Horizontal broaching machine: how it cuts metal

• 1
  Key takeaways
• 2
  What the horizontal broaching machine actually does
• 3
  Rise per tooth and the limits of one pull
• 4
  Why the horizontal layout suits long parts
• 5
  Which metals broach well and which fight back
• 6
  When broaching beats milling, and when it does not
• 7
  Broaching compared with milling for internal profiles

A horizontal broaching machine is a machine tool used for cutting metal by dragging a long, toothed bar through a fixed workpiece. The bar is the broach. Each tooth stands slightly taller than the one before it, so the cut is divided into many small steps that happen in one continuous stroke. Nothing rotates. The workpiece stays locked while the broach travels.

On a horizontal machine the pull axis is parallel to the floor. The broach is gripped at the front by a pull head, passed through a pilot hole or a pre-machined slot in the part, then drawn through at 3–30 m/min depending on material. Because the bar is supported along its length, a horizontal layout handles broaches that would sag under their own weight in a vertical machine.

The name describes the machine, not the finish. The finish comes from the last few teeth, which have a smaller rise and a wider land. Those teeth shave rather than hog, and they set the final surface. Typical results sit around Ra 0.8–1.6 μm on steel and better on aluminium, with no secondary operation needed on the broached face.

Every tooth does a fixed job. Roughing teeth take the bulk of the stock, semi-finishing teeth correct the profile, and finishing teeth size the surface. If a tooth chips, the damage shows up on every part that follows until the broach is resharpened. That is why broaching shops track stroke count rather than calendar time.

The controlling number on any broaching job is rise per tooth, sometimes called chip load. It is the height difference between one tooth and the next. For carbon and alloy steel, 0.04–0.10 mm per tooth is a normal band. Push it higher and the tooth edge starts to break down. Drop it lower and you need more teeth, which makes a longer and more expensive broach.

Total stock removal is the sum of every rise in the bar. If a keyway needs 6 mm of depth and the rise is 0.06 mm, the broach needs roughly 100 cutting teeth plus finishing teeth. That length has to fit the machine stroke and the part length together. This is the hard limit people hit first: the broach grows, and the machine may not have the travel.

Cutting force scales with the number of teeth in contact at one time, not the total tooth count. A narrow keyway might engage 8 teeth at once, while a wide spline might engage 20. Force per tooth on steel runs in the range of 100–300 N per mm of edge width. Add those up and you get the pull capacity the machine must supply.

Heat is the quiet problem. All that work happens in a few seconds, so the broach and the part both grow. Shops flood the cut with cutting oil, not soluble coolant, because the oil carries heat away and lubricates the tooth flank at the same time. Starved lubrication shows up as a torn surface, not as a broken tool.

• 1
  Rise too highTooth chipping, rising spindle load, torn finish on the last teeth.
• 2
  Rise too lowBroach gets long, cost climbs, and the machine may run out of stroke.
• 3
  Too few finish teethProfile drifts and surface finish becomes inconsistent part to part.

Gravity decides a lot in broaching. A vertical machine holds the broach above the work, and a long bar hangs. A horizontal machine lays the bar along its axis, so the broach is supported by the pull head and the guide. That is why long keyways, splines, and internal profiles usually run horizontally.

Clamping follows the same logic. On a horizontal machine the fixture holds the part against the pull force along one axis. A shaft with a 500 mm bore can be clamped on its outside diameter and pulled through in a straight line. On a vertical machine the same part needs a taller fixture, more headroom, and more care to keep the bore aligned with the ram.

Part size on the machines we run reaches 4,000 mm in the longest axis. That does not mean every broaching job should be that large. Long parts are also heavy, and handling weight is a real cost. When a part crosses roughly 40 kg, we look at whether the broach can be pulled through a stationary part or whether the part should move instead.

Horizontal broaching also fits parts that cannot be rotated. A welded housing, a cast gearbox, or a frame with an irregular outline can sit flat on the table. There is no need to spin it for indexing. That single fact removes a whole class of fixture problems.

Broaching likes materials that shear cleanly. Free-machining steels such as 1018 and 1045 cut with a continuous chip and hold a good finish. Aluminium alloys including 6061, 7075, and 2024 broach fast and cool, though they need sharp teeth because the material tends to smear. Brass and bronze sit in the same easy group.

Stainless is a middle case. Grades 303 and 416 broach well because of the sulfur addition. Grades 304, 316, and 316L work-harden under the tooth, so the rise per tooth has to be high enough to stay under the hardened layer. If the broach rubs instead of cutting, the surface hardens and the next tooth has a harder job than the one before.

Titanium and nickel alloys are the hard end. Ti-6Al-4V and Inconel generate high cutting forces and hold heat at the edge. We slow the stroke, reduce rise per tooth, and expect shorter broach life. On these alloys we also check whether broaching is the right process at all, since a 5-axis milling operation may remove the same stock with less tooling risk.

Hardened parts need a decision early. Above roughly 40 HRC, broaching becomes a grinding or EDM job unless the profile was cut before heat treatment. Broaching a hardened spline usually means the broach wears on the flank, and the tooth geometry changes as it wears.

Broaching wins on repeat volume. A broach is a dedicated tool, so the setup cost is real. Once it exists, cycle time for an internal spline can be 10–30 seconds, which no end mill matches. The break-even point usually sits in the low thousands of parts for a simple profile, and lower if the alternative needs several tools and a long cycle.

Milling wins on flexibility. A 5-axis machining center can cut an internal profile, change to a different part number, and change back without a new tool. If the design is still moving, or if the profile changes between revisions, broaching locks you in. A design change after the broach is made means a new broach.

Accuracy is not the deciding factor as often as people assume. A good broach holds ±0.005 mm on the profile and repeats it part after part, because the tool is the gauge. Milling can reach the same tolerance, but it depends on the operator, the tool wear, and the setup. Broaching trades flexibility for repeatability.

There is a middle path. For prototypes and low-volume runs we mill the internal profile on a 3-axis or 4-axis machine with a long-reach cutter, then measure the result. If the part later moves into production at volume, the profile data is already proven and a broach can be quoted against a known geometry.