Heat Treat Mistakes

Heat Treatment Troubleshooting: Costly Mistakes and How to Catch Them

A troubleshooting guide to the errors that wreck heat treat numbers, from soak-time-per-inch mistakes to load density and quench errors, each with a symptom and a numeric fix.

The most common error is loading a furnace by piece count instead of load density. A rack rated for 1,200 lb of gears at 45 lb per cubic foot will not deliver the same result packed to 70 lb per cubic foot. Symptom: the outer parts hit temperature in 40 minutes while the core of the load lags 25 to 40 minutes behind, and hardness scatter jumps from plus or minus 1 HRC to plus or minus 4 HRC. Root cause: convection paths choke and radiant view factors drop. Fix: hold density near 40 to 50 lb per cubic foot for still-air furnaces. Check it with the Furnace Load Density calculator before you commit a batch.

Soak time errors trace back to a wrong per-inch rule. Practitioners quote 'one hour per inch' as gospel, but that number is for through-heating carbon steel in a radiant furnace, not a 1.5 inch H13 die block. Symptom: incomplete austenitizing shows up as soft spots reading 8 to 12 HRC below target. Root cause: soak was timed from door-close, not from when the part surface reached setpoint. Fix: start soak only after the load thermocouple reaches within 10 degrees F of setpoint, then hold 30 minutes per inch of ruling section for alloy steel, roughly double the carbon-steel figure. Model it in the Annealing Cycle Time calculator.

Thermocouple placement is the quiet killer of every energy and cost number. Symptom: your Furnace Energy Cost output looks 15 to 20 percent low because the control couple sits near the burner, reading 1,560 degrees F while the load sits at 1,510. Root cause: control couple versus load couple mismatch, plus drift on an uncalibrated type K couple that can read 15 to 30 degrees F high after 2,000 hours. Fix: run a survey with a load couple in the coldest zone, and recalibrate on a 90-day cycle. A 20 degree F control error on a 1,700 degree F cycle inflates gas burn by roughly 3 to 5 percent per point of overshoot.

Ignoring furnace utilization corrupts every per-part figure. Symptom: your Heat Treat Cost per Part number doubles between two identical jobs run a week apart. Root cause: one batch filled 90 percent of the working volume, the other ran at 35 percent, but the fixed 6-hour cycle burned the same 480 cubic feet of gas either way. Fix: track loaded volume against rated volume and never quote off a full-load assumption. A furnace at 40 percent utilization spreads the same energy and labor over 40 percent of the parts, so cost per part climbs 2.5x. Confirm with the Furnace Utilization and Batch Heat Treat Capacity calculators.

Quench mistakes hide until distortion or cracking appears. Symptom: parts warp beyond a 0.010 inch flatness limit or show quench cracks at stress risers. Root cause: quench media chosen by habit, not severity. Water has an H value near 1.0 to 2.0 and will crack a hardenable alloy that only needs oil at H 0.25 to 0.50. A second cause is bath temperature: fast oil run at 180 degrees F instead of the specified 120 to 140 degrees F loses 20 to 30 percent of its cooling power. Fix: match Grossmann H value to hardenability and log bath temperature every batch. Price alternatives with the Quench Media Cost calculator.

Tempering gets shortchanged when operators count ramp time as soak. Symptom: as-quenched brittleness survives into service, and impact values sit 40 percent below spec. Root cause: a 2-hour temper timed from door-close gives maybe 70 minutes of true soak on a heavy load that needs a full 2 hours at temperature after equalization. Fix: time temper soak from load-couple equalization, and apply the one-hour-per-inch minimum with a 2-hour floor. Double tempering is not optional for high-alloy tool steels: skipping the second cycle leaves 5 to 15 percent retained austenite untempered. Cost the cycles honestly with the Tempering Cycle Cost calculator.

Scrap gets buried in overhead instead of costed to the process that made it. Symptom: a line runs a 3 percent scrap rate on paper but margins bleed anyway. Root cause: rework and downgraded parts are not counted, and the true loss on a hardened part includes all upstream machining value, not just the raw bar. A part scrapped after grinding can carry 8 to 12 times its material cost. Fix: cost scrap at fully loaded value at the point of failure, not at raw-material value. Use the Heat Treat Scrap Cost calculator, and if hardness scatter drives the scrap, quantify it with the Hardness Variation calculator before blaming the furnace.

Unit slips create errors that survive review because the number looks plausible. Symptom: an energy quote comes in 3.4x low. Root cause: mixing therms and kWh. One therm equals 29.3 kWh, so a cycle you logged as 400 kWh of gas is really 1,364 kWh of thermal energy if the meter reads therms. Another classic: specific heat of steel is 0.11 Btu per lb per degree F, not 0.11 kJ, a 2.1x error. Fix: carry units through every line and sanity-check against a known figure, roughly 250 to 400 Btu per lb to austenitize and quench typical steel. When two calculators disagree by a clean factor like 3.4 or 2.1, suspect a unit conversion first.

Published 2026-07-01.