Analysis and Use of the CNC Tower Thread Cutting Method

A tower thread is still a standard 60° V, trapezoidal or Whitworth profile turned on a lathe. What makes it awkward is the part it sits on: a tall shaft, column or tower-shaped body where the thread runs some distance from the chuck and often far from the tailstock support. The cutting edge meets the flank at an angle, so the radial force pushes the tool away from the work. That push shows up directly in the thread's pitch diameter and flank angle.

Depth of cut per pass decides how much of that force appears at once. A 60° external thread with 2.0 mm pitch removes roughly 2.6 mm² of material per side over the whole profile, and no lathe tool survives taking that in one pass. The question is how you split it up. Direct infeed splits it straight down the middle, so both flanks cut at the same time. Oblique infeed feeds in at an angle, so one flank does most of the work.

Tool nose radius and thread height set the starting point. For a 60° form, thread height is 0.6134 × pitch, and the tool nose radius should sit between 0.1 and 0.15 × pitch. Get those two numbers right and the pass schedule becomes predictable. Get them wrong and you will chase pitch diameter all afternoon with no repeatable result.

G32 is a single-pass threading block. You program the start point, the end point and the feed rate equal to the thread pitch, and the control does exactly that one pass. Nothing is calculated for you. Every depth increment is a number the programmer writes into the program, which is both the strength and the risk of the method.

In absolute programming X and Z give the target point; U and W give an incremental move from the current position. A typical first pass might be 0.3 mm depth, then 0.25 mm, then 0.18 mm, then 0.12 mm, then a spring pass at the same depth. Those numbers come from experience with the material, not from the control.

Cutting happens on both flanks at once. Chip load is shared, so the edge sees a balanced radial load, but the contact area is double what a single-flank method produces. On a 3 mm pitch in 4140 at 45 HRC, that double contact is where chatter starts.

The method rewards short, stiff setups. When the thread is within 3 × diameter of the chuck, G32 gives you the tightest control over final pitch diameter. Beyond that the same program will drift.

G76 hands the pass schedule to the control. You supply the thread height, the first depth of cut, the minimum depth of cut, the finishing allowance, the number of finishing passes, the tool tip angle and the chamfer width, and the control works out every intermediate pass along an oblique path.

Common addresses: M for the number of finishing passes, A for the included tip angle, q for the first depth of cut in microns, ∆dmin for the minimum depth of cut, and D for the finishing allowance. Some controls use a separate address for chamfer width at the start of the thread.

One flank carries most of the load. That reduces the contact area per pass, which lowers cutting force and spreads wear across a smaller zone. The trade is a slight imbalance: the loaded flank wears faster than the idle flank, and a thread cut this way can show a small flank angle error if the tool is not reset often enough.

Where G76 earns its place is on long thread runs, coarse pitches and materials that work harden. The control will not let a pass get too deep, so a programming mistake turns into a slower cycle rather than a broken insert.

For a 4 mm pitch in 316L, a typical G76 cycle might start at 0.35 mm, taper down to a 0.05 mm minimum, then take two finishing passes at 0.04 mm. The cycle time is longer than G32, but the insert lasts.

Tool on center, every time. A threading insert set 0.1 mm above center will rub and cut an oversized pitch diameter; set 0.1 mm below and the flank angle opens up. A height gauge and a test cut settle it faster than any calculation.

Rigidity decides the maximum usable depth of cut. A 4,000 mm maximum processing size lathe can turn a long shaft, but the overhang still has to be managed with a tailstock or steady rest, or the thread will come out with a visible lead error.

Spindle speed follows the material and the thread height. A common starting point is a surface speed of 100–150 m/min for aluminium, 60–90 m/min for carbon steel, and 30–50 m/min for stainless. Coarse threads need the lower end of each band.

Cooling matters more than most people expect on deep threads. Flood coolant aimed at the leading flank keeps the insert from thermal cracking, especially on a G32 cycle where each pass re-enters the same groove.

Check pitch diameter with a thread micrometer or a three-wire set, not with a nut. A nut tells you the thread fits one mating part. A micrometer tells you whether the next thousand will.