideal for situations like this, and can result in an extremely low reject rate.

 

[stanweld]

Is the tube seamless? If so the autogenous weld may be subject to underbead cracking/microfissuring. The overlapping weld could have cracked the underlying deposited weld metal. And since the weld is essentially a seal weld, small cracks formed may have led to failure during test.

Have you discounted porosity as the initiator of the fracture? Porosity can be common in seal welds, especially in threaded components due to out gassing during tie-in.

 

[chicopee]

From You description of the butt weld prep- square instead of bevelled edge- it sounds as if you are welding two pieces of metal with reinforcement only. Is that right?