Rapid Tempering and Stress Relief Via High-Speed Convection Heating
Rapid stress-relieving and tempering using salt-bath processing and induction heating are well-known processes, but these terms are rarely used in conjunction with processes conducted in a conventional convection oven. However, rapid-heating technology using a proprietary high-speed convection oven with turbulent flow is used to achieve not only faster processing times, but also improved quality and properties.
Rapid heating is defined as a heating method that accelerates conventional furnace heating . Heat transfer rates of up to 30 times those achieved in conventional convection furnaces are possible . In the past, rapid heating technology was primarily applied in the forging industry where steel is heated to 1000 to 1250 C (1830 to 2280 F), but it is less commonly encountered in the heat-treating industry. Rapid heating is enjoying increasing use for stress relieving, but it still is in its infancy for tempering processes.
Stress relieving typically is used to remove residual stresses that have accumulated from prior manufacturing processes. The process involves heating to a temperature below A(c1) (for ferritic steels) and holding at that temperature for the time required to achieve the desired reduction in residual stresses, followed by cooling at a rate slow enough to avoid creating excessive thermal stresses. No microstructural changes occur during stress relieving. Stress relieving results in a significant reduction of yield strength in addition to reducing residual stresses to some "safe" value in crack-sensitive materials.
Typically, stress-relieving times for specific alloys are obtained from standards such as those listed in Table 1 [3-5], which were developed for conventional convection-heated batch ovens.
Stress-relieving results depend on temperature and time, which are correlated through Holloman's parameter, P  given in Equation 1 where T is temperature in K, t is time (h) and C is the Holloman-Jaffe constant, which is calculated from equation 2 where P is a measure of the thermal effect of the process. Processes having the same Holloman's parameter show the same effect.
A similar commonly used expression to evaluate the stress relief of spring steels is the Larson-Miller equation [3-5,7], shown in Equation 3.
Pyromaitre developed software (Pyrograph) that uses the Larson-Miller equation to model production heating requirements for selecting a high-speed stress-relieving oven. Figure 1 shows an example of a typical output.
Tempering, also known as "drawing," is the thermal treatment of hardened and normalized steels to obtain desired mechanical properties including improved toughness and ductility, lower hardness and improved dimensional stability. When steel is hardened, the as-quenched martensite is very hard and brittle. During tempering, as-quenched martensite is transformed into tempered martensite, which is composed of cementite spheroids (carbides) dispersed in a soft ferrite matrix, resulting in reduced hardness and increased toughness. The objective is to allow hardness to decrease to the desired level and then to stop the carbide decomposition by cooling. The extent of the tempering effect is determined by temperature and time [8,9].
Tempering involves heating hardened steel to some temperature below the eutectoid temperature . Tempering is performed as soon as possible after the steel cools to between 50 and 75 C (120 and 170 F) and room temperature to reduce the potential of cracking. Tool steels that cannot be tempered immediately after quenching should be held at 50 to 100 C in an oven until they can be tempered . Tempering can be conducted at any temperature up to the lower critical temperature [A(c1)].
Typically, tempering times are a minimum of approximately one hour. Thelning has reported a rule of thumb of 1-2 hours/in. of section thickness after the load has reached a preset temperature . After heating, the steel is cooled to room temperature in still air. Recommended tempering conditions and heat-treating cycles for a wide range of carbon and alloy steels is provided in SAE AMS 2759.
Tempering times and temperatures also can be calculated using various methods. The Larsen-Miller equation, although originally developed to predict creep data, is a common method used to predict the tempering effect of medium/high alloy steels . Bofors  reported that a Holloman-Jaffe constant (C = 20) is appropriate for all steels, while Grange and Baughman reported that C = 18 should be used .
Sinha  shows that the Holloman-Jaffe constant varies with carbon content and desired hardness (Fig. 2a). The incremental contribution to hardness of each alloying element of a steel may be determined from Table 2 . Hardness (DPH) is calculated by multiplying the concentration of each of the alloying elements (within the range shown) times the factor for that element at a given constant C, and all of the values then are added together to provide the DPH. The interrelationship between tempering time and the Holloman-Jaffe parameter at different tempering temperatures is shown in Fig. 2b .
Spies [6,14] gives the interrelationship between tempering temperature, time and steel chemistry as shown in Equation 4 where HB is the Brinell hardness after hardening and tempering, H(h) is the Rockwell (HRC) hardness after hardening, and T(t) is the tempering temperature in C. The equation was developed for the conditions: H(h) = 20-65HRC, 0.20-0.54% C, 0.17-1.40% Si, 0.50-1.90% Mn, 0.03-1.20% Cr and T(t) = 500-650 C (930-1200 F).
High Speed Convection
Tempering can be performed in convection furnaces, salt baths and even by immersion in molten metal. Induction tempering and flame heating also are used but will not be discussed further here. Convection furnaces are the most common systems, and it is important that they have fans and/or blowers to provide for uniform heat transfer when heating the load. Typically, convection tempering furnaces are designed for use at 150 to 750 C (300 to 1380 F).
One of the most important, critical deficiencies of most conventional stress-relieving and tempering ovens is a lack of temperature uniformity in the material being heated. This is illustrated in Fig. 3, which shows the actual temperature of the part depends on its placement in the basket in the oven. A high-precision oven not only requires uniform temperature, but also uniform heat transfer. Turbines are used to achieve rapid heat transfer, but this is not sufficient to achieve uniform heat transfer. The latter depends on the number and position of turbines used, the shape of the heating chamber and placement of the component on the belt to optimize surrounding airflow. Thus, heating time and temperature uniformity of the part is optimized.
Optimal stress relief for a given material is not a fixed value of one specific time and temperature as suggested in Table 1. This is illustrated by x-ray diffraction data shown in Figs. 4a and 4b [3-5]. These data show that five minutes at 800 F is equivalent to 770 F using a Pyro high-speed stress relief oven. Also, ten minutes at 700 F in the same oven is equivalent to 25 minutes in a conventional batch oven at the same temperature.
Continuous-process simulation in a batch oven
Typically, stress-relieving and tempering processes are developed in the laboratory using a conventional batch convection oven. However, it is not unusual to obtain relatively poor correlation between the two ovens. One reason is because it is essential to not only model the actual process temperature, but also to model the heat up and cool down processes. Such modeling can be performed in a Pyro Try-Out Oven, a high-speed convection batch oven that has a very tight temperature tolerance of +/-3 C (+/-5 F) throughout the heating zone. The desired heat-up and cool-down process also can be programmed. The oven allows simulating any in-line stress-relieving process for any oven of any size. Figure 5 shows an example of attainable heating and cooling curves.
Stress relieving and tempering examples
A comparison of high-speed stress relief (7 min in a Pyro high-speed stress-relief oven) and conventional stress relief (35 min) for automotive-engine valve springs illustrates the Pyro oven capabilities. In both cases, the stress-relief temperature was 220 C (430 F). The tensile strength for the high-speed process was 275.5 ksi versus 275.9 ksi (1.89 GPa vs. 1.9 GPa) for the conventional process. The as-received wire samples had an average tensile strength of 275.9 ksi. The average hardness was essentially the same: 50.7 and 50.8 HRC for rapid stress relief and conventional, respectively. Average microindentation hardness of the high-speed process was 0.6 HRC higher than the conventional process.
Ten valve springs from each process were fatigue tested for 50-million cycles, with two failures for each stress relieving process. Load losses were measured at two heights: Load 1 at 1.880 in. and load 2 at 1.200 in. Test results are summarized in Table 3. Based on failure and load-loss results, valve-spring fatigue life produced by high-speed stress relief is as good or better than that produced by the conventional process.
Results of x-ray diffraction measurement comparison of automotive tension-belt springs (Fig. 6) illustrate that essentially the same amount of stress relief is achieved using both the conventional and the high-speed process.
Although soaking times and temperatures are generally fixed by the steel chemistry, substantial reductions in process times can be achieved by accelerating the heat-up time by designing more efficient heat transfer between the heated atmosphere and the load by using high-speed convective, turbulent flow which also provides for significant improvements in temperature uniformity throughout the load. The use of a heat transfer simulation (Pyrograph) software facilitates the process design process. A new batch try-out oven can successfully model any in-line continuous process. Substantial process design efficiency and property improvements that use less floor space and provide greater productivity are possible using a high-speed stress relief or tempering process.
Auto-engine valve springs can be stress relieved using a high-speed stress relieving process in a Pyro oven, and various high-speed tempering processes have been develop including automotive axles, hypoid gears and CV joints. For example, using rapid stress-relieving technology, the total stress relieving time for 16-mm (about 5/8 in.) diameter CrSi wire can be reduced to 10 min or less.
- N. Fricker, K.F. Pomfret, and J.D. Waddington, Commun. 1072, Inst. Of Gas Engrg., 44th Annu. Mtg., London, 1978
- G.E. Totten, G.R. Garsombke, D. Pye and R.W. Reynoldson, :Heat Treating Equipment, Steel Heat Treating Handbook, eds. G.E. Totten and M.A.H. Howes, Marcel Dekker Inc., New York, p 293-481, 1997
- M. Grenier and R. Gingras, High Speed Stress Relief, Proc. SMI Tech. Symp., Chicago, p 125-128, 1999
- M. Grenier and R. Gingras, Advances in High Speed Stress Relief, Proc. SMI Tech. Symp., Chicago, p 100-103, 2001
- M. Grenier, High-Speed, High-Precision Stress Relieving, Springs, Vol. 41, No. 5, p 68-71, Oct. 2002
- B. Liscic, Steel Heat Treatment Handbook , Steel Heat Treatment, Eds. G.E. Totten and M.A.H. Howes, Marcel Dekker Inc., New York, p 527-662, 1997
- A.K. Sinha, Ferrous Physical Metallurgy, Basic Heat Treatment, Butterworths, Boston, p 403-440, 1989
- M.A. Grossmann and E.C. Bain, Principles of Heat Treatment, Tempering After Quench Hardening, ASM, p 129-175, 1964
- G. Krauss, Steels: Heat Treatment and Processing Principles, Tempering of Steel, ASM Intl., p 206-261, 1990
- K. Kasten, A Primer of Terminology for Heat Treat Customers: Heat Treating, p 32-39, Feb. 1980
- E. Tarney, Heat Treatment of Tool Steels, Tooling & Production, p 102-104, May 2000
- K-E. Thelning, Steel and Its Heat Treatment (2nd Ed.) Heat Treatment - General, Butterworths, London, p 207-318, 1984
- D.J. Naylor and W.T. Cook, Heat Treated Engineering Steels, Mat. Sci. & Tech., Vol. 7, p 435-488, 1992
- H.J. Spies, G. Munch, and A. Prewetz, M¿glichkeiten der Optimierung der Auswahl vergutbarer Baustahle durch Berechnung der Hart-und-vergutbarkeit, Neue Hutte, Vol. 8, No. 22, p 443-445, 1977