Three ways to put a thread in metal
A thread callout on a drawing says what the thread must be, not how it should be made, and there are three production routes that get there. The thread can be cut with a tap, generated with a rotating single-point tool on a milling machine, or formed by squeezing material between rollers rather than removing any.
The three routes do not produce equivalent threads. They differ in the strength left in the component, in the tolerance that can be held, in the tooling cost and in the material they will even work in. Choosing on price alone, or letting the machine shop choose after the fact, is how a part ends up with a thread that strips in service or a tapping operation that breaks tools all afternoon.
Tapping: fastest per hole, least forgiving
Tapping is the cheapest way to put a thread in a hole that already exists. A tap is a threaded tool that is driven into a pilot hole and cuts or forms the thread as it goes, and on a CNC machine it is a fast, well-understood cycle. For a part with a dozen small holes in aluminium or mild steel it is almost always the right answer.
Its weakness is that it has no way to correct its own path. The tap follows the pilot hole, so hole position, hole diameter and hole depth all have to be right first, and a tap that meets a hard inclusion or a hole that has been drilled undersize will break rather than cut. A broken tap in a nearly finished part is one of the more expensive failures in a machine shop, because removing it risks the part the work was spent on.
Tapping also concentrates stress. The process removes material and leaves a sharp-rooted profile, which is why a tapped hole in a soft alloy is the thread most likely to be specified with an insert instead.
Thread milling: one tool for every diameter
Thread milling uses a small single-point cutter on a helical path, generating the thread rather than forming it in one pass. The obvious advantage is flexibility: a single thread mill can produce any diameter the machine can reach, because the diameter is set by the programmed path and not by the tool.
The second advantage matters more in practice. Because the tool is small and the cut is light, a thread mill handles large-diameter threads, hard materials and blind holes with poor chip clearance far better than a tap. If a thread mill breaks it leaves a smaller obstruction in a larger hole, and the operation can often be recovered. It is slower per thread, and it consumes machine time, which is what you are paying for.
Thread rolling: formed, not cut
Thread rolling displaces material rather than removing it. The blank is squeezed between dies or rollers, and the metal flows into the thread form. Nothing is cut away, so the grain of the material is not severed - it is re-oriented along the thread profile, which is why a rolled thread is generally stronger in fatigue than a cut thread of the same nominal size.
The trade-off is that rolling is a forming operation and needs the right material. Ductile metals such as low-carbon steel and aluminium roll well. Hardened or brittle material, and castings with porosity near the surface, do not. Rolling also needs a blank diameter that is right for the thread - usually between the pitch diameter and the major diameter - so it has to be planned into the process rather than substituted at the last step.
What the method does to the thread's strength
Three effects separate the routes. First, whether material is removed or displaced: a cut thread severs the material's grain, a rolled thread refines it. Second, the root radius: a formed thread typically has a more generous root and a smaller notch effect, which is where fatigue cracks start. Third, surface finish inside the thread, which is smoother on a rolled thread and matters for galling and for sealing.
The practical consequence is that a thread carrying a repeated load should not be specified without asking how it will be produced. For a static joint the difference is usually not worth the cost of changing the process. For a thread that sees vibration or cycles, the rolled thread is often the difference between a joint that survives and one that does not.
Matching the method to the material and the hole
Small holes in soft, ductile material are the natural home of tapping. Large threads, hardened material and awkward blind holes favour thread milling, and long production runs of a thread that will be repeatedly loaded favour rolling, where the volume can absorb the setup.
Hole geometry constrains the choice too. Thread milling reaches a blind hole with limited chip clearance better than a tap, and it can thread close to a shoulder where a tap cannot run out. Rolling needs a blank with somewhere for the displaced material to go, so it is done before the thread area is finished, not after.
What to put on the drawing
Specify the thread standard and size, the class of fit, the depth of full thread and the position tolerance of the hole - and where the thread will be loaded, say so. Marking the thread with a note that it must be rolled, or that it must not be tapped, is unusual but entirely legitimate when the joint depends on it.
The most common omission is the class of fit. A thread note with no class is made to whatever the shop considers normal, and two parts made to different assumptions may not assemble or may be loose. Naming the standard and the class costs a line and removes the ambiguity.
References
Thread callouts are written against a published standard: either the Unified Thread Standard or the ISO metric screw thread. The dimensioning discipline that governs the hole position and any feature control frame is ASME Y14.5.
Frequently asked
Is a rolled thread always stronger than a cut thread?
For fatigue loading, usually yes - rolling displaces the material instead of severing its grain and leaves a better-formed root radius. For a static joint the difference is generally not decisive. Rolling also needs a ductile material and the right blank diameter, so it is a process you plan for rather than a substitution you make at the end.
When should a thread be milled rather than tapped?
When the thread is large, when the material is hard, when the hole is blind with poor chip clearance, or when several thread sizes must come off one tool. Thread milling is slower per thread but it is far more tolerant of the conditions that break taps, and a broken thread mill is a smaller problem than a broken tap.
Do I need to specify how the thread is produced?
Only where the joint depends on it. A note that a thread must be rolled, or must not be tapped, is legitimate and worth writing when the thread carries a repeated load. For ordinary static joints, specifying the standard, size and class of fit is normally enough.