The ProblemA manufacturer of AGMA Class-3 gears wanted to switch raw material from bar stock to forgings. During qualification trials, normal inspection practices using a test piece (an actual gear segment) produced from bar stock passed case depth and hardness (surface and core) requirements. However, validation testing on an actual part (produced from a forging) was found to have a variation in hardness in the root of the gear teeth. It was necessary to determine the root cause of the problem and to try to design a salvage cycle for the parts already produced.
The InvestigationThe investigation into the cause of the root-hardness variation in these gears began by gathering information about the raw material as well as the heat-treatment practices. Testing followed on both gears made from bar stock and from forgings (Fig. 1). Both parts had been carburized to a targeted effective case depth of 0.070 inch (flank) and 0.035 inch minimum (root).
The material certification sheets for the bar stock and forgings indicated that although the chemistry was within specification limits for an SAE 8822 material, the forgings had lower hardenability values (J4 and J8 positions). A comparison of the carburizing profiles by microhardness methods supported this finding, as the bar stock responded very differently than the forgings.
Several “dips” in hardness were consistently observed in the forgings (between the fourth and sixth indentations and between the ninth and 11th position) and are considered abnormal. Typically, this is an indicator of a significant change in the microstructure. However, the effective case depth (50 HRC) of both samples appeared to be essentially the same. This suggested additional testing be conducted in the root area of the forged gears (Fig. 2).
These results indicated that the effective case depth (50 HRC) of the forging is very dependent on the location sampled. It should be noted that the selection of the test location was random and done on a polished (not etched) sample. The additional test locations were adjacent to the original selected test zone. Focusing in on a depth of 0.030-0.040 inch (Fig. 3) provided more evidence.
First Bainite TestIn order to help determine the severity of the problem, parts were water quenched in order to obtain maximum case hardness in accordance with SAE J1975 (Case Hardenability of Carburized Steels). “Distance to the first appearance of bainite” (DFB) is defined as the location on water quenching where the hardness just begins to drop off.
For the greatest resistance to impact, data suggest that the as-quenched microstructure must be substantially free from bainite or pearlite. The presence of very small amounts of bainite in the case has also been reported to reduce fatigue resistance.
A detailed study of the occurrence of bainite in carburized end-quench hardenability specimens involved some 81 alloys carburized at 925°C (1700°F), cooled to 845°C (1550°F) and end-quenched. Case and hardness profiles were determined at a depth corresponding to 0.90% carbon. The amount of bainite as a function of distance from the quenched end of the bar was determined (Fig. 4). It should be noted that one cannot detect the presence of small amounts of bainite from hardness data alone.
Multiple regression analysis developed an empirical relationship for predicting DFB from composition (Equations 1a, 1b). The regression equations appear below, and they are valid at the 0.9% C level in the case for steels containing 0.5-1.1% Mn, 0-1.5% Ni, 0-1.0% Cr, and 0-0.5% Mo. Alloy contents are entered in weight percent:
(1a) DFB (in millimeters from the quenched end) = 54.79(Mo2) + 6.4(Cr2) – 76.1(MnNi) + 118.8 (MnMoNi) + 106.1 (MnMoCr) +15.5(MnNiCr) + 52.9(MoNiCr) + 1.18
(1b) DFB (in sixteenths of an inch from the quenched end) = 35.4(Mo2) + 4.0(Cr2) – 49.7(MnNi) + 74.8 (MnMoNi) + 66.9(MnMoCr) + 9.8(MnNiCr) + 33.3(MoNiCr) + 0.7
Alloy interactions influence the presence of bainite in the carburized case and come into play when modifying a carburizing-steel composition. The regression equations provide a convenient method of predicting the effect of changes in composition on DFB. They can also aid in the establishment of a minimum alloy content to assure a bainite-free microstructure in the carburized case.
Test ResultsTesting also revealed that the flank achieved a consistent effective case depth of 0.070-0.080 inch, while the root varied from 0.030-0.036 inch. The first bainite test (Table 1) indicated a quenching problem with the forgings.
Oil-quenching tests were performed on similar gear components manufactured from SAE 43B17 to view the impact of a material change on hardenability response (Fig. 5). This investigation was triggered by the discovery that some of the 8822 forgings had boron additions. The profiles found are very similar in slope, but the point of first bainite has been moved to about 0.045-inch depth in the 43B17 material.
By comparing oil and water quenching, the increased hardenability of 43B17 (Fig. 6) is not as drastic (as when water quenched) but does increase the reliability of the heat treatment and was a good alternative material choice to consider moving forward. First bainite in a boron steel can depend on carbon content, which must be taken into consideration.[5,6]