Electric Transaxle Noise, Vibration, and Harshness (NVH) Guide for EV and AGV OEMs

What NVH Actually Means for a Driven Axle

NVH stands for noise, vibration, and harshness. In a electric transaxle, these three properties decide whether an operator notices the machine at all. Noise is the audible tone a resident or worker hears. Vibration is the physical shake transferred through the frame, seat, or floor. Harshness is the subjective roughness of how the sound and feel change during acceleration. For a low-speed EV builder or an AGV integrator, NVH is not a luxury metric. It is the difference between a product that feels engineered and one that gets returned after the first week of use.

Why Electric Transaxles Get Complained About More Than Combustion Axles

A gasoline drivetrain hides a lot. The engine produces broadband masking noise that covers gear whine and bearing hum. An electric axle removes that cover. With no engine idling in the background, the AGV drivetrain vibration and high-pitched gear tones become the dominant sound signature of the whole machine. Low-speed logistics vehicles, floor scrubbers, and garden tractors all run at exactly the speed range where gear meshing orders are most audible. That is why a transaxle that passes a bench torque test can still fail a showroom demo. The mechanical function is fine; the perception is not.

Primary Noise Sources Inside an Electric Transaxle

Most complaints trace back to three sources. First, gear meshing generates a tonal order at the tooth-passing frequency. Second, rolling bearings produce broadband and defect-related tones. Third, the motor itself radiates electromagnetic noise even when the gears are perfect. Separating these during a fault hunt matters, because the fix for a gear problem is not the fix for a motor problem. A practical first step is to run the motor on a spin rig without the gear set, then with the gear set, and compare spectra.

Gear Meshing Frequency and Precision Grade

The mesh frequency equals the number of teeth multiplied by the shaft rotational speed. A 17-tooth pinion at 3,000 rpm produces a 51,000-order event every minute. The quality of that event depends on the gear grade. AGMA and ISO both publish accuracy classes; AGMA 8-9 is common for cost-sensitive utility vehicles, while AGMA 10-12 is expected when the axle sits inside a premium cabin or a quiet warehouse AGV. Tighter gear meshing quality reduces the amplitude of the sidebands around the mesh frequency, and that is what the human ear reads as “smooth” versus “whiny.” Specifying the grade in the drawing is cheaper than reworking a complaint batch.

Motor Electromagnetic Noise: Radial Force Waves and Cogging Torque

The electric machine adds its own signature. Stator slots and rotor poles interact to create radial force waves that push the housing at specific orders. When a structural mode of the housing lines up with one of those orders, the whole axle sings. Cogging torque, the reluctance ripple present even at zero current, shows up as a low-speed creep judder that operators feel at takeoff. Reducing motor harmonics EV content starts at the electromagnetic design stage: skew angle, slot-pole combination, and winding layout all shift the order content before a single part is machined.

Vibration Isolation: Mounts That Decide Comfort

Even a good axle radiates some energy. The mount system is the last line of defense. Rubber bushings are the default for utility vehicles because they are cheap and tune easily by durometer. Hydraulic mounts damp low-frequency shake better and are common in ride-on equipment where the operator sits close to the axle. Active mounts with tuned mass dampers exist but rarely pay off below premium segments. The mistake most OEMs make is treating the mount as a generic rubber puck rather than a tuned component matched to the axle’s dominant order.

How to Test NVH Before You Accept a Batch

A transaxle noise test should happen at three levels. The component bench test spins the axle under load and records sound pressure at fixed microphone positions. A semi-anechoic chamber removes room reflection so you can trust the numbers. The full-vehicle or full-machine test captures what the end user actually experiences, including frame and panel resonance. Microphone arrays help localize a tone to a specific bearing or gear cluster when a single spectrum is not enough. Skipping the full-machine step is the usual reason a “passed” axle still disappoints in the field.

Standards Buyers Should Reference in Specs

ISO 362 covers exterior pass-by noise for road vehicles and is a useful external target. SAE J1470 describes measurement of interior noise in passenger cars and gives a repeatable method for position and weighting. Most AGV and industrial-vehicle makers also keep an internal standard because their use case, a quiet warehouse or a hospital corridor, is stricter than any public regulation. The point is not to cite a standard for its own sake but to give the supplier a measurable acceptance line.

Common NVH Complaints and What They Usually Mean

Startup judder almost always points to cogging torque or a rotor imbalance felt at low speed. Acceleration whine is a gear mesh order that grows with rpm, typically a grade or alignment issue. Idle resonance, a low boom at standstill with the motor energized, is usually a mount or housing mode coupling with a motor order. Each symptom maps to a different engineering root cause, so a complaint log with rpm and load context is worth more than a vague “too loud” note from the field.

Writing NVH Acceptance Into the Procurement Contract

Do not leave NVH to “reasonable quality.” Put numbers in the purchase order. State the maximum A-weighted sound pressure at a defined rpm and load, the allowed gear mesh order amplitude, the required gear accuracy class, and the balance grade of the rotor. Define the test method and the fixture so both sides measure the same thing. A supplier can only guarantee what is written; an undocumented expectation becomes a polite argument at payment time.

Tip: Require a signed NVH test report at three load points (no-load, half-load, rated-load) before the final 20% of payment is released, so tonal defects surface before they reach your customers.

Rotor Balance and Mechanical Tolerances

Imbalance is the silent contributor behind many “mystery” vibrations. A rotor balanced to G6.3 may be fine for a fan but marginal for an axle mounted near a seated operator. Moving to G2.5 or G1.0 changes the felt shake more than buyers expect. Bearing seats, shaft runout, and housing bore concentricity also set the floor for achievable NVH. Tightening these mechanical tolerances upstream is far cheaper than adding foam and dampers downstream to cover a sloppy base build.

NVH Problems That Trigger Recalls and Warranty

Field returns tied to NVH are rarely about absolute loudness. They are about a tone that changes over time, a bearing that starts to sing after 200 hours, or a mount that hardens in the cold and transmits a boom. These are wear and material issues, not design intent, and they are expensive because the product is already installed. A simple accelerated endurance spin on a sample batch catches most of them for a fraction of a recall cost. Treat NVH as a reliability signal, not just a comfort number.

Frequently Asked Questions

What NVH level is acceptable for a warehouse AGV?

Most warehouse operators target below 65 dBA at one meter during travel, but the stricter requirement is usually the absence of a pure tonal whine rather than the overall level. A flat broadband sound is tolerated far better than a sharp order tone at a fixed pitch.

Can a quiet gear set fix motor electromagnetic noise?

No. Gear improvements reduce mesh orders only. If the dominant complaint is a hum that exists even when the axle is spun without load, the source is the motor’s radial force waves or drive switching frequency, and the fix belongs in the electromagnetic or inverter design.

How much does gear accuracy grade affect cost?

Moving from AGMA 8 to AGMA 11 can add 15 to 40 percent to gear cost depending on volume and material, but it often removes the need for secondary damping. For cabin-adjacent or indoor applications, the upgrade usually pays for itself in fewer complaints.

Is a semi-anechoic chamber necessary for incoming inspection?

Not always. A controlled bench with a fixed microphone distance and a reference fixture gives repeatable comparison between batches. A full chamber matters more for design validation; for incoming QC, consistency of method beats absolute lab grade.

Why does the same axle sound different in two machines?

Because the frame, panels, and mount points create their own resonance. An axle that is quiet on a rigid test cart can boom on a thin stamped chassis. Always validate on the actual production frame, not just the supplier’s rig.

Should NVH be tested hot or cold?

Both. Rubber mounts and lubricants change character with temperature. Cold-start whine and a hardened mount boom are common field complaints that a warm-only test will never reveal.

What balance grade should an EV transaxle rotor meet?

For operator-proximate utility and AGV axles, G2.5 is a sensible starting target, with G1.0 for premium or seated-operator machines. The choice depends on how close the rotor sits to the person judging the comfort.

Post time: Sep-09-2026