A TIG torch held at a shallow angle against a circumferential joint, the shell of a muffler clamped to a ring, a gloved hand steadying the cup, and a small hard arc between the tungsten and the work. The cup is a large pink alumina type. Below the joint a length of threaded rod and a clamp hold the assembly in place.
The weld is going around a circle, which is the least forgiving geometry in the shop, because the position changes as you travel. And the electrode at the centre of the arc has a colour it should not have.
A golden bead is the finish line. It says the heat input landed in a band that produced a clean, evenly spaced, lightly coloured weld, and on stainless that is what a controlled arc looks like.
A blue tip is not a result. It is a condition of the tool, and it is readable before, during and after the weld. That is what makes it more valuable to the person running the torch: the bead tells you what you did, the tip tells you what the machine was doing while you did it.
Anyone can photograph a good bead. The tip is the part of the picture that was not staged.
On stainless, a light straw to gold tint along the bead and a narrow band either side is the ordinary appearance of a well-shielded weld that cooled without sitting in the heat too long. It is a sign that the gas was reaching the joint and that travel kept moving.
It is worth being precise about this because the same colours mean something different on titanium, where the palette is a stricter instrument and a saturated blue is a warning rather than a finish. On stainless, blue and purple are not automatically a defect, and chasing a perfectly gold bead on a part that does not need it is a way to spend time on cosmetics. Gold here is evidence of control, not a certificate of soundness.
Tungsten is only inert while the gas is around it. Take the shield away and the hot tip oxidises, and the oxide shows as colour: straw first, then blue, then a dull grey-white that no longer conducts the way it should.
The usual routes to that colour, in the order they are found in practice:
The colour itself is not the problem. What it indicates is that the tip is no longer the shape and cleanliness the arc was set up for.
An oxidised or contaminated tungsten loses its ability to hold a stable arc in one spot. The arc starts to wander, sometimes splits into two feet, and the puddle follows it. Penetration becomes uneven along a joint that looks acceptable from above. In the worst version, tungsten transfers into the weld and sits there as a hard, brittle inclusion that will not show on the surface and will not pass a bend test.
This is why a golden bead and a burnt tip can exist in the same photograph: the bead is the average of the pass, and the electrode condition is a local event. A joint can run mostly well and still carry the consequence of the two centimetres where the arc wandered.
A tungsten is treated as something you use until it stops working. It is closer to a cutting tool: its geometry decides what the arc does, and its geometry is destroyed by use.
Grinding a used electrode back to a point is normal. Grinding it back repeatedly along the same axis is not, because the tip develops a taper that no longer matches the angle the arc was set for, and the arc follows the taper. When a tip has gone blue and pitted, the correct move is to cut the contaminated length off and re-point, not to dress the discolouration away and hope.
The point itself matters more than most operators are taught. A long fine taper concentrates the arc and narrows the puddle; a blunt angle spreads it. Grinding across the axis rather than along it leaves scratches that the arc will follow. None of this is visible in a finished weld, and all of it is visible in the arc.
The pink cup in the frame is alumina, chosen because it shrugs off heat and spatter. It is also a ceramic that cracks, chips and spalls, and every one of those is a leak in a system whose whole job is to keep air away from hot metal.
A cup with a chip at the rim will still look fine on the bench and will still weld, badly, with a shielding problem that appears only on one side of the joint. The same is true of a gas lens whose screen has been blinded by spatter, which is the quietest failure in the whole setup because the flow gauge still reads correctly.
The joint in the photograph is circumferential, which means the position is changing continuously. At the top of a round shell the puddle wants to sag away from you. At the bottom it wants to run into the arc. The same current and the same travel speed produce different beads at three o'clock and at nine o'clock, and the discolouration band tells you which parts of the circle got more heat than the rest.
The clamp below the joint is doing two jobs at once: holding the ring square to the shell, and pulling heat out of the seam so the shell does not walk as it shrinks. A fixture that holds but does not conduct is half a fixture.
Three things on a finished joint are read from the outside, and they are worth knowing before you accept one:
None of these replace a bend test or a section. All of them are free, and all of them are visible in a photograph like this one, which is why the tip in the corner of the frame is worth as much attention as the bead in the middle of it.
FUPOWER has been TIG welding titanium, stainless steel and aluminium in Shanghai since 2009, with pressing, forming, mandrel bending, perforation, casting, CNC machining and silicone hose production in four facilities under one roof. If a joint has to hold, the setup is where it is decided.
🟡 A golden bead is a result, not a diagnosis — it records what the pass produced, and it can look correct while the setup has already drifted.
🔵 A burnt tip is the diagnosis — electrode colour reads the shielding condition before, during and after the weld.
🧰 Oxidation has a short list of causes — stick-out, cup size, gas flow in either direction, a blocked lens, a cracked cup, or post-flow cut short.
⚡ A compromised electrode moves the arc — wander, a split arc, uneven penetration, and occasionally tungsten transferred into the weld.
🔬 Tungsten inclusions are visible — small bright specks in the bead, hard and brittle, and they do not polish out.
📏 The tip is a cutting tool — taper and grind direction set the arc shape, and repeated re-pointing quietly changes both.
✂️ Cut back, do not dress — a blue, pitted tip loses its contaminated length rather than having the colour ground away.
🏺 The cup is part of the gas system — alumina resists heat and spatter, and every chip or crack is an invisible shielding leak.
🌬️ Too much flow is a failure mode — turbulence at the cup mouth draws air in, and the gauge still reads correctly.
🔄 A circle changes position continuously — top and bottom of a round shell want different beads from identical settings.
🧱 A clamp that holds but does not conduct — is half a fixture; the ring has to stay square and shed heat.
📷 Three things read from outside — tungsten specks, the width of the discolouration band, and ripple spacing that collapses where travel stopped.