Heat input is the amount of arc energy delivered into a length of joint. Not the total energy in the part, not the peak temperature at any instant, but the energy per unit of travel. It is the single figure that best predicts both the colour you see on stainless and the way a part walks out of shape after it cools.
The reason it earns that status is that it survives changes in equipment. Two welders on different machines, with different torch styles, can produce the same heat input if the balance of current, arc voltage and travel speed lands in the same place. That is what makes it a transferable description of a weld rather than a description of a machine setting.
It is also why a procedure can specify a heat input ceiling and leave the current range fairly wide. The ceiling describes the outcome the qualification protected. The current range describes one way of staying under it.
Three things decide heat input: how much current the arc carries, how much voltage it drops across the arc gap, and how fast the torch moves. Current and voltage multiply into power; travel speed divides it back down into energy per length. Everything else, torch angle, gas composition, cup size, is second order by comparison.
Arc voltage is the one welders influence without noticing. Voltage rises with arc length, so a welder running a long arc is running higher power at the same current setting, and he is spreading it over a wider area at the same time. The two effects partly cancel in appearance and do not cancel in distortion.
Travel speed is the one that gets abused. It is also the variable that most directly trades against current, which is why the pair should always be discussed together rather than one at a time.
In practice nobody adjusts heat input directly. They adjust current, or they adjust how fast they move, and the second is far more responsive. A welder who feels the pool going cold will slow down before he touches the dial, and slowing down raises heat input even though the machine has not changed.
That habit is usually right for the puddle and wrong for the part. Slowing down to keep fusion gives a wider heat-affected zone and more total energy in the section, which is exactly the condition that produces a long, slow bow in a thin-wall tube. The puddle looks better. The part comes out worse.
The disciplined alternative is to fix travel speed by rhythm and let current carry the fusion. It feels less natural at first because the welder is no longer modulating with his hand, but it makes the result repeatable, and repeatability is the whole point of a documented procedure. You can see how consistent rhythm shows up in finished tube work on our stainless steel exhaust system pages.
Colour on stainless is a surface oxide, and its progression from straw through bronze to blue and grey tracks how hot the surface got and how long it stayed there in air. It is genuinely informative and it is not a heat input reading. A fast pass at high current can leave a narrow, pale band beside a heat-affected zone that is wider than it looks.
The confusion arises because colour is the only feedback most shops have without instrumentation. Reading it well means reading it as a gradient along the weld and across it, not as a shade at one point. A tight straw band that fades quickly into parent metal suggests a concentrated arc and a quick exit. A broad blue field with soft edges suggests the part soaked.
Titanium raises the stakes, because the oxide colours there appear at lower energy and progress to a dull grey that indicates contamination rather than simple oxidation. On titanium, colour is closer to a verdict than an indicator, which is why the shielding discussion in our titanium exhaust system notes keeps coming back to gas rather than to current.
A weld does not bend a part because it is hot. It bends the part because the hot region tries to expand against colder metal around it, yields while it is soft, and then contracts on cooling with nothing to push back against. The amount of movement tracks the size of the hot region and the steepness of the gradient around it.
Heat input controls both. Higher input widens the region that goes soft and flattens the gradient, so contraction is spread over more material but acts on a bigger volume. Lower input concentrates the effect, which sounds better and is not automatically so, because a narrow, steep gradient produces its own angular pull.
On thin wall the same heat input that behaves on a heavier section will burn through, because there is less material to conduct energy away from the pool and less thickness to absorb the keyhole. The fix is not simply to reduce current. Reducing current alone forces a slower travel speed, and the slower travel speed puts the energy back in.
The workable combination on thin wall is a small, focused arc, a fast travel and a current high enough to fuse without dwelling. It is a narrower operating window than most welders are used to on thicker material, and it is unforgiving of hesitation. This is the region where a good fit matters most, because a gap that must be filled by dwell will destroy the balance.
A tube is a closed shape, so heat input is never applied to a thermally neutral body. By the time the welder comes back around to the start, that region has cooled and is now a heat sink, while the region just welded is still hot. The same settings produce a different result at each clock position.
Experienced welders compensate by changing travel speed slightly as they progress, or by sequencing the joint so the closing section lands where the part can absorb it. Neither of those is a deviation from the procedure if the heat input ceiling is respected, and both are more effective than turning the machine down halfway through.
Heat input is not something a buyer can measure on a delivered part, but the consequences are visible and the process behind them can be asked about directly. A supplier who understands the variable will answer these without reaching for a brochure.
Tigindustry has been welding exhaust and intake parts in Shanghai since 2009, across four factories and a plant of about 53,819 sq ft, and heat input discipline is what keeps repeated parts identical. You can see the results on our stainless steel exhaust mufflers and exhaust pipe pages, or start from shop by material and TIG stainless steel welding.
