Lack of fusion is the defect that makes welding inspection difficult, because it has no reliable surface signature. The bead can be the right height, the toes can be smooth, the colour can be correct, and the joint underneath can be two pieces of metal sitting against each other with nothing joining them.
What makes it worse is that the defect is often small in area and continuous in length. A joint can be fused for most of its circumference and unfused along one side, and nothing on the outside will show it. The part looks finished and behaves like a part that was never welded.
This is why procedures, fit-up standards and destructive testing exist in the first place. They are not there because welders are careless. They are there because the most consequential defect in TIG work is also the least visible.
The classic location is the side wall of a groove, where the weld metal sits against base metal that never got hot enough to melt. The pool wets the surface, freezes against it, and forms a joint that is mechanical rather than metallurgical. It holds until it is loaded.
The root is the second location. A root that never tied in leaves a gap under the bead, and on a tube that gap is a leak path and a crack starter. The face of the weld can be perfect while the root is open along its whole length.
The third location is between passes. A bead laid onto a previous bead that was not cleaned, or that had cooled into a shape the arc could not wet, can sit on top of it without fusing. That is why multi-pass joints fail in the middle of the deposit rather than at the edges.
Fusion requires the base metal to actually melt, which requires enough energy to reach it and stay there long enough. A long arc, a wide arc or a fast pass can all deliver heat to the surface without pushing enough into the joint wall.
The reason this is so common on thin wall is that the welder is deliberately keeping energy low to avoid burning through. That caution is correct, and it produces a bead that sits on the surface rather than in it. The compromise is a tight, focused arc at a current high enough to melt the wall and a travel speed fast enough to leave before the wall collapses.
Current alone is not the answer. Adding current with a long arc simply widens the bead and spreads the energy further from the joint. The useful combination is a short arc, a small pool and a deliberate effort to point the arc at the joint line rather than at the middle of the gap. Our titanium fab components pages show the kind of assemblies where that matters most.
Wetting is the visible sign that the base metal has melted and the pool has flowed onto it. A properly wetted toe has a shallow angle where the bead meets the parent metal. A toe that stands up steeply, or that shows a line where the bead ends abruptly, was never fused.
Travel speed controls whether wetting has time to happen. Moving too fast means the arc passes before the wall has melted, and the metal freezes as a shape sitting on the surface. Moving too slowly means the pool grows and the weld becomes a puddle that overheats the wall without necessarily tying in at the edges.
Watching the toe rather than the crown is the habit that prevents this. The crown can be dressed afterwards. The toe is the joint, and it is either wet in or it is not.
A gap that is too tight gives the arc nowhere to go. The two edges sit against each other, the arc plays across the top, and the welder fuses the surface while the joint line below stays cold. The result looks like a weld and is a seam. Our stainless steel exhaust system pages show joints built to avoid that.
A gap that is too wide is the opposite problem and produces the same defect from a different direction. The welder fills the channel with rod, the pool sits in the middle, and the side walls never see enough energy because the metal is being added faster than the arc can melt the parent.
Both are fit-up failures that present as welding failures. The welder can sometimes compensate, and the compensation is always a compromise on fusion. Getting the gap right before the torch is lit is the only reliable fix.
On titanium and on stainless, an oxidised surface will not fuse. The oxide sits between the pool and the base metal, and the pool wets the oxide rather than the metal. The bead looks continuous, and the joint is a layer of ceramic.
This is why cleaning between passes matters so much on multi-pass work. A pass that has been left with a coloured, oxidised surface must be removed before the next pass, not welded over. Brushing is enough on light oxide; heavier discolouration needs to be ground back.
On titanium the same principle applies to the shielding. A joint that lost gas coverage while it cooled has an oxidised layer that the next pass will not fuse through, which is how a part can pass a visual check and fail a bend. Our titanium exhaust system work treats interpass cleaning as part of the weld cycle.
Visual inspection will not find lack of fusion. It will find the surface consequences of the process that caused it, which is a different and weaker thing. Finding the defect itself requires either destruction or a volumetric method.
Bend testing exposes it on a coupon, because the unfused side wall opens into a tear as the strip is folded. A macro section shows it as a straight line where the weld metal meets parent metal with no fusion boundary, and it shows how much of the wall is affected.
On production parts, dye penetrant will find surface-breaking lack of fusion and nothing else. Radiography and ultrasonic methods find it internally, and both need a technique written for the geometry rather than a generic procedure. The honest position for most fabricators is that prevention is cheaper than detection, because detection at this level costs more than the weld.
Because this defect evades the eye, the questions that matter are about how the supplier prevents it rather than how he inspects for it.
Tigindustry has been building welded assemblies in Shanghai since 2009, with more than 50 employees across four factories and a plant of about 53,819 sq ft, and fusion control is treated as a fit-up problem as much as a torch problem. You can see the range on our TIG titanium welding parts and exhaust pipe pages, or browse by alloy under shop by material.
