Soft Shafts After Hardening and Tempering
A heat-treatment bay must supply 40 splined shafts, Ø30 × 220 mm, made from EN19 (AISI 4140) bar with a mill certificate showing 0.40 % C, 1.05 % Cr and 0.20 % Mo. The drawing calls for 40–45 HRC after hardening and tempering, with 0.30 mm radial grinding allowance left on all diameters. The parts are charged cold into an electrically heated box furnace without atmosphere control, soaked 45 minutes at 860 °C, quenched into a 900-litre oil tank at 55 °C with no agitation, then tempered for 2 hours at 540 °C. Results: hardness 33–36 HRC on the surface, a decarburised layer of 0.12 mm confirmed by microhardness traverse, and two shafts cracked through the keyway corner during quenching. The furnace thermocouple was replaced last month and has not been calibrated since. The oil has not been tested for two years and may contain water. The buyer needs a decision today: rework, or remake from bar.
1. Separate the possible causes of the low hardness reading and describe how you would prove which one is operating.
Read the worked answer
Three candidate causes must be separated, because they lead to different remedies.
- Tempering too severe. For a 0.40 % C, 1 % Cr, 0.2 % Mo steel, 540 °C tempering gives roughly 33–38 HRC, which matches the readings exactly. Published tempering curves put 40–45 HRC near 440–480 °C for this analysis. This is the most probable single cause of a uniform 33–36 HRC across the whole batch.
- Decarburised skin. A 0.12 mm low-carbon layer reads soft on any surface hardness test and would give low values even if the core were correct. Proof: grind 0.3 mm off one sample and re-test, or run a microhardness traverse from surface to core.
- Inadequate austenitising or quench severity. A mis-reading thermocouple could mean the parts never reached 860 °C, or water in the oil could give soft spots. Proof: check hardness distribution across the section (uniform low = tempering or austenitising; patchy = quenching), verify furnace temperature with a calibrated survey thermocouple, and check the oil for water and for cooling-curve deterioration.
The mill certificate rules out wrong chemistry, so that hypothesis is closed.
2. Decide between rework and remake, and give the rework route with parameters, using the grinding allowance in your reasoning.
Read the worked answer
Rework is viable. The decarburised layer is 0.12 mm deep and the radial allowance is 0.30 mm, so grinding 0.30 mm per side removes the whole decarburised zone and still leaves the finished size (the allowance is there for exactly this). Had the decarburisation exceeded the allowance, the parts would be scrap.
Route:
- Scrap the two cracked shafts. A quench crack at a keyway corner will not close and will grow on re-hardening; the corner also needs a larger fillet radius or chamfer before any remake.
- Grind 0.30 mm per side to clean up, then verify by microhardness or by a spot test that no soft skin remains.
- Re-austenitise at 850–870 °C with protection - protective atmosphere, controlled atmosphere or a proprietary protective coating, since the furnace has none.
- Quench into agitated oil at 50–60 °C; agitate the load, do not bundle the shafts, and temper immediately, never letting parts sit un-tempered.
- Double temper, 2 hours each, at 450 °C as a starting point, and adjust from a coupon tested on the same cycle to hit 40–45 HRC.
- Check hardness after grinding, not on the as-quenched skin, and check 10 % of the batch plus all first-off parts.
3. What system-level changes would you insist on before the next batch is released?
Read the worked answer
Instrumentation: calibrate the new thermocouple and survey the furnace with a calibrated probe at three points; a furnace without a valid temperature survey is not a heat-treatment furnace, it is an oven.
Atmosphere: with no atmosphere control, every batch of this steel will decarburise. Either fit atmosphere control, use protective coating or pack the parts, or accept and plan for a machining allowance larger than the decarburisation depth and inspect the layer each batch.
Quench: test the oil for water content and cooling characteristics and set a routine sampling interval; add agitation, because a still 900-litre tank cannot give reproducible cooling around a 30 mm section.
Process discipline: the tempering temperature must be backed by a coupon test on the actual furnace, not by a wall chart; one test piece per cycle is cheap insurance against a 40-piece rework. Record austenitising temperature, soak time, quench start delay and tempering parameters on a route card, and require hardness approval after finish grinding before the batch moves to assembly.