1. Quick lookup by process or material
Enter a process name (quenching and tempering / normalizing / carburizing / nitriding / induction hardening / solution + aging) or a material grade (45 / 40Cr / 20CrMnTi / GCr15 / 42CrMo / 6061 / 304) to get the heating temperature, cooling method, achievable hardness and inspection requirements.
First separate three cases: ① Through (bulk) heat treatment (annealing / normalizing / quenching / tempering / Q&T) changes the whole section; ② Surface heat treatment (induction / flame / laser hardening) hardens only the surface and leaves the core untouched; ③ Thermochemical treatment (carburizing / carbonitriding / nitriding / QPQ) changes the surface chemistry before hardening. When a drawing simply says "HRC50", the service life of parts made by these three routes differs completely — always confirm which one is meant.
2. The four basics: annealing / normalizing / quenching / tempering
| Process | Heating temperature | Holding and cooling | Resulting structure | Hardness change (45 steel as example) | Purpose |
|---|---|---|---|---|---|
| Full annealing | Ac3 + 30–50 °C (45 steel 820–840 °C) | Furnace cool after holding at ≤50 °C/h to 500 °C, then withdraw | Ferrite + pearlite | Softest, HB ≤197 | Removes casting / forging structural defects, refines grains, improves machinability |
| Spheroidizing anneal | Ac1 + 20–30 °C (GCr15 780–800 °C) | Slow cool at 20–30 °C/h to 650 °C after holding, then withdraw | Spheroidal pearlite | HB 179–207 | Softens high-carbon and bearing steels; improves machinability and quench distortion |
| Stress-relief annealing | 550–650 °C (below Ac1) | Hold 2–4 h, furnace cool | Structure unchanged | Essentially unchanged (1–3 HRC lower) | Removes welding, machining and cold-work residual stress; prevents distortion and cracking |
| Normalizing | Ac3 + 30–50 °C or Accm + 30–50 °C (45 steel 850–870 °C) | Air cool (forced air / mist for large parts) | Fine pearlite + ferrite | 10–20 HB higher than annealing HB 170–210 | Refines grains, homogenizes structure, improves machinability and prepares for quenching |
| Quenching | Hypoeutectoid: Ac3 + 30–50 (45 steel 840 °C) Eutectoid / hypereutectoid: Ac1 + 30–50 (GCr15 830–850 °C) | Water / oil / salt bath / polymer quenchant; cooling rate must exceed the critical rate | Martensite + retained austenite | 45 steel HRC 55–58 GCr15 HRC 62–65 | Gives high hardness and wear resistance; prepares for tempering |
| Low-temperature tempering | 150–250 °C | Hold 1–3 h, air cool | Tempered martensite | Maintains HRC 58–64 (1–2 HRC drop) | Relieves quench stress and brittleness: tools and dies / carburized parts / bearings |
| Medium-temperature tempering | 350–500 °C | Air cool after holding | Tempered troostite | 45 steel 400 °C → HRC 40–44 | Springs and spring clips — balanced elasticity and toughness |
| High-temperature tempering (Q&T) | 500–650 °C | Air cool after holding | Tempered sorbite | 45 steel 560–600 °C → HB 220–250 (HRC 20–25) | Shafts, gears, connecting rods — the best overall mechanical properties |
Q&T is not quenching plus any tempering: The value of Q&T lies in through hardening. If the section is not hardened through (ferrite + pearlite in the core), the hardness after high tempering may pass but impact energy and fatigue life are distinctly low. That is why large sections (φ>40) use 42CrMo rather than 45 steel — the criterion is hardenability, not hardness.
Hardenability quick reference (critical hardening diameter, empirical values)
| Material | Critical diameter, water quench | Critical diameter, oil quench | Suitable section |
|---|---|---|---|
| 45 steel | φ15–20 mm | φ8–12 mm | Small shafts, pins, small blocks; large sections can only be surface hardened |
| 40Cr | φ30–38 mm | φ22–28 mm | Medium-section shafts, gears, connecting rods |
| 42CrMo | φ55–65 mm | φ40–50 mm | Large shafts, large gears, wind-power / hydraulic parts |
| 20CrMnTi | — | φ30–40 mm (after carburizing) | Carburized gears with a strong, tough core |
| GCr15 | About φ25 | φ20–25 mm | Bearing rings and rollers (limited by wall thickness) |
Note: the critical diameter varies with quench agitation, grain size and carbon content. This table is for initial material selection, not for acceptance. Where hardenability matters, specify "core hardness" on the drawing or accept per GB/T 225 (Jominy end-quench test).
3. Heat treatment and hardness of common materials
The table is generated from the same page data as the lookup above. Hardness values are typical ranges for conventional processes; actual values follow furnace test pieces and part sampling.
4. Surface and thermochemical treatment (surface hardening only)
| Process | Temperature / medium | Effective case depth | Surface hardness | Core | Features and applications |
|---|---|---|---|---|---|
| Gas carburizing | 900–930 °C, kerosene / methanol + propane, carbon potential 0.8–1.1 %C | 0.3–2.0 mm (gears usually 0.8–1.2) | HRC 58–63 after quenching + 180 °C tempering | Retains low-carbon toughness, HRC 30–40 | Hard surface and tough core, resistant to contact fatigue; gears, pins, chains. Grinding is needed after carburizing and distortion is significant |
| Carbonitriding | 820–880 °C, carburizing atmosphere + ammonia | 0.1–0.5 mm | HRC 58–64 | — | Low temperature, low distortion, wear and fatigue resistant; small-module gears, thin-wall parts, sewing machine parts |
| Gas nitriding | 500–560 °C, dissociated ammonia, 20–60 h | 0.2–0.6 mm | HV 900–1100 (roughly HRC 68 and above) | Q&T condition (HB 250–300) | Minimal distortion — can be the final operation with no further machining; the specialty of 38CrMoAl; long cycle and high cost |
| Plasma (ion) nitriding | 480–560 °C, vacuum glow discharge, N₂/H₂ | 0.1–0.5 mm | HV 800–1100 | Q&T condition | Faster than gas nitriding, can be locally masked, white layer controllable; long shafts, large parts, stainless steel (oxide film need not be removed first) |
| Induction hardening | Surface induction heating to 850–950 °C then immediate water spray | Medium frequency 1–10 kHz: 2–6 mm High frequency 100–500 kHz: 0.5–2 mm | HRC 50–62 | Retains the Q&T structure | Journals, tooth flanks, guideways, sprockets; Q&T must come first and there must be a transition between hardened layer and core |
| Flame hardening | Oxy-acetylene flame at 850–950 °C, immediate water spray | 2–6 mm | HRC 50–58 | Normalized / Q&T condition | Large one-offs, on-site repair, racks; quality depends on the operator and uniformity is poor |
| Laser hardening | Surface momentarily 1000–1300 °C, self-quenching | 0.2–1.0 mm | HRC 55–62 | Substrate unaffected | Least distortion, precise local hardening, no quenchant needed; die edges, thin walls, deep cavities |
| QPQ salt-bath composite | 570 °C nitriding + oxidizing salt bath, 1–4 h | Compound layer 5–20 μm, total case 0.1–0.3 mm | HV 600–1000 | — | Wear + corrosion + anti-galling with low distortion; replaces black oxide / chrome plating; shafts, valve parts, moulds |
| Vacuum heat treatment | Same as conventional processes (quench / temper / carburize) in a vacuum furnace with high-pressure gas quenching | Same as conventional | Same as conventional, with no surface decarburization | — | Bright, no decarburization, low distortion; tool steel, precision parts, titanium alloys, stainless steel |
How to size the carburized case (estimating without tables)
- Gears: Case depth ≈ (0.15–0.25) × normal module. Module 2 → 0.3–0.5 mm; module 3 → 0.5–0.75 mm; module 5 → 0.75–1.25 mm. Take the upper limit for heavy loads and the lower value for thin walls.
- Pins / bushings: Take 0.3–0.8 mm according to contact stress and wear, leaving 0.1–0.2 mm for grinding (the case depth must include the grinding allowance).
- Too shallow: crushing and spalling; too deep: the carbon gradient flattens, residual compressive stress drops, fatigue resistance actually gets worse, and cost rises.
- Determining case depth: carburizing uses "case harden depth (CHD)", the distance from the surface to 550HV1 (GB/T 9450); nitriding measures total case depth and compound layer thickness metallographically per GB/T 11354.
5. Heat treatment of stainless steels and non-ferrous metals
| Material | Hardenable by quenching? | Process | Typical hardness | Key points |
|---|---|---|---|---|
| 304 / 316 austenitic stainless steel | No (no martensitic transformation) | Solution treat 1010–1120 °C, water quench (rapid cooling prevents carbide precipitation) | Solution-treated HB ≤187 | Can only be strengthened by cold work (cold rolled 1/2H → HB ≈250); stress relief at 350–400 °C, avoiding the 450–850 °C sensitization range |
| 420 / 3Cr13 martensitic stainless steel | Yes | 1000–1050 °C oil quench + 200–300 °C temper | HRC 48–53 | Tools, valve cores, pump shafts; corrosion resistance falls as the tempering temperature rises |
| 440C / 9Cr18 | Yes | 1030–1050 °C oil quench + 150–200 °C temper | HRC 58–60 | Bearings, tools, nozzles; much retained austenite after quenching, so cryogenic treatment is needed |
| 17-4PH (05Cr17Ni4Cu4Nb) | Yes (precipitation hardening) | 1040 °C solution + 480 °C × 1 h aging (H900) | HRC 40–45 | Low distortion with balanced strength and toughness; raising the aging temperature to 620 °C → HRC 32–38 (H1075) |
| 6061 aluminium alloy | Yes (age hardening) | 530 °C solution, water quench + 175 °C × 8 h artificial aging (T6) | HB 95, σb ≥310MPa | The mainstream temper for extrusions and CNC parts; T651 = solution + stretch stress relief + aging, with little machining distortion |
| 7075 aluminium alloy | Yes | 470 °C solution, water quench + 120 °C × 24 h (T6) | HB 150, σb ≥570MPa | Aerospace structures; poor corrosion resistance, so T73 or Alclad is usual |
| 5052 aluminium alloy | No | Work hardening only (H32 / H34) | H34 HB 60–77 | Sheet metal, fuel tanks, extrusions; annealed O temper is the softest — local annealing before bending |
| 2024 aluminium alloy | Yes | 495 °C solution, water quench + natural aging (T4) | HB 120, σb ≥470MPa | Riveted structures; often Alclad to improve corrosion resistance |
| A356 cast aluminium | Yes | 540 °C solution + 155–165 °C × 4–6 h (T6) | HB 70–100, σb 260MPa | Die cast and gravity cast parts; die castings contain porosity and blister at high solution temperatures, so T5 is usual |
| TC4 (Ti-6Al-4V) | Yes (transformation + aging) | Anneal 700–800 °C; or 950 °C solution + 540 °C aging (STA) | Annealed HRC 33–38; STA HRC 40–45 | Vacuum or inert atmosphere is mandatory; the oxygen-contaminated layer must be pickled off |
The most common mistake:"Quench to HRC50" on a 304/316 drawing is impossible — austenitic stainless has no martensitic transformation and quenching only solution-softens it. Either change the material (420/440C/17-4PH) or the process (cold work, nitriding). Domestic drawings often write "solution treatment" as "quenching"; inspect against the solution-treated condition (corrosion resistant, non-magnetic, soft).
6. Hardness conversion and strength estimation
| HRC | HBW (approx.) | HV (approx.) | Estimated tensile σb for carbon steel | Typical correspondence |
|---|---|---|---|---|
| 20 | 226–238 | 238–247 | 750–800 MPa | Upper limit of Q&T 45 steel; carburized core of low-carbon steel |
| 25 | 253–266 | 266–285 | 850–900 MPa | Q&T 40Cr; general structural parts |
| 30 | 286–302 | 302–316 | 960–1020 MPa | Q&T 42CrMo; grade 8.8 bolts |
| 35 | 327–345 | 345–361 | 1100–1160 MPa | Grade 10.9 bolts; high-strength shafts |
| 40 | 371–392 | 392–409 | 1250–1320 MPa | Grade 12.9 bolts; pre-hardened mould steel |
| 45 | 415–442 | 442–458 | 1400–1490 MPa | Hot-work tool steel (typical H13 range) |
| 50 | 481–513 | 513–530 | 1620–1730 MPa | Induction-hardened tooth flanks and cutting edges |
| 55 | 560–599 | 599–618 | 1890–2020 MPa | Quenched 45 steel; connecting-rod dies |
| 60 | 654–697 | 697–720 | 2150–2250 MPa | Bearing steel, cold-work dies, carburized case |
| 65 | 739–795 | 795–820 | 2490–2680 MPa | Upper limit of quenched GCr15; converted nitrided-case HV |
Conversion is only a reference, never an acceptance basis: Use GB/T 1172 (conversion values for hardness and strength of ferrous metals) for accurate conversion. Three hard rules: ① conversion tables hold only for carbon and low-alloy steels — not for stainless, cast iron or non-ferrous metals; ② the same hardness can differ by 30 % in strength between materials, so never back-calculate strength from hardness to check a strength value; ③ below HRC 20 the error grows sharply, so measure HBW or HV instead. Estimate tensile strength as σb ≈ (3.3–3.4) × HBW (MPa).
7. Heat treatment defects and inspection
| Defect | Symptom | Common cause | Inspection / prevention |
|---|---|---|---|
| Oxidation and decarburization | Low surface hardness; ferrite in the surface layer | No protective atmosphere, poor furnace sealing, oily parts | Measure the decarburized layer metallographically (GB/T 224); switch to protective atmosphere / vacuum / anti-oxidation coating; the decarburized depth must be less than the subsequent machining allowance |
| Overheating | Coarse grains and coarse martensite; lower impact toughness | Excessive temperature or holding time | Grain size rating (GB/T 6394), normally ≥5; normalize to recrystallize and refine, then re-quench |
| Burning | Grain boundaries melted and oxidized; parts break with a tap | Temperature near the solidus, furnace out of control | Metallography; beyond repair and must be scrapped — the costliest accident in a heat-treat shop |
| Insufficient hardness | Below the specified value after quenching | Underheating, slow cooling, insufficient hardenability, surface decarburization, overloading the furnace | Multi-point hardness testing (GB/T 230.1); check quenchant temperature, agitation and transfer delay |
| Soft spots | Locally low hardness (patches or spots) | Vapour blanket not broken, parts touching, oil or scale on the surface | Grid-sample hardness; increase quenchant agitation, space parts apart, clean before quenching |
| Excess distortion | Bending, ovality, flatness out of tolerance | Uneven heating and cooling, poor loading, residual stress, large length-to-diameter ratio | Measure runout / flatness; hang slender parts in a pit furnace; martempering / austempering; press-quenching fixtures |
| Quench cracking | Cracks at sharp corners, hole edges and section changes | Too fast cooling, abrupt section changes without fillets, no preheat, high carbon content | Magnetic particle / fluorescent penetrant inspection; in design enlarge fillets and avoid sharp corners and blind holes; slow cooling near the Ms point |
| Excessive retained austenite | Low hardness; dimensions unstable over time | Excessive quench temperature, low Ms point (high-carbon, high-alloy) | Metallography + X-ray retained-austenite measurement; cryogenic treatment (-60 to -120 °C) + multiple tempers |
| Case too shallow / too deep | Case depth out of drawing tolerance | Unstable control of temperature, time, carbon / nitrogen potential | Furnace test pieces + metallography / hardness-gradient method (550HV1 limit); a test coupon in every load is the only reliable way |
| Case spalling | Surface flaking and crushing | Excessive white layer (nitriding), excessive carbon forming network carbide, core hardness too low | Examine white layer and carbide morphology metallographically; control carbon and nitrogen potential; nitrided parts must be properly Q&T in the core |
Inspection checklist (for the inspection instruction sheet)
| Item | Method / standard | Acceptance point |
|---|---|---|
| Surface hardness | Rockwell HRC (GB/T 230.1); Vickers HV for small parts / thin layers (GB/T 4340.1) | Average 3–5 points, 0.1–0.2 mm from the surface or per drawing; HRC is forbidden on thin walls (it penetrates) |
| Core hardness | HRC or HBW | Mandatory for carburized and induction-hardened parts; shows whether the base is properly Q&T |
| Effective case depth | Hardness-gradient method (550HV1) or metallography | Carburizing GB/T 9450; nitriding GB/T 11354; furnace test pieces are mandatory |
| Microstructure | Martensite grade, grain size, decarburized layer, carbide morphology | Grain size ≥5; rate martensite + retained austenite in carburized parts per GB/T 25744 |
| Distortion | Runout, flatness, dimensions | First piece after heat treatment is mandatory; 5–10 % sampling in production |
| Cracks | Magnetic particle (MT) / fluorescent penetrant (PT) | Focus on quenched parts and ground surfaces (grinding cracks); 100 % inspection on critical load-bearing parts |
8. How to write it on the drawing (hardness / case depth / process)
| Requirement type | Recommended wording | Wrong wording and consequence |
|---|---|---|
| Through Q&T | Q&T to 220–250 HBW | "HRC22–25" alone — mixing HB and HRC with no process given, so the supplier may hit the hardness by normalizing |
| Through hardening | Quench + temper to 45–50 HRC | "Heat treat HRC50" without material or tempering temperature — properties cannot be reproduced |
| Local surface hardening | Induction harden tooth flanks to 50–55 HRC, effective case depth 1.0–1.4 mm (550HV1) | "Surface harden" alone with no case depth — too shallow crushes, too deep chips the teeth |
| Carburize and harden | Carburize and harden: surface 58–62 HRC, CHD 0.8–1.2 mm, core 30–42 HRC | "Carburize 1 mm" without hardness or core requirements — the supplier may carburize without hardening |
| Nitriding | Plasma nitride: surface HV 550–650, case ≥0.30 mm, white layer ≤0.02 mm | "Nitride HRC60" — the nitrided case is only tens of microns, so HRC penetrates to the substrate; HV is mandatory |
| Stress relief | Stress-relief anneal after rough machining, 600 °C × 2 h | Left out, so finishing follows welding / roughing directly and distortion shows up at assembly |
| Stainless steel | Solution treat, HB ≤187 | "Quench HRC50" — physically impossible, so the supplier can only fake the report |
Where to call it out and tolerances: Hardness normally goes in the technical requirements ("3. Q&T to 220–250 HBW"); local requirements are called out with a leader to the surface concerned. Normal hardness tolerance is ±3 HRC or an HBW range of 20–30. When one part carries several different hardness values, define the zone boundaries (thin solid line plus locating dimensions), otherwise inspection cannot decide where to measure.
9. Heat treatment equipment and cost characteristics
| Equipment | Suitable parts / processes | Cost characteristics | Economic batch |
|---|---|---|---|
| Box resistance furnace | Annealing, normalizing and tempering of small parts; occasional quenching | Lowest, charged per furnace load | One-off to small batch |
| Pit furnace | Long shafts, rods, lead screws (hung to prevent bending), carburizing | Low to medium | Medium batch |
| Salt bath furnace | Quenching of small and medium parts, high-speed steel, local heating | Medium (salt consumption and high environmental treatment cost) | Medium batch |
| Mesh-belt / roller-hearth continuous furnace | Quench and temper of fasteners, bearings, standard parts | Lowest cost per part | Large batch (continuous production) |
| Vacuum furnace | Tool steel, precision parts, titanium, stainless steel — bright quenching | High (small load, long cycle) | Medium batch |
| Induction heating equipment | Local hardening of journals, tooth flanks, guideways | Medium (coils and fixtures are expensive, cycle time per part is short) | Best value in large batches |
| Continuous carburizing line / multi-purpose furnace | Batch carburizing of gears and pins | Medium (charged by case depth × time) | Large batch |
| Plasma nitriding furnace | Long shafts, large parts, stainless steel, moulds | Medium-high | Medium batch |
Quoting rules of thumb: heat treatment is usually charged by weight × unit price or by furnace load, with carburizing / nitriding priced up according to case depth (doubling the case depth multiplies the time by about 3–4). A single prototype nitriding job often costs more than the part itself — for small quantities prefer QPQ or surface hardening.
10. Linking heat treatment with machining (process chain)
- Leave enough allowance: Quench distortion is about 0.1–0.5 % (more for high L/D, thin walls, abrupt section changes). Leave 1.5–3 mm before quenching, 2–4 mm for Q&T parts, and "case depth + 0.1–0.2 mm grinding allowance" for carburized parts.
- Typical process chain: Blank (cast / forged / extruded) → normalize or anneal → rough machine → stress relieve → Q&T → semi-finish → quench / carburize → grind → surface treatment. For nitrided parts, nitriding goes after grinding and before polishing.
- After quenching only grinding, wire EDM, sinker EDM and lapping are possible: drilling and tapping above HRC 45 is essentially impossible (even carbide drills chip), so threads, keyways and oil holes must be completed before quenching.
- Slender parts (L/D>10): Must be hung in a pit furnace and quenched vertically, otherwise bending far exceeds the allowance. Lead screws, guideways and long shafts fall in this class.
- Thin-wall parts (wall <3 mm): Prefer vacuum quenching / martempering / press quenching, or switch to nitriding or laser hardening (least distortion).
- Welded assemblies: Stress relief (550–650 °C) after welding and before finishing, otherwise parts distort within days; avoid putting final fitting surfaces near welds.
- Nitriding comes last: Sizes grow 2–10 μm after nitriding (compound layer growth), so precision fits need allowance or post-nitriding lapping; do not plate or blacken after nitriding.
- Order of heat treatment and surface treatment: Quench / temper → hydrogen bake-out (high-strength parts) → blasting / polishing → plating / blackening / anodizing. The reverse (plate then quench) burns the coating and compounds hydrogen embrittlement.
Three notes for designers: ① Higher hardness is not better — first decide the split between "tough core, wear-resistant surface", then choose the process; ② a surface-hardening requirement without case depth and core hardness is no requirement at all; ③ fillets in the design, no sharp corners or blind holes, no abrupt section changes — this prevents cracking better than any quenchant.
11. Distortion after slitting: residual stress release and wire-EDM slotting
Applies to: bored parts, sleeves, bushings, split clamping collars, flange rings, thin-wall rings — parts slit or slotted by wire EDM after heat treatment, where bore size, roundness or flatness goes out of tolerance.
1) Why it distorts: not "cut crooked" — the stress is being released
| Mechanism | What happens | Typical symptom |
|---|---|---|
| Residual stress release (main cause, over 70 %) | During quenching the surface cools first and the core later, so surface compression and core tension lock each other in a closed ring and balance macroscopically. Slit the ring and it becomes a C-shaped spring; the constraint is gone, stress rebalances, and the ring opens or closes. | Bore grows or shrinks overall, roundness worsens; the slit opens or closes; several parts in the same lot move the same way by similar amounts |
| Transformation stress (retained austenite / aging) | Retained austenite keeps transforming to martensite during storage or tempering and expands in volume; after slitting the constraint is reduced, transformation is freer and dimensions keep drifting. | Passes right after slitting, keeps changing 24–72 h later; repeated measurements differ |
| Cutting heat stress | The discharge zone reaches thousands of degrees, forming a white layer (recast layer) and local tensile stress on both sides of the kerf; thin walls with insufficient rigidity are pulled out of true. | Whitened kerf edges, chipping, microcracks; local collapse at the bore mouth and a "chamfer" near the slit |
| Clamping force release | Before slitting the part is held in a vice, clamps or on a magnetic chuck — it is already being forced out of shape — and it springs back when released. | Distortion depends on clamping position; the part moves as soon as it is released, and a different fixturing gives a different number |
One-line test: If it "passed before slitting, failed after, and parts in the same lot move the same way by similar amounts", it is almost certainly residual stress release — the responsibility lies with heat treatment (insufficient tempering / no stress relief), not with wire EDM. If the directions are mixed and the magnitudes scattered, check fixturing, uneven wall thickness and grinding stress first.
2) Slitting distortion estimator
Calculated for a ring with uniform wall thickness, linear stress through the wall and one through-slit. Use it to judge magnitude and responsibility, not for acceptance.
Formula:ΔD ≈ D²·σ / (E·t); slit opening ≈ 3.14·ΔD (the same distortion seen another way: the whole circumference change piles up at one slit).
Typical residual stress: quenched and untempered 400–700 MPa; quenched + 180–200 °C temper 200–350; Q&T (500–600 °C temper) 80–150; after stress-relief annealing <50. Actual distortion is usually 30–60 % of the formula value (the stress profile is not perfectly linear and the kerf has width).
3) Order of magnitude: how thick a wall can take it
| Case | D / t | σ (MPa) | Estimated ΔD | Equivalent to |
|---|---|---|---|---|
| φ25 thin-wall sleeve, quenched and low tempered | 25 / 3 | 300 | 0.30 mm | IT7 for φ25 is 0.021, 14× over |
| φ40 sleeve, Q&T | 40 / 5 | 150 | 0.23 mm | IT7 = 0.025, 9× over |
| φ40 sleeve, quenched and low tempered | 40 / 5 | 320 | 0.50 mm | 20× over— slitting scraps the part |
| φ60 thick-wall ring, quenched and low tempered | 60 / 10 | 300 | 0.52 mm | IT7 = 0.030, 17× over |
| φ60 same size, fully stress relieved | 60 / 10 | 80 | 0.14 mm | Still 4× over, but can be ground back with allowance |
| φ120 large flange ring, Q&T | 120 / 15 | 200 | 0.93 mm | Close to 1 mm— must be slotted before finishing |
The pattern:ΔD is proportional to D² and inversely proportional to t . So large-diameter thin-wall parts are the most sensitive: double the diameter and distortion quadruples; halve the wall and it doubles again. "Improving wire-EDM accuracy" does nothing for these parts.
4) Assigning responsibility: who caused the distortion
| Symptom | Main cause | Responsible stage | Action |
|---|---|---|---|
| Passes before slitting, bore moves 0.05–0.3 mm after, same direction across the lot | High residual stress: insufficient tempering, delayed tempering after quench, no stress relief | Heat treatment | Add a stress-relief temper (20–30 °C below the original tempering temperature, hold 2–4 h, furnace cool); change the routing for the next lot |
| Passes right after slitting, keeps changing after 24–72 h | Retained austenite transformation + stress relaxation (aging distortion) | Heat treatment | Cryogenic treatment (-60 to -120 °C) + two or more tempers; stabilize before finishing |
| Already oval, tapered or uneven in wall thickness before slitting | Uneven quench cooling, poor loading, off-centre blank | Heat treatment / machining | Quenched parts cannot be cold straightened (they will crack) — only grinding back with allowance, or switch to press quenching / martempering |
| Whitened kerf sides, microcracks, chipping | Electrical parameters too high, single pass only, dirty dielectric | Wire EDM | Multiple passes (rough cut leaving 0.05–0.1, then 2–3 skim passes); check white-layer depth metallographically and remove it by electropolishing or grinding if needed |
| Top and bottom of the same part distort differently | Uneven wall thickness, different cooling at top and bottom | Machining / heat treatment | Measure wall-thickness variation (coaxiality) first and return out-of-tolerance parts to machining; in heat treatment switch to vertical hanging and stronger agitation |
| Changing the fixturing changes the distortion | Clamping force release; inconsistent measurement datum | Inspection / process | Specify a uniform clamping force and measuring condition in the inspection procedure (see Section 6) |
5) The cure is process sequence, not precision
Standard routing for split / thin-wall sleeves:
| Op. | Operation | Key points |
|---|---|---|
| 1 | Rough turn / rough mill | Leave 1.5–2 mm, removing the as-cast skin and uneven layer |
| 2 | Q&T or normalize | Homogenizes the structure and prepares for final hardening; distortion at this stage does not matter |
| 3 | Semi-finish turn | Leave 0.3–0.5 mm |
| 4 | Stress-relief anneal 600–650 °C | A key operation: 2–4 h then furnace cool, clearing the stress from roughing and Q&T |
| 5 | Finish turn / rough grind | Leave 0.05–0.1 mm on the bore for finishing after slotting |
| 6 | Quench + temper | Temper within 1–4 h of quenching; for thin walls prefer martempering / austempering / press quenching |
| 7 | Wire-EDM slotting | After quench and temper, before finish grinding — the distortion has already happened and there is still stock to correct it |
| 8 | Stress-relief temper (20–30 °C below the tempering temperature) | Slotting itself introduces cutting stress, so stabilize once more |
| 9 | Finish grind / hone the bore | Grind while clamped in a process sleeve, so the ground size is the assembly size |
Three hard rules:
① Slot before quenching and the part is scrap. The slot is a stress raiser — it opens during quenching and cracks very easily.
② Quenched parts must not be cold straightened. Straightening cracks them; distortion can only be ground back with allowance or avoided by press quenching.
③ The drawing must state the measuring condition. A split part's bore can differ by more than 0.1 mm between free and clamped states; without a statement, buyer and supplier will argue forever.
① Slot before quenching and the part is scrap. The slot is a stress raiser — it opens during quenching and cracks very easily.
② Quenched parts must not be cold straightened. Straightening cracks them; distortion can only be ground back with allowance or avoided by press quenching.
③ The drawing must state the measuring condition. A split part's bore can differ by more than 0.1 mm between free and clamped states; without a statement, buyer and supplier will argue forever.
6) How to measure slit parts and how to write the drawing
| Measuring condition | Applies to | Note |
|---|---|---|
| Free state | Parts that are split by design and not clamped in assembly | CMM / inside micrometer directly; plug gauges do not apply (they spread a split part) |
| Clamped (assembly) state | Clamping sleeves, expanding mandrels, clamp collars — shaped by clamping force | Use a process sleeve or expanding mandrel, and the clamping force / torque must be written into the specification; fit the CMM data to the assembly constraint |
| Before / after slitting comparison | Verifies residual stress level and validates the process | Measure twice at the same section, same point and same temperature; record the interval (0 h / 24 h / 72 h) |
Drawing callout example:
Wrong wording: writing only
φ40 H7 (measured in clamped assembly state, clamping torque 25 N·m, gap 3 mm)Heat treatment: quench + 200 °C temper, hardness 58–62 HRC; after slotting, stress relief at 180 °C × 3 hWrong wording: writing only
φ40 H7— with no measuring condition, slit parts always fail if judged in free state, and there is no clamping-force basis if judged in clamped state.7) Problems wire EDM itself leaves behind
| Item | Value / symptom | Inspection and countermeasure |
|---|---|---|
| White layer (recast layer) | 5–20 μm in a single pass; hard but brittle, containing microcracks | Measure white-layer depth on a metallographic section; use multiple passes on critical parts, or remove the layer by electropolishing / grinding |
| Dimensional accuracy | Wire EDM (slow) ±0.003–0.005 mm; (fast) ±0.01–0.02 mm | Kerf width 0.2–0.35 mm must be deducted in strength calculations |
| Cracking of quenched parts | Above 55 HRC with insufficient tempering, parts crack during or after cutting | Confirm adequate tempering (normal hardness gradient) before cutting; cut from outside inward and leave support at the last break-through |
| Corner collapse near the kerf | Thin walls <2 mm; discharge concentrates at the edge | Reduce peak current and add skim passes; blunt critical edges first |
One more point that is easily overlooked: Cutting a specimen for metallography or hardness releases stress the moment it comes free, so the measured size no longer represents the part itself. To judge size, measure before slitting , or use a furnace test piece from the same load for structure and hardness.