In this article, the long-term perspective of distortion capability is studied; distortion data from the start of production in 2011 was compared with distortion data from 2018; 25 consecutive loads were analyzed in each case.

volkControlling distortion during the case-hardening process is of key importance when manufacturing gears. By effective control of distortion and the variation of distortion, significant costs in post-heat-treatment machining processes can be avoided. Especially for e-drive gears such as internal ring gears or final drive ring gears, significant cost-savings can be achieved.

If distortion is controlled in such a manner that ISO class 8 is guaranteed after case hardening, the grinding operation gets obsolete, and parts may be honed only.

The combination of low-pressure carburizing (LPC) and high-pressure gas quenching (HPGQ) offers the potential to provide better control of distortion compared to other process-combinations such as atmospheric carburizing with oil quenching.

This article analyzes distortion values of gear components from a planetary set of a six-speed automatic transmission over a long period of time. The gears were analyzed in terms of circularity, helix average and helix variation. It is demonstrated that distortion data stays stable and predictable even over a long period of time when applying optimized heat-treatment process parameters and if the process steps in the manufacturing chain before heat treatment are frozen and robust.

When combining this heat-treatment technology with appropriate geometrical inspection, this will result in guaranteed ISO class 8 geometry after heat treatment.

The article gives directions on how this goal can be achieved by combining an advanced heat-treatment process with advanced gear-inspection technology.

1 Introduction

Heat-treatment distortion has a strong cost impact, since distorted components need to be hard- machined after heat treatment. Better control of distortion means:

  • Less cycle time per part in hard machining.
  • Less hard-machining capacity needed.
  • Less tooling cost for hard machining.

For some applications, hard machining can be completely eliminated with an excellent control of distortion.

With the introduction of E-mobility, proper distortion control has become even more important than it has previously. Distorted gear components cause noise in the transmission. Especially battery-operated electric vehicles (BEVs) and all other electrified vehicles (including hybrids) require a low-noise transmission with high-precision components. E-drive components typically require a grinding operation to correct heat-treat distortion, which is followed by a honing operation to achieve the proper surface finish of the gear teeth. Sufficient distortion control can eliminate the need for the grinding operation.

The components that were evaluated in this study are assembled in passenger cars with conventional combustion engines. The level of distortion control allows to skip all hard machining operations so that the gears go directly after heat treat into the transmission assembly.]

2 Distortion mechanisms and high-pressure gas quenching (HPGQ)

The relevant mechanisms that cause distortion of components during heat treatment have been described extensively in literature [Hee17]. Walton [Wal92] published the numerous potential factors that are influencing distortion in more detail. (See Figure 1.)

Figure 1: Potential factors influencing distortion [Wal92].

By applying the technology of low-pressure carburizing (LPC) and high-pressure gas quenching (HPGQ), heat-treat distortion can be significantly reduced. LPC is a case-hardening process that is performed in a pressure of only a few millibars using acetylene as the carbon source in most cases. During HPGQ, the load is quenched using an inert gas-stream instead of a liquid quenching media. Usually, nitrogen or helium are used as quench gas [Loe05, Heu15].

HPGQ offers a tremendous potential to reduce heat-treat distortion. Conventional quenching technologies — such as oil- or polymer-quenching — exhibit inhomogeneous cooling conditions. Three different mechanisms occur during conventional liquid quenching: film-boiling, bubble-boiling, and convection. Resulting from these three mechanisms, the distribution of the local heat-transfer coefficients on the surface of the component is very inhomogeneous. These inhomogeneous cooling conditions cause tremendous thermal and transformation stresses in the component and subsequent distortion. During HPGQ, only convection takes place, which results in much more homogenous cooling conditions [Heu10].

Significant reductions of distortion by substituting oil-quench with HPGQ have been published [Alt05]. Another advantage of HPGQ is the possibility to adjust the quench-intensity exactly to the needed severity by choosing quench-pressure and quench-velocity. Typical quench pressures range from 2 bar to 20 bar. The gas velocity is controlled by a frequency converter. Typical gas velocities range from 2 m/s to 20 m/s depending on the part geometry and the steel grade of the component. If the HPGQ chamber is equipped with a gas-flow reverse system, the gas-flow direction can be alternated during the quench between “top to bottom” and “bottom to top.” This alternating flow direction results in more uniform cooling curves within each part and within the load from part to part. Therefore, further reductions in height and variation of distortion can be achieved.

Figure 2 shows a typical industrial system for the HPGQ process.

Figure 2: High-pressure gas quenching (HPGQ) chamber of a heat-treatment system.
Figure 3: Comparison of roundness after atmospheric carburizing and oil quenching vs. LPC and HPGQ.

A comparison between the roundness of gear shafts after traditional heat treatment with atmospheric carburizing plus oil quenching vs. LPC plus HPGQ is given in Figure 3. Switching to LPC and HPGQ led to a significant reduction in distortion. As a consequence, the fall-out rate was reduced from 0.124% to 0.004% for this component.

3 Process chain of manufactured components

3.1 Components

In this article, the distortion control of three internal ring gears from a 6-speed automatic transmission is analyzed. Figure 4 shows these internal ring gears from a 6-speed automatic transmission.

Figure 4: Illustrations of FD internal ring gear, output internal ring gear; reaction internal ring gear (from left to right).

The process of LPC using acetylene as carbon-source and HPGQ using helium as a quench gas followed by tempering is applied in serial production for these gears.

They are made of 5130H material, and their dimensions can be found in Table 1. The case-hardening depth CHD after heat treat is specified as 0.3 … 0.6 mm and surface hardness is specified as 79 … 83 HRA (equivalent to 56 … 63HRC). The applied LPC process provides surface hardness values between 59 and 61 HRC.

Table 1: Dimensions of internal ring gears.

Design and manufacturing specifications for these components have not changed over the years after start of production.

3.2 Process chain of manufacturing

Figure 5 shows the process chain of the manufactured components. Each process step except for washing may have an effect on the distortion potential of the internal ring gears. Therefore, it is important to first reduce these potentials to the possible minimum before start of production — namely to reduce residual stresses and material inhomogeneities of the blanks. Then, with the start of production, the process chain should be a frozen process, and parameters need to be kept stable.

Figure 5: Process chain of manufacturing of internal ring gears.

Indeed, the process chain was not changed after start-of-production for these gears. The LPC-heat-treatment process has neither been changed over the years after start of production.

4 Measurement of distortion

Since start of production in 2011, the geometry of the gears is being controlled before and after heat treatment using a CNC analytical-gear-checker. Three samples are pre-measured per heat-treat load. One sample is placed in the top corner, one sample in the middle-middle, and one in the bottom-middle of the load. These positions were chosen to cover the “extreme” positions within the load. Figure 6 shows the set-up of a typical load. These pre-measured samples are being measured after heat treat again to obtain the distortion values. Control limits are calculated to establish the distortion control limits.

Figure 6: Load of internal ring gears.

The gears are analyzed for circularity, helix average left flank, helix average right flank, helix variation left flank, and helix variation right flank. For each sample, four equally spaced teeth are inspected. Both left flank and right flank are examined per tooth.

Helix average measurement traces the four equally spaced teeth at the pitch diameter, then averages the four readings. Helix variation is the range of the four readings.

Circularity is measured by probing all teeth on the pitch diameter at the mid-point of the face width.

The range within the three samples per load represents the within-load variation. The load-to-load variation is obtained when comparing the average value of the three samples from one load with the average value of three samples from other loads.

All values are stored in a quality-database, which is part of the CAQ-system (computer-aided quality  system). The distortion data is stored together with load-number, production-traveler-number, and goods- receiving number. This allows the correlation of distortion values with the chemical composition of the steel and with the metallurgical data (microstructures and hardness-data) after heat treat.

5 Analysis of long-term distortion capability

5.1 FD Internal Ring gear

Figure 7 shows a comparison of the circularity of 25 consecutive loads from the start of production in 2011 and 25 consecutive loads from seven years later in 2018. The upper graph shows the circularity of the sample from top-corner, the middle graph shows the values from the middle-middle, and the lower graph from the bottom-middle of the load. No significant differences could be found between 2011 and 2018, which shows that distortion is stable and predictable over such a long period of time.

Figure 7: FD internal ring gears; circularity of 25 consecutive loads each from the year 2011 and 2018 in three different positions within the load (USL=upper specification limit and LSL=lower specification limit).

Figures 8 to 11 show comparisons of 25 consecutive loads from the start of production in 2011 and 25 consecutive loads from seven years later in 2018 regarding helix values. Each upper graph shows the helix values of the sample from top-corner, each middle graph shows the values from the middle-middle and each lower graph from the bottom-middle of the load. No significant differences could be found, which shows that distortion is stable and predictable regarding helix values both in average and in variation.

Figure 8: FD internal ring gears; “helix average (FHb) left flank” of 25 consecutive loads each from the year 2011 and 2018 in three different positions within the load (USL=upper spec. limit and LSL=lower spec. limit).
Figure 9: FD internal ring gear; “helix average (FHb) right flank” of 25 consecutive loads each from the year 2011 and 2018 in three different positions within the load (USL=upper spec. limit and LSL=lower spec. limit).
Figure 10: FD internal ring gear; “helix variation (Vbf) left flank” of 25 consecutive loads each from the year 2011 and 2018 in three different positions within the load (USL=upper spec. limit and LSL=lower spec. limit).
Figure 11: FD internal ring gear; “helix variation (Vbf) right flank” of 25 consecutive loads each from the year 2011 and 2018 in three different positions within the load (USL=upper spec. Limit and LSL=lower spec. limit).

5.2 Output Internal gear

Figure 12 shows a comparison of 25 consecutive loads from the start of production in 2011 and 25 consecutive loads from seven years later in 2018. Again, the upper graph shows the circularity of the sample from top-corner, the middle graph shows the values from the middle-middle, and the lower graph from the bottom-middle of the load. No significant differences could be found, which shows that distortion is stable and predictable.

Figure 12: Output internal gear; circularity of 25 consecutive loads each from the year 2011 and 2018 in three different positions within the load (USL=upper spec. limit and LSL=lower spec. limit).

Figures 13 to 16 show comparisons of 25 consecutive loads from the start of production in 2011 and 25 consecutive loads from seven years later in 2018 regarding helix values. Each upper graph shows the helix values of the sample from top-corner; each middle graph shows the values from the middle-middle,  and each lower graph from the bottom-middle of the load. No significant differences could be found, which shows that distortion is stable and predictable regarding helix values both in average and in variation.

Figure 13: Output internal gear; “helix average (FHb) left flank” of 25 consecutive loads each from the year 2011 and 2018 in three different positions within the load (USL=upper spec. limit and LSL=lower spec. limit).
Figure 14: Output internal gear; “helix average (FHb) right flank” of 25 consecutive loads each from the year 2011 and 2018 in three different positions within the load (USL=upper spec. limit and LSL=lower spec. limit).
Figure 15: Output internal gear; “helix variation (Vbf) left flank” of 25 consecutive loads each from the year 2011 and 2018 in three different positions within the load (USL=upper spec. limit and LSL=lower spec. limit).
Figure 16: Output internal gear; “helix variation (Vbf) right flank” of 25 consecutive loads each from the year 2011 and 2018 in three different positions within the load (USL=upper spec. limit and LSL=lower spec. limit).

5.3 Reaction Internal gear

Figure 17 shows a comparison of 25 consecutive loads from the start of production in 2011 and 25 consecutive loads from seven years later in 2018. Again, no significant differences could be found, which shows that distortion is stable and predictable.

Figure 17: Reaction internal gear; circularity of 25 consecutive loads each from the year 2011 and 2018 in three different positions within the load (USL=upper spec. limit and LSL=lower spec. limit).

Figures 18 to 21 show comparisons of 25 consecutive loads from the start of production in 2011 and 25 consecutive loads from seven years later in 2018 regarding helix values. Each upper graph shows the helix values of the sample from top-corner; each middle graph shows the values from the middle-middle, and each lower graph from the bottom-middle of the load. No significant differences could be found, which shows that distortion is stable and predictable regarding helix values both in average and in variation.]

Figure 18: Reaction internal gear; “helix average (FHb) left flank” of 25 consecutive loads each from the year 2011 and 2018 in three different positions within the load (USL=upper spec. limit and LSL=lower spec. limit).
Figure 19: Reaction internal gear; “helix average (FHb) right flank” of 25 consecutive loads each from the year 2011 and 2018 in three different positions within the load (USL=upper spec. limit and LSL=lower spec. limit).
Figure 20: Reaction internal gear; “helix variation (Vbf) left flank” of 25 consecutive loads each from the year 2011 and 2018 in three different positions within the load (USL=upper spec. limit and LSL=lower spec. limit).
Figure 21: Reaction internal gear; “helix variation (Vbf) right flank” of 25 consecutive loads each from the year 2011 and 2018 in three different positions within the load (USL=upper spec. limit and LSL=lower spec. limit).

6 ISO class 8 after case hardening

If heat-treat distortion is controlled in such manner that ISO class 8 is guaranteed after case hardening, the grinding operation gets obsolete, and parts may be honed only.

Substantial cost savings are achieved when the grinding operation is eliminated.

As shown in the above, the LPC and HPGQ process was successfully applied on internal ring gears, achieving stable and predictable distortion values over a long period of time. For the examples described earlier, the circularity tolerance is currently 200 microns (=tolerance required for transmissions used with internal combustion engines). If these gears had a class 8 requirement, the circularity tolerance would be 78 microns. For output internal gears, class 8 is already met. However, for FD internal gears and reaction internal gears, some values are still worse than ISO class 8.

Further progress in distortion control has been made in recent years with the continuous improvement of:

  • Reversing quenching (changing gas flow direction).
  • Dynamic quenching (varied quench intensity over time).
  • Improved fixture design.

When applying these improvements, most gears from all three types described will meet class 8 (max. 78 microns circularity) with only very few flyers outside class 8.

To clearly identify those flyers, the optimized LPC and HPGQ process should be combined with appropriate geometrical inspection to achieve guaranteed ISO class 8 geometry after heat treatment.

The so called “ISO class 8 after case hardening approach” is described in the following:

The gear manufacturer outsources heat treatment to a commercial heat-treat service provider who is capable of geometrical inspections. First, the process chain of manufacturing and the heat-treatment process is optimized in a joint effort between gear manufacturer and heat-treat service provider. Then, before start of production, a statistical analysis is performed from as many loads as possible to determine the initial fallout rate. Only if this fallout rate is small enough, “ISO class 8 after case hardening approach” becomes feasible from an economic perspective.

During regular production, the heat treater sorts 100% of the treated parts using a roll-checker. Only good parts (= geometrically checked class 8 – parts) are returned to the gear manufacturer. The gear manufacturer pays only for the heat treatment of good parts, not for the heat treatment of parts outside class 8.

If, for some reason, the initial fallout rate is exceeded during regular production, a joint root cause analysis is started to fix the root cause.

The benefits of this “ISO class 8 after case hardening approach” for the gear manufacturer are as such:

  • 100 percent good geometrically inspected parts are received.
  • No need to measure and sort parts inhouse before honing.
  • Substantial cost savings since grinding operations are eliminated.

7 Summary

The LPC and HPGQ process was successfully applied on internal ring gears for a 6-speed automatic transmission. From each production load, three samples are geometrically inspected, and the distortion data is stored in a quality database.

In this article, the long-term perspective of distortion capability was studied. Distortion data from the start of production in 2011 was compared with distortion data from 2018. Twenty-five consecutive loads were analyzed in each case.

The distortion values proof to be stable and predictable. This means that the process chain of manufacturing including steel-making, casting, forming, and green machining and the heat-treatment process were successfully frozen and maintained stable over a long period of time.

In addition, a new approach is outlined in this article to guarantee ISO class 8 after case hardening. Substantial cost savings are achieved when the grinding operation is eliminated, and parts may be honed    only. The grinding operation gets obsolete, and parts may be honed only if ISO class 8 is guaranteed after case hardening.

This new approach combines low distortion case hardening with 100% sorting of the gears after heat treatment.

The gear manufacturer benefits from this “ISO class 8 after case hardening approach” since:

  • 100 percent good geometrically inspected parts are received.
  • There is no need to measure and sort parts inhouse before honing.
  • Substantial cost savings are achieved since grinding operations are eliminated. 

References

  1. [Alt05]: Altena H., Schrank, F. und Jasienski, W.: Reduzierung der Formänderung von Getriebeteilen in Gasaufkohlungs-Durchstoßanlagen durch Hochdruck-Gasabschreckung. In: HTM 60 (2005)1, S. 43-50
  2. [Hee17]: Heeß Karl et al: Maß- und Formänderungen infolge Wärmebehandlung von Stählen, 5. vollständig überarbeitete Auflage Expert Verlag 2017
  3. [Heu10]: Heuer, V.; Loeser, K.; Faron, D.R.; Bolton, D.: Low distortion heat treatment of transmission components; AGMA Technical Paper 10FTM04, 2010. ISBN 978-1-55589-979-0
  4. [Heu15]: Heuer, V., Loeser, K. Schmitt, G.: Improved Materials and Enhanced Fatigue Resistance for Gear Components; AGMA Technical Paper 15FTM02, 2015. ISBN 978-1-55589-008-7
  5. [Loe05]: Loeser, K.; Heuer, V.; Schmitt, G.: Auswahl geeigneter Abschreckparameter für die Gasabschreckung von Bauteilen aus verschiedenen Einsatzstählen. In: HTM 60 (2005)-4, page 248- 254
  6. [Wal92]: Walton, H.: Dimensional changes during hardening and tempering of through-hardened bearing steels. Quenching and Distortion control (Conference Proceedings). ASM International, 1992, S. 265-273