Photos first ------->
http://imgur.com/a/Ubu7ZThis is just the ROOT PASS in the photo
This was a project I was on not too long ago that called for CJP welds on 3" 1/4 wall square tubes with a reinforcing fillet weld. This was only meant to hold 180 pounds but who cares its only your tax dollars paying for the aluminum! Anyways the spec called for full pen and it had photos of examples where there was just a sign of full pen the joint was ok. Now I had to borescope all these joints from the inside through drilled holes and a small wire camera. These are the samples prior to the project starting.
Now my question. If a weld is too convex it is rejected on profile. Now what about the penetration being so much that it acts as a stress riser of a weld just the same? Aluminum is kind of a ***** in this type of welding situation and the root itself doesnt actually lift up.
Curious your input everyone.
-J
If that is as you say a CJP T-Joint
Reject for overlap.
No need to measure convexity, or look at profile examples...
If there is a code standard where that condition is not considered overlap, I'm unaware of it.
That is the root pass only in the photos. So you would for sure call that overlap even if it is outside the weld profile entirely? I can see this two ways.. If you were to look where the weld actually is... as in the joint where the full pen weld goes and drew the line at the CJP then it is still good if you don't count the excess. The other I can say its a giant weld pouring out the back! but about that being overlap.. over lap would have to come from the corner or possible the flat edge right? this is coming straight off the back plate. even an air gap under it. Just making odd looks at it haha.
J
To me it look like all the penetration was on just one side of the base material.
M.G.
That reinforcement in the pictures is *not* outside the weld profile entirely.
It is the weld ! The weld "actually is" where the welder put it, and therefore shall be inspected where the welder put it.
Overlap does not have to come from one place or another... It simply is there or it is not there. It is always rejectable with zero tolerance in any of the AWS "D" codes.
The process is so out of control that they are placing the "weld" in a place it does not belong and in a shape/profile that is rejectable.
Edit:
That is a very interesting joint... I've never seen an aluminum CJP single bevel T welded from one side.
Is the process GTAW or GMAW?
What was the root face measurement prior to welding on the beveled base metal piece?
GTAW, glad I could show you something new haha
Even though our codes don't specify 'pretty' or 'ugly' when it comes to welds....SOMEBODY HIT THAT BABY WITH A BIG OLD UGLY STICK!! WOW.
I don't see how that could ever be acceptable at the root with the ID material appearing as it does. Take that sample and cut it so you can bend one side over backwards. I'll bet the results will speak for themselves... LACK OF FUSION at the very least.
Brent
![](/mwf/attachments//avatars/4269-2369.jpg)
In addition to the problems already noted with all of that melt thru piled up on the inside, I'm afraid there are problems in the inside corners as well....those appear to have the opposite problem.
The thing that really gets me about stuff like this is that there were prolly PQR's done prior to this test that included tensiles. Where the heat input was carefully controlled, in a more typical groove joint, to get acceptable mechanical properties from the base/weld metals.
Then they throw all that heat input control out of the window with a production mock-up, and if the macros would have come out nice, they would have put a stamp on it and started producing a product that could not possibly have the mechanicals in the weld zone and HAZ that was originally engineered and proven with the tensiles.
I understand that the original post stated that the assembly must hold only 180 lbs. Ok, than why the complicated CJP from one side?
We don't know why is was so over built, that's what their customer wanted. They couldnt get that root sealed without blowing out the back like this. it would fill out but not on the end runs into the corners.
For a split second I thought this was a friction weld, but then realized there were square corners....
No one has asked the magic question: What welding standard applies?
Most welding standards address the weld profile one way or another. They may call it root convexity or root concavity, melt through or concave root, reentrant corners, face or root reinforcement, something. As pointed out by Brent, we may not have criteria for ugly, so we have to assess the weld using criteria provided by the applicable welding standard.
Many of the comments posted have merit with regards to the cause and consequences of the root condition. I would point out that the joint may have been cleaned insufficiently, i.e., oxidized metal, and excessive amperage in an attempt to ensure CJP. The shape of the root surface isn't that unusual for aluminum when the root surface is not provided with inert root shielding gas. Granted, aluminum can be welded without root shielding gas, but the surface contour is going to suffer. Incomplete cleaning, i.e. vigorous wire brushing with a stiff bristled stainless steel brush, just prior to welding can cause incomplete fusion.
The acceptability of the surface condition of the root is somewhat dependent on the nature of the applied stress on the connection. If the weldment is subject to a static load, the "ugly" root surface will have little if any effect on the performance of the connection. If the weldment is subjected to cyclic loads, it is another matter than needs to be addressed to ensure CJP is achieved and the surface contour meets the acceptance criteria. Of course all is for naught if the welder has insufficient skill and employs poor technique.
The fact of the matter is that the inside of the small tube is probably not accessible for VT, so how does one check the root surface condition other than in the laboratory during the qualification process? All sorts of defects can exist with no one the wiser if it cannot be examined visually. Another Alfred E. Newman moment.
Best regards - Al