“Established standards for transmission and driveline NVH are not sufficient for EV and hybrid applications,” says Mark Findlay, Managing Director of DSD (Drive System Design), a U.K.-based automotive engineering consultancy.
There is no excuse for gear noise in either EVs or hybrids, Findlay asserted in an interview with Automotive Engineering. Customer expectations of required levels of cabin refinement do not allow unseemly sounds rising from the “whine cellar” in any production vehicle, whether premium or budget.
Because the background cabin noise in an EV, or a hybrid operating in electric mode, is significantly lower than that of a vehicle with an internal combustion engine, transmission noise levels that were acceptable in conventional vehicles are no longer satisfactory. The issue is further complicated because electric motors operate at much higher speeds and produce higher-frequency tones that are more noticeable to vehicle occupants and can become wearing on longer journeys.
EV/hybrid NVH issues may originate as a result of the gear geometry, the design of the transmission housings or supporting structures, or even the mounting bushings or surrounding sub-frame, explained Findlay. “It would therefore be a mistake to suggest that it can be solved solely through optimization of gear micro geometry, or any other singular approach," he said.
Pitfalls to be avoided
At a component level, the potential contribution to NVH from the gear design is judged predominantly by the peak-to-peak transmission error (TE)—the departure from uniform relative angular motion of a pair of meshing gears. Findlay noted that this is influenced by "all deviations from an ideal tooth form, the conditions under which the gears operate and individual component stiffness in the system."
However, the most common factor is how the gear mesh operates under load. Its behavior is subject to the instantaneus deflections of all the elements in the system, including shafts, bearings, plus the housing which carries the bearings. This is why the system approach—considering every element of the drivetrain as a potential contributor to NVH—is so important to achieving an effective solution, he said.
Detailing the difficulties of creating an acceptably quiet EV drivetrain, Findlay added that at a system level, the vibration caused by the TE acts as the key source of excitation for the dynamic response of that drivetrain. Effectively acting as an amplifier or loudspeaker, the design of the transmission housing and supporting structures dictate the way in which the TE is translated into a vibrational response and subsequent noise.
There are pitfalls to be avoided if satisfactory production levels of NVH are to be achieved by designers and engineers.
“A transmission housing designed with only structural requirements in mind may act to amplify the excitation from TE that is otherwise within targets, creating unacceptable noise," Findlay explained, noting that the effect can be further magnified if the meshing frequency of the gears aligns with a natural frequency of the housing or supporting structures, resulting in resonance and increased noise.
He describes DSD as “one of the pioneers” of this approach, having used whole system simulation since 2007. He said that as software tools have become more powerful, understanding both their strengths and limitations is essential: “It’s analogous to a fine violin that can only be heard at its absolute best in the hands of a classical violinist!”
Tackling prototype issues
Achieving lower NVH levels for any vehicle inevitably requires improved understanding of the underlying issues in order to isolate root causes. While hardware and software providers continually increase the power of their analysis products, it falls to specialist users such as DSD, to develop their application.
In Findlay’s view, while early analysis can filter out many potential NVH issues, due to the complexity of the systems, there may still be issues that arise in the prototype stage.
“We have developed techniques that let us get closer to identifying any issues at the concept stage but at the moment it is best used in fast response applications," he said. "There are so many dependent variables that the model generates more data than can be assessed within realistic timescales. You can generate a complete picture of the modal damping of, say, a transmission case, but it may couple with just one of many bending cases."
Success thus requires a greater focus on the specific area of interest, which inevitably means responding to an issue found in actual hardware. This approach has proven highly cost-effective and quick to provide a solution, he said. Typical investigations “only amount to tens of thousands of dollars” yet can save OEMs many times that amount. Final manufacturing costs are often unaffected and the solutions typically incur just a small tooling change, he added.
Findlay cited the example of a premium EV with a transmission noise arising from certain gear meshes. DSD developed a solution for the OEM within six weeks, which reduced the critical surface velocity of the transmission case by a factor of 48.
“The noise source was eliminated," he said. "A combination of measurement and simulation was used to establish an improved casing design and new macro and micro geometry for all the gears affected." This gear geometry is now in volume production.
As applications of electric systems continue to burgeon within automotive design and engineering, ongoing research at DSD aims to turn commonly used analysis tools into more effective design tools, able to identify and eliminate undesirable system level interactions reliably at the design stage, before hardware exists.
In the immediate future, Findlay believes that even the smallest noise concern will become unacceptable on premium electrified vehicles, while the relatively low cost of overcoming any issues could see budget EVs and hybrids quickly closing the gap.
“With the level of knowledge available today, there is no excuse for gear noise," he asserted. "The priority for engineers is to integrate the effects of statistical manufacturing tolerances into the predictive processes to ensure the robustness of all solutions under series production conditions.”