Why Should Machine Tools Relieve Stress?

• 1
  Symptom, likely cause, and what to do
• 2
  Where residual stress comes from before the tool ever touches the part
• 3
  Why should machine tools relieve stress before high-precision cutting
• 4
  How stress relief is actually done
• 5
  Thermal drift in the machine is a different problem with the same symptom
• 6
  Clamping and fixturing can imitate stress release
• 7
  Step by step: controlling stress in a precision part

Castings, forgings, and weldments cool unevenly. The surface solidifies first while the core is still hot and plastic. When the core finally contracts, the skin holds it back, and the part locks in a stress field with no external load applied. Nothing is wrong with the material. The stress is a memory of how it cooled.

Machining then removes material unevenly. Cut 3 mm off one face of a plate and the internal balance is broken. The remaining material relaxes toward a new shape, and the part moves after the vise opens. This is why a part can measure perfect on the machine and fail on the CMM the next morning.

Heat treatment adds its own stress. Quenching a 4140 block from austenitizing temperature creates a steep thermal gradient and a phase change that both push stress into the part. Tempering reduces it but does not remove it evenly across a thick section.

The engineering consequence is simple. Any process that changes shape, temperature, or material distribution leaves stress behind. High-precision work has to plan for that release, not hope it stays quiet.

• 1
  Cooling gradientsSurface and core contract at different times.
• 2
  Uneven stock removalOne-sided cuts unbalance the stress field.
• 3
  Heat treatmentQuench and temper leave a gradient through the section.
• 4
  WeldingLocal heating and shrinkage pull the joint and the base metal.

Tolerance decides how much movement you can absorb. A bracket at ±0.1 mm can be machined from raw stock and shipped. A part at ±0.005 mm cannot. At that level, a 0.02 mm relaxation after unclamping is a scrap event, not a rework.

The rule in our shop is straightforward. If the drawing tolerance is tighter than 0.05 mm and the blank is a casting, forging, or weldment, the part gets a roughing pass, a stress relief step, and a finishing pass. Skipping the middle step usually costs more than it saves.

Thin walls and long slender parts are the worst cases. A 2 mm wall in a 6061 housing can move 0.03–0.08 mm after the clamps come off if the roughing cut was aggressive. Reducing radial engagement and taking a light finishing pass after relaxation keeps the wall where the model says it should be.

Stress relief is not only about the part. The machine tool itself has thermal stress. A spindle that has run for 20 minutes is not the same length as a cold spindle. That is a separate mechanism, covered below, but it shows up as the same symptom.

• 1
  Under 0.05 mmPlan rough, relieve, finish on cast or welded blanks.
• 2
  Wide toleranceOne-pass machining from stock is often enough.
• 3
  Thin wallsRelaxation after roughing is larger than the tolerance.

Thermal stress relief is the most common method. The blank is heated slowly to a temperature below the critical range, held long enough for creep to redistribute stress, and cooled slowly in the furnace. For carbon and alloy steels the hold is often 1–2 hours per 25 mm of section thickness, with a furnace cool. The exact cycle depends on the alloy, so we follow the material supplier's recommendation.

Vibratory stress relief uses a controlled vibration at a sub-resonant frequency applied to the part for 10–30 minutes. It is faster and avoids scale and distortion, which matters on large weldments. It is less thorough than a furnace cycle, so we reserve it for parts that cannot be heated or where the tolerance is moderate.

Natural aging is the slow option. Leave the rough-machined part on the rack for weeks and let time do the work. It is still used for large castings and for granite and iron machine bases. For a job with a 3–5 day ship window, natural aging is not a plan; it is a reason the schedule slips.

A practical shortcut on many parts is to rough machine, remove the clamps, let the part sit for a few hours, measure, and then finish. The relief is partial but the movement is captured before the final pass. This works well on aluminum, less well on large steel castings.

• 1
  ThermalFurnace cycle, slow heat and cool, best for steel castings.
• 2
  Vibratory10–30 minutes, good for large weldments.
• 3
  Natural agingWeeks of rest; only for long schedules.
• 4
  Rough-and-restPartial relief for aluminum parts on short lead times.

A cold machine and a warm machine cut differently. The spindle grows 10–30 μm as it reaches steady state, and ballscrews grow along their length. On a 4,000 mm travel machine the screw can move more than a tight tolerance allows. That drift looks like stress release in the part, but the part is fine.

The fix is a warm-up cycle. We run the spindle and the axes through a programmed routine for 30–60 minutes before the first finishing cut. After warm-up, the operator re-zeros the work offset. On long production runs we check the first part, then re-check every 2 hours and log the drift.

Coolant temperature matters as much as spindle speed. If the chiller runs at 20 °C and the shop air is 30 °C, the machine casting and the part sit at different temperatures and the measurements disagree. We measure critical features at 20 °C ± 1 °C and give the part time to stabilize.

If you see a size trend across a run rather than a single jump, suspect thermal growth before you suspect the material. A trend that reverses after a lunch break is almost always heat, not stress.

• 1
  Warm-up30–60 minutes before finishing, then re-zero.
• 2
  Trend not jumpSize drift across a run points to thermal growth.
• 3
  Measure at 20 °CLet the part stabilize before final inspection.

A part that springs when the vise opens is not always releasing internal stress. Sometimes the clamp bent it in the first place. Over-tightening a thin wall pushes material into the cut zone, the tool removes it, and the wall springs back to a shape that is now undersize.

We use soft jaws machined to the part profile, torque the clamps to a set value, and support thin sections with adjustable jacks where possible. For a 2 mm wall, a clamp force of a few hundred newtons is enough to move the part more than the tolerance. Reducing force changes the result more than changing the toolpath.

For thin plates, vacuum chucks and adhesive fixturing hold the part flat without side load. They do not work for every geometry, but when a part has a flat datum and a thin section, they remove clamping from the list of variables.

One useful test: measure the part while it is still clamped, then after unclamping, then after 24 hours. Three numbers tell you whether the movement is elastic, stress-driven, or still developing.

• 1
  Soft jawsMachined to the profile, contact over a wide area.
• 2
  Torque controlSet clamp force instead of tightening by feel.
• 3
  Vacuum and adhesiveHold thin plates without side load.