Aluminum "cold roll" weld defect 1

[IFRs]

I've heard the term "cold roll" used to describe an aluminum fillet weld defect. It seems to be where the toe of the fillet is not concave but exhibits a profile more perpendicular to the material being joined. Does this sound familiar, is there a better description out there? We are using pulsed push-pull MIG traveling on a geared track welding 0.063 5052-H32 to 0.20 6061-T6 using 0.035 and 0.125 4043 straight Argon in a well fixtured shop environment. We do about 720 feet per day per fixture, this appears to be a common issue. Not sure how to avoid it, looking for guidance on if it is a rejectable defect and what the consequences are of leaving it there ( it has been suggested that the 0,063 base metal will fracture at the toe of the weld where cold-roll exists ). Thanks!

 

[Ron]

This is "cold lap". Where the toe does not fuse, but the material "rolls" over the surface as is cools. Has to do with technique (slow down!).

 

[tbuelna]

The condition you describe is called "overlap". The toe of a fillet weld should exhibit neither excessive convexity nor undercutting. Overlap is due to the base metal at the fillet toe not being heated enough to fully overcome the surface tension in the filler metal puddle allowing it to diffuse out smoothly at the boundary. Overlap is not quite as extreme as a cold lap. With a cold lap there is no fusion between the filler and base metal, and the filler is simply laying on top of the base metal. With overlap, there is fusion between the filler and base metal but it is mostly superficial, and the abrupt transition in cross section at the excessively convex fillet weld toe can create stress concentrations.

Based on the description you provided of your fillet weld joint, specifically that you have a large mismatch in thickness of the two parts (.063" and .200"), it seems a bit odd that you are having an overlap condition at the fillet toe of the thinner part but not at the fillet toe of the thicker part. Can you provide a sketch of the weld joint?

 

[IFRs]

Your comments are very helpful. In my 37 years of making aluminum parts, I have not seen this ( perhaps not looking for it ) and this is the first situation where it may contribute to a failure. I have fixtured and welded many miles of this type of joint, mostly with 0.040 thick welded to 0.090. Attached is a crude drawing of the weld joint ( desired on left and cold-roll or overlap on the right ) and two pictures of one section where this defect was found. This is a particularly egregious example, in a hand welded section. The backside shows almost no telegraphing, the .063 material is not cracked although there is a shadow line. In most cases, the machine weld is very consistent and good looking ( as opposed to this one ) and the weld transitions into this basic condition and then back out. I'll get pictures of a machine welded section tomorrow.

 

[IFRs]

Rats, the attachment did not work, I'll try again...

 

[racookpe1978]

Slowing down will help. You can also try angling the wire feed a little more vertical so the "heat" is directed slightly more into the lower metal section. Not too much! You do want the heat to go into the upper metal as well, but as you see in the left sketch, that upper corner is already melting sufficiently to slump over and go down.

 

[metengr]

http://www.twi-global.com/technical...ed-joints-a-review-of-the-practicalities-066/

 

[metengr]

A reference book for welding;

http://www.kobelco.co.jp/english/welding/events/files/2011Defect.pdf

 

[gtaw]

Try wire brushing the joint just prior to welding.

Aluminum oxide melts at a much higher temperature than the base metal. It isn't a cure-all, but it cannot hurt.

Best regards - Al

 

[Ron]

With a stainless steel brush, of course!

 

[gtaw]

Yes, with a clean stainless steel bristled brush. Good point!

Best regards - Al