A boss is a local raised feature that carries a bolt, a sensor, a port or a mounting point. It can be cast into the component or welded onto it, and the choice between those two routes is not a matter of preference. Each carries a different set of consequences in strength, inspection, distortion and cost, and the right answer depends on which of those consequences the component can absorb.
Casting a boss into the part means the feature is formed in the same operation as everything else. There is no joint, no filler metal and no heat affected zone. The load path from the boss into the wall is continuous, and the geometry can be shaped to distribute stress rather than concentrate it. That is a genuine structural advantage, and it is the reason cast bosses are preferred on parts that see vibration or thermal cycling.
Welding a boss on means the feature is added after the main body exists. That is flexible and often faster to arrange, particularly at low volume or when the boss position is still being developed. It also introduces a joint, and a joint is where the interesting things happen.
A welded boss has a fusion boundary, a heat affected zone and a local region of residual stress. Each of these is a potential initiation point for a crack under cyclic loading, and each has to be inspected. The inspection is not free, and on a part with several bosses it multiplies.
The heat affected zone also softens or alters the material locally, which matters on components that are heat treated for strength. A cast component can be heat treated as a whole after casting. A welded assembly has to be heat treated after welding, or accepted with a local property variation that the design has to tolerate.
Distortion is the other consequence. Welding a boss onto a thin wall pulls the wall as the weld cools, and on a large flat face that pull can be enough to take the face out of flatness. Casting the boss avoids the pull entirely because there is no local thermal cycle after the part is formed.
The economics of casting a boss improve with the number of bosses and the number of parts. A single boss on a single prototype is a welding job. Eight bosses on a production component is a pattern feature, and the marginal cost of each additional boss in the pattern is small once the pattern exists.
There is also a consistency argument. Cast bosses are in the same place on every part, to the same tolerance, with the same surrounding geometry. Welded bosses depend on fixture, operator and procedure, and while those can be controlled, the control is an ongoing activity rather than a property of the tooling.
On a component that has to assemble repeatedly, that consistency has value beyond the unit cost. A boss that is consistently placed is a boss that the assembly fixture can rely on. A boss that varies is a boss that generates rework at the end of the line, where rework is most expensive.
Welding is not the wrong choice. It is the right choice in specific circumstances: at very low volume, where a pattern cannot be justified; where the boss position is still being developed and the design will change; where the boss has to be a different material from the body; and where the component is large and simple enough that fabrication is cheaper overall than tooling.
There is also the repair case. A cast boss that has been damaged can sometimes be replaced by a welded one, and a boss that was missed in the design can be added to an existing casting. These are legitimate uses, and they are different from choosing the process at the design stage.
The honest comparison is between a casting with the boss included and a fabrication with the boss welded, priced across the whole production run rather than per unit. The casting has a high fixed cost and a low marginal cost. The fabrication has a low fixed cost and a marginal cost that includes every joint, every inspection and every correction.
Where a boss is cast, the drawing needs to specify it as a cast feature, with draft, with a fillet at the base and with the section thickness considered against the wall it sits on. A cast boss that is drawn as a plain cylinder with a vertical wall is a boss that will cause a draft problem and a hot spot at the same time.
Where a boss is welded, the drawing needs to specify the joint, the inspection and the post weld treatment. These are not details to be left to the fabricator, because they determine whether the boss performs as intended.
Components that mix both routes, such as some of the assemblies in the custom fabricated parts range, need both sets of requirements on the same drawing. That is more work at the specification stage and considerably less work at the inspection stage.
Casting a boss instead of welding one is therefore a decision about where the complexity goes. Casting moves it into the pattern, where it is paid once. Welding leaves it in the part, where it is paid every time.
A cast boss concentrates mass, and the wall beneath it has to feed that mass or accept the shrinkage. It is a section question before it is a boss question. TIG aluminium welding avoids the concentration but puts a joint in the same place.
A boss exists to locate something, so its position is a fit requirement rather than a casting convenience. Casting holds that position through the tooling, which matters more as the number of bosses rises. aluminium fuel tank range and aluminium water tank range assemblies depend on exactly that consistency.
A damaged cast boss can sometimes be replaced by a welded one, which is a legitimate repair and a poor design default. TIG stainless steel welding and titanium fabricated components both offer that route when a casting has to be recovered rather than scrapped.
