Behind the innovation: The road to rewriting current sensing

Key takeaways:

• The TMCS2100 is TI's first multiaxial coreless Hall-effect current sensor for traction inverters in electric vehicles, eliminating the need for magnetic cores without sacrificing current measurement accuracy.
• Unlike conventional coreless current sensors that measure the magnetic field surrounding the current in one direction, the TMCS2100 measures it in two directions, making it less sensitive to vibration and more tolerant of physical displacement.
• The new multiaxial coreless Hall-effect current sensor addresses a long-standing trade-off in traction inverter design: achieving coreless, precise current sensing that reduces crosstalk between motor phases and torque ripple, contributing to better driving range and a smoother ride.
• Developed through a collaboration between TI's Kilby Labs and TI's sensing team, the TMCS2100 enables automakers to achieve the precision they need without size and weight trade-offs, while fitting into smaller, more efficient traction inverters.

For years, the same problem perplexed – and frustrated – thousands of engineers designing traction inverters for electric vehicles.

In traction inverters, achieving precise motor control requires precise current measurement. That has long meant confronting a fundamental trade-off; while a magnetic core can provide the necessary current measurement accuracy, it adds weight and size to every design. Remove the magnetic core, and vibration alone could compromise the current readings.

TI’s engineers didn’t want to accept the trade-off as the final answer; instead, they turned it into an invitation to innovate.

“Our customers expressed a clear need for a current-sensing solution in traction inverters that was coreless and accurate,” said Jason Cole, vice president and general manager of Sensing Products at TI. To develop the solution, TI’s sensing business team partnered with Kilby Labs, TI’s advanced R&D engine that collaborates with TI’s technology businesses to tackle engineering challenges that don’t yet have clear answers.

Turning a technological challenge into a solution

TI began working with two automotive engineers already facing the traction inverter trade-off, bringing actual design constraints into the project from the outset. The designers’ primary question became TI’s starting point: Could a coreless solution deliver the accuracy that traction inverters must have, without the compromises engineers had long accepted?

“Ideally, you want to have a few key collaborators who can work closely with you,” said Lei Ding, systems manager at Kilby Labs. “Then you can have a small team that quickly prototypes and evaluates the concept, and evolves it.”

The collaboration shaped the initial direction of the project. Early concepts relied on an external microcontroller to correct measurement errors after the fact. But discussions between Kilby Labs and the sensing team pointed toward a more ambitious goal: a sensor that could correct measurement errors internally.

From there, the project took a different path. Rather than refining their first design, TI’s engineers began rethinking how a sensor could measure current in a traction inverter.

The team soon delivered its prototype to a customer’s lab in Germany, where they spent three days testing it under conditions that the sensor would actually face: different current levels, different frequencies, different modulations.

Sometimes, these tests led to more questions. For example, how would the tests hold up in a real system? But interestingly, uncertainty became part of the process rather than a warning sign.

“Instead of looking for easy examples that would prove an idea works, we took the opposite approach: looking for the cases that would show it would fail,” said Dok Won Lee, process development engineer manager at Kilby Labs. “That gives you an answer quickly.”

One question kept resurfacing. Existing coreless sensors measure the magnetic field surrounding the current in one direction. What if a sensor measured it in two?

A smarter sensor, built together

Until now, no solution on the market could execute complex current sensing and account for the multi-axis nature of a signal. TI engineers knew that if they wanted to reduce unintended influences that would compromise accuracy, they needed to adopt a multiaxial sensor design.

A multiaxial sensor design avoids the industry’s long-standing trade-off. By reading the magnetic field in two directions, the sensor becomes less sensitive to vibration. Multiaxial measurement capability also provides more accurate and reliable magnetic sensing, increasing the device’s ability to tolerate physical displacement. Better precision reduces crosstalk between motor phases and torque ripple, the jerky behavior that affects both an electric vehicle’s driving range and how smoothly it rides.

“It felt very rewarding, because this was not a simple problem,” Dok Won said. “It was high risk, and high reward. There were a lot of uncertainties that could have led to a dead end. But we were able to lead into productization.”

The multiaxial sensor eventually became a real product: the TMCS2100 multiaxial coreless Hall-effect current sensor for traction inverters.

“This is just a piece of the innovation,” said Jerry Shi, general manager of HEV/EV and powertrain at TI. “Our customers will be able to take it even one step further. When placed into vehicles, this device will help automakers achieve the precision they need without size and weight trade-offs, while fitting into smaller, more efficient traction inverters.”

The development of the TMCS2100 may have taken many directions amid several cycles of reviews, refinement and readjustment. But the journey ended in one place.

“We’re engineers inventing for engineers. Our goal was to understand our customer’s problem and cocreate a solution they didn’t know was even possible,” Jason said. “This technology embodies that.”