CNC Treatment Experience: Heat Treatment and Machining

• 1
  Why the CNC Treatment Experience Starts Before the Toolpath
• 2
  What Heat Treatment Actually Does to the Part
• 3
  Where Distortion Comes From and How to Cut It
• 4
  Machining Before or After Treatment: A Decision Rule
• 5
  What Changes on the Machine After Treatment
• 6
  Finishing and Inspection After the Furnace
• 7
  Treatment Route vs Machining Sequence

A hardened 4140 shaft and the same shaft in the annealed state are not the same part. Hardness changes how the tool enters the material, how much heat leaves with the chip, and how much the part moves after you unclamp it. The CNC treatment experience can be summed up in one line: decide the heat treat route before you decide the toolpath.

Every heat treatment cycle does two things at once. It changes the microstructure, which is what the drawing asks for. It also releases or adds internal stress, which is what the drawing never mentions. On a 200 mm long 4140 shaft, that second effect can move the part 0.05–0.15 mm along the axis after quenching and tempering.

So the sequence question is not a shop preference. It is a tolerance question. If your tightest callout is ±0.05 mm and the part will be through-hardened, you must leave grinding stock and machine after treatment. If the callout is ±0.2 mm and the part is only stress relieved, rough machine, treat, then finish in one light pass.

The trade is always the same. Treat first and you cut hard material with slow speeds and short tool life. Treat last and you chase distortion with extra stock and extra operations. Neither is free. The right answer comes from the tolerance, the hardness, and the part's length-to-width ratio.

Annealing heats the steel above its critical temperature, holds it, then cools it slowly in the furnace. Hardness drops to roughly 150–200 HB. Internal stress from casting, forging, or prior roughing is largely relieved. This is why annealing is often the first operation on a forging, long before any finish cut.

Normalizing is the same idea with air cooling. It refines grain structure and gives a more uniform hardness than annealing, still machinable at 180–220 HB. For a part that will be machined to final size and not hardened afterward, normalizing is usually enough to keep dimensions stable.

Quenching is the opposite direction. The part is heated above critical temperature and cooled fast in oil, water, or polymer. Martensite forms, hardness jumps to 50–62 HRC depending on carbon content, and volume changes. That volume change is not uniform, because thick sections cool slower than thin ones.

Tempering follows immediately. Reheating to 150–650 °C trades some hardness for toughness and relieves the stress that quenching locked in. Skip tempering and the part becomes a crack waiting for a load. The tempering temperature, not the quench, is what really sets the final working hardness.

Distortion has three sources. The first is thermal gradient: thin webs cool faster than thick bosses, so they shrink first and pull the part out of shape. The second is transformation: martensite takes up more volume than the austenite it came from, and the expansion happens unevenly. The third is residual stress from the machining that came before treatment.

Rough machining leaves stress in the surface layer. When the furnace releases that stress, the part bends. A 300 mm long aluminum or steel plate roughed with a heavy 3 mm depth of cut can move 0.1 mm or more during stress relief. Rough with lighter passes, or specify a stress-relief cycle between roughing and finishing.

Fixturing during quench matters too. Parts laid flat on a tray warp differently than parts hung vertically. For long shafts, vertical hanging with a controlled quench is the standard answer. For thin plates, press quenching between plates holds flatness during the critical first seconds.

Design helps more than any process tweak. Symmetric sections, uniform wall thickness, and generous fillets all reduce the gradient. If a rib must be on one side, adding a balancing rib on the other side costs less than the straightening operation it saves.

Use three numbers to decide. Final hardness, tolerance band, and part length. Hardness above 45 HRC rules out most carbide turning at reasonable tool life, so grinding or hard milling after treatment becomes the finish operation. Hardness below 35 HRC usually allows a single finishing cut after treatment with coated carbide.

Tolerance band decides stock. Leave 0.3–0.5 mm per surface for grinding, 0.15–0.25 mm for a finishing mill or turn pass after treatment. On a part held to ±0.005 mm, the finishing cut must come after the last thermal cycle, no exceptions. No sequence can hold that band across a quench.

Part length decides whether you need an intermediate stress relief. Under 100 mm, distortion is usually small enough to handle with stock. Over 300 mm, or any part with a length-to-diameter ratio above 8, plan a stress-relief cycle between roughing and finishing.

For case-hardened parts, the rule shifts. Carburizing and nitriding build a hard skin 0.2–1.0 mm deep over a soft core. Only the skin changes, so distortion is smaller than through-hardening. Machine to near-final size, leave 0.02–0.05 mm for a light grind, and keep sharp internal corners out of the design.

Hard material pushes you toward slower surface speed and smaller chip loads. On 50 HRC tool steel, coated carbide turning runs around 60–90 m/min surface speed with 0.1–0.15 mm/rev feed. On the same steel at 30 HRC, you can run 150–200 m/min. That difference is why treat-first routes cost more machine time.

Tool choice follows hardness. Below 40 HRC, standard coated carbide works. Between 40 and 55 HRC, use cubic boron nitride or ceramic inserts for turning, and solid carbide with a hard coating for milling. Above 55 HRC, grinding is usually cheaper than any cutting tool.

Cutting forces rise with hardness, so clamping and workholding matter more. A part that was stable at 25 HRC can chatter at 50 HRC with the same fixture. Reduce radial engagement, shorten tool overhang, and check that the vise or chuck has enough grip without deforming the finished surface.

Heat is the other change. Hard materials carry heat into the tool instead of the chip. Air blast or high-pressure coolant keeps the edge alive. On finishing passes where surface finish matters, keep depth of cut small and constant so the tool does not rub.

Scale and decarburization sit on the surface after quenching. Scale must come off before any finish cut or you will chip the tool on a hard, brittle layer. Bead blasting or a pickling step removes it and gives a uniform starting surface.

Decarburization is the quiet problem. It leaves a soft layer 0.05–0.3 mm deep that looks like good steel. If you grind through it, you expose the hard core. If you do not remove it, the surface hardness reading comes back low. Specify the decarburization limit on the drawing, or leave enough stock to remove it entirely.

Grinding introduces its own stress and heat. A burn mark means the surface was overheated and the microstructure changed locally. Use light passes, a dressed wheel, and enough coolant. For a part held to ±0.005 mm, measure after the part reaches room temperature, not straight off the machine.

Final inspection should happen after the part has stabilized. Check hardness at several points, not one, because a quench bath gives different results at the center and the edge. Record the location of each reading so a low number can be traced to a section thickness rather than blamed on the material.