Key Takeaway: Texas Instruments has unveiled a single-chip motor control solution capable of running 6 simultaneous FOC loops with just 500-nanosecond closed-loop latency — eliminating the need for multiple MCUs in humanoid robots and industrial automation systems. The solution integrates motor control, communication interfaces, and functional safety (SIL3) on one chip, reducing wiring complexity by over 50% and enabling 24+ axes of coordinated motion from a single processor.
The robotics and industrial automation industries are undergoing a fundamental shift in how motor control systems are designed. At the TI Mobility & Robotics Seminar 2026 held on July 9 in Seoul, Texas Instruments Korea Director Heo Jeong-hyeok unveiled a real-time control technology strategy that centers on three breakthrough capabilities: 500-nanosecond ultra-low latency computation, single-chip multi-axis control, and single-pair Ethernet communication architecture. This article breaks down what these advances mean for factory automation engineers, motion control designers, and anyone building the next generation of robotic systems.
Why 500-Nanosecond Latency Changes Everything
To appreciate the significance of TI’s announcement, you need to understand the closed-loop delay problem in motor control. In a typical field-oriented control (FOC) system, the processor must read current sensors via an ADC, perform Clarke and Park transforms, run a PID controller, compute space-vector PWM, and output new duty cycles — all within one switching cycle.
Traditional microcontrollers achieve closed-loop latencies in the range of 5 to 50 microseconds. While this works for simple two-axis systems, it becomes a bottleneck when coordinating dozens of axes simultaneously. Consider a humanoid robot: a single arm requires 7 axes (3 shoulder, 1 elbow, 3 wrist), and the full body can demand 24 or more coordinated axes. At 50-microsecond latency per axis, the cumulative delay across a 24-axis system introduces noticeable lag and vibration.
TI’s solution compresses this entire pipeline — from ADC sensing sampling to PWM control output — into a single closed loop controllable within 1 microsecond, with core computation achieving 500-nanosecond latency. This is accomplished through a combination of hardware-level optimizations:
- 125-nanosecond ADC IP: On-chip analog-to-digital converters with industry-leading conversion speed eliminate the primary bottleneck in the feedback loop.
- Built-in Trigonometric Math Unit (TMU): Reduces traditional trigonometric operations from over 100 clock cycles to just 7 clock cycles — a 14x improvement that directly accelerates Clarke/Park transforms.
- Single-chip 6-FOC execution: The high-end MCU architecture (such as the AM26 series and C2000 DSP family) can simultaneously run six complete FOC control loops on a single chip.
- Real-time AI acceleration core: An embedded neural processing unit handles time-series data such as current, voltage, and temperature for inverter efficiency optimization and predictive maintenance.
For factory automation engineers, this means smoother multi-axis motion, reduced vibration, and higher throughput without the complexity of distributing control across multiple processors.
Single-Chip Multi-Axis: From 3 Chips to 1
The current industry standard for controlling a 7-axis robotic arm typically requires 3 to 7 separate microcontrollers — one for the shoulder cluster, one for the elbow, one for the wrist, and additional ICs for safety monitoring. This multi-chip approach introduces several problems:
- Synchronization overhead: Each chip must communicate with others via SPI, CAN, or Ethernet, adding latency at every handoff.
- Wiring complexity: More chips means more PCBs, more connectors, and more potential failure points.
- Power consumption: Multiple MCUs draw more aggregate power than a single optimized processor.
- Cost: Each additional chip adds BOM cost, assembly cost, and testing cost.
TI’s single-chip approach consolidates all of this. The AM26 series MCU can simultaneously execute 6 FOC control loops, handle communication protocols, manage safety monitoring, and process AI inference — all on a single die. The practical impact is significant:
For a humanoid robot with 24+ axes: Instead of 4 to 8 separate MCUs plus safety ICs, a designer can use 2 to 4 AM26 chips, each handling 6 axes. The inter-chip communication is simplified through the built-in EtherCAT or single-pair Ethernet interfaces, reducing the total wiring harness volume and weight by an estimated 50% or more.
For industrial CNC and motion systems: A single AM26 chip can replace a multi-axis motion controller plus individual servo drives, consolidating the control hierarchy and reducing the total number of components in the control cabinet.
Communication Architecture: CAN Is Giving Way to EtherCAT and Single-Pair Ethernet
One of the less-discussed but equally important aspects of TI’s announcement is the communication architecture migration. Heo Jeong-hyeok specifically highlighted the transition from traditional CAN bus to EtherCAT and single-pair Ethernet (T1 standard) for next-generation robotics.
This migration is driven by three factors:
1. Bandwidth Requirements
CAN bus maxes out at 1 Mbps (or 5 Mbps for CAN FD), which is insufficient for real-time synchronization of 24+ axes. EtherCAT delivers 100 Mbps with deterministic timing, and single-pair Ethernet (100BASE-T1 and 1000BASE-T1) provides gigabit speeds over a single twisted pair.
2. Wiring Reduction
Single-pair Ethernet transmits data over just one pair of twisted wires instead of the traditional four-wire configuration, reducing wiring harness volume and weight by 50%. For humanoid robots where every gram and millimeter matters, this is a game-changer.
3. Diagnostic Capabilities
T1 Ethernet includes built-in diagnostic mechanisms that can predict potential cable breakage — something CAN bus cannot do. In industrial environments where cable failure can cause costly downtime, this predictive capability adds significant value.
TI has begun supplying 100BASE-T1 and 1000BASE-T1 gigabit TSN physical layer chips, enabling designers to implement this architecture today. The combination of single-chip multi-axis control with single-pair Ethernet creates a complete system-level solution that dramatically simplifies robot wiring.
Three Functional Safety Topologies for Different Risk Levels
Functional safety is no longer optional in industrial and collaborative robotics. TI has released three topology structures that meet ISO 13849 and SIL3, PLd/PLe safety standards, giving designers flexibility to match safety architecture to application requirements:
Topology 1: Main MCU + 2 Parallel Embedded MCUs (Plif)
A main control MCU operates alongside two parallel-running embedded MCUs that act as safety monitors. Upon detecting a physical fault, the system forcibly switches to a safe state. This topology is already pre-certified by TUV Rheinland, reducing time-to-certification for OEMs.
Topology 2: Dual-MCU in the Same Chip
The safety channel and application computation are separated within the same silicon die. This reduces hardware volume and material costs while maintaining safety integrity — ideal for cost-sensitive applications where the full external safety MCU approach is overkill.
Topology 3: Single MCU + Intelligent PMIC (Lockstep)
A single MCU combines with an intelligent power management IC, driving the R5F safety core in lockstep mode. This achieves a 99% fault detection rate in a single-chip configuration, making it suitable for humanoid robot joints where space is at a premium.
For motion control engineers evaluating safety options, the key takeaway is that functional safety is no longer a separate subsystem requiring additional chips, PCBs, and wiring. TI’s integration of safety into the motor control MCU itself reduces both system complexity and bill-of-materials cost.
What This Means for Indian Electronics and Factory Automation
India’s electronics manufacturing sector is at an inflection point. With the ECMS (Electronic Components and Manufacturing Systems) scheme boosting domestic component production, and the government pushing for higher value-addition in electronics manufacturing, solutions like TI’s single-chip motor control become particularly relevant.
For Indian OEMs building CNC machines, pick-and-place systems, conveyor automation, or emerging robotic platforms, the benefits are direct:
- Lower BOM cost through chip consolidation
- Simplified PCB design with fewer components and interconnections
- Faster time-to-market with integrated safety reducing certification timelines
- Future-proof communication with EtherCAT and T1 Ethernet support
- Predictive maintenance built into the silicon through the AI acceleration core
Companies like Neway & Velton, working in servo motors, motion control systems, and conveyor automation, can leverage these single-chip solutions to develop more compact, cost-effective products for the Indian and export markets.
Related Reading
- Electronics & Microcontrollers — More embedded systems articles on justLast
- Technologies — Industry trends and analysis
Frequently Asked Questions
What is the closed-loop latency of TI’s new single-chip motor control solution?
TI’s new solution achieves a closed-loop delay controllable within 1 microsecond, with core computation latency of 500 nanoseconds. This is approximately 10 to 100 times faster than traditional MCU-based motor control systems.
How many motor axes can TI’s single-chip solution control simultaneously?
The AM26 series and C2000 DSP family can simultaneously run 6 complete FOC control loops on a single chip. For a 24-axis humanoid robot, this means 4 chips can handle the entire motion control workload, compared to 8 or more chips in traditional architectures.
What safety standards does TI’s motor control solution meet?
TI offers three functional safety topologies meeting ISO 13849 and SIL3, PLd/PLe standards. The TUV Rheinland pre-certified topology (Main MCU + 2 parallel embedded MCUs) provides the fastest path to safety certification for OEMs.
Can TI’s single-chip solution replace a traditional PLC + servo drive architecture?
Yes, for axis counts up to 6 per chip. The AM26 series integrates motion control, communication, and safety on a single die, effectively combining the functions of a motion PLC and individual servo drives. For larger systems, multiple chips communicate via EtherCAT for coordinated multi-axis motion.
How does single-pair Ethernet (T1) benefit robotics wiring?
Single-pair Ethernet transmits data over one pair of twisted wires instead of four, reducing wiring harness volume and weight by approximately 50%. It also includes built-in diagnostic mechanisms for predictive cable failure detection, which CAN bus cannot provide.
Is this solution available for purchase or evaluation today?
The AM26 series and C2000 product family are available through Texas Instruments’ authorized distributors. Evaluation kits and development boards can be obtained through TI’s product pages. The solution was showcased at the TI Mobility & Robotics Seminar 2026 in Seoul on July 9, 2026.
This article was originally published on justLast.in on July 14, 2026. Sources: TI Mobility & Robotics Seminar 2026 (Seoul, July 9), Texas Instruments Korea, Engineer News Network, Industry USA.
