High-resolution current and voltage sensing feedback design
It is considered the biggest challenge in designing motor control systems.
TI C2000 with built-in high-performance ADC reduces system costs.

Automation applications are integrating more industrial robots and automated machines.
A robot can pick up an engine block, precisely place it on a vehicle chassis, and return to its original position to perform another task with the same precision every time. However, these tasks are impossible without precise motor control.

Designing high-resolution current and voltage sensing feedback is a major challenge in motor control systems. To avoid potential analog signal loss or interference, you will often want to digitize the feedback signal as close to the sensor as possible.

In this case, the digital feedback signal can cause potential timing problems, as the timing window narrows as the clock speed increases and the sampling rate increases. It also exacerbates the problem of different trace lengths between clock and data signals.

A webinar held on October 30th at e4ds.com provided a comprehensive overview of factory automation applications, including TI's C2000 family of products that can address these design challenges. Details of the article can be found in the webinar.


Solutions to High-Resolution Current and Voltage Sensing Feedback Design Challenges
The C2000 product family can be divided into the following groups based on performance and peripheral device characteristics.

TI's most recently released F28004x products and the 37x family have excellent code compatibility because their peripheral IPs are of the same generation.

TI C2000 MCUs are currently available in two product families: Piccolo for low- and mid-end products and Delfino for high-end products.

The 37x family, ideal for factory automation and servo controllers, features improved cores for computational acceleration and extensive adoption of analog IP. The F28004x product family can be used in home appliances and powertrain applications for electric vehicles (EVs) or hybrid EVs by adopting most of the IP of the 37x family while reducing the price.

The 37x family delivers up to 800 MIPS of compute performance through a dual-core CPU and dual CLA. The 37x family is the first to feature a TMU module capable of performing trigonometric functions and division in one to two cycles, maximizing FOC computational capabilities for motor control.

Additionally, it has up to four independent ADC converters, enabling high-speed sampling of up to 14M per second, and up to eight channels of window comparators for overcurrent/overvoltage tripping functions.

The F28004x family incorporates up to seven channels of programmable operational amplifiers (OP-AMPs), a feature not found in the 37x family. Therefore, when using a shunt to measure phase current during motor control, the shunt's output signal can be directly connected to the MCU without the need for a separate external OP-AMP for signal amplification.

Additionally, it has three independent ADC modules, enabling sensing of three-phase currents at the same time when controlling a three-phase motor, thereby contributing to improved control performance.

The F28004x family can operate at up to 100 Hz using an internal PLL and boasts a computational performance of up to 200 MIPS, including CLA. While most existing Piccolo family members consisted of fixed-point cores, the F28004x family features an FPU capable of hardware-based floating-point processing. This allows users to easily develop C code in floating-point format.


C2000 MCUs overcome the shortcomings of 32-bit FPGAs and ASICs.
Many industrial inverter and servo drive manufacturers add separate FPGAs or ASICs to implement functions not available in commercial 32-bit MCUs.
However, implementing encoders or sigma-delta filter modules for position sensors in FPGAs or ASICs increases system cost and development complexity. This is due not only to the increased price of the FPGA itself, but also to the additional cost of a separate power supply IC and external ADC converter. There is also the problem of needing to secure additional specialized personnel to develop VHDL code for FPGA.
If a single SoC with a built-in high-performance ADC is used and no external logic IC is required, system cost can be reduced compared to the MCU+FPGA architecture.

C2000 MCUs, which can be equipped with TI's DesignDRIVE technology, are commercial MCUs that implement a high level of system integration and have the advantage of shortening development time and costs. Materials that allow you to experience TI's DesignDRIVE technology are shared on TI's website.


How to use C2000 MCUs with EtherCAT
There are two main ways to use EtherCAT, one of the most widely used communication methods in factory automation, with C2000 MCUs.
The first method involves connecting Backoff's ET1100 to the C2000 via EMIF or SPI. TI provides driver files for accessing the ET1100's registers, and an EtherCAT slave stack optimized for the C28x is available from the EtherCAT Technology Group (ETG). The TI reference design TIDM-DELFINO-ETHERCAT simplifies EtherCAT design on C2000 MCUs.

A second approach is to implement EtherCAT slaves and masters using TI's AMIC110 communications processor, which is compatible with the ET1100. The AMIC110 uses an ARM Cortex A8 core and supports not only EtherCAT but also virtually all industrial communication protocols. It can be connected to the C2000 MCU via SPI communication, and the register MAP for EtherCAT is compatible with Backoff's ET1100.

The CANOpen protocol, which utilizes CAN, another communication method widely used in recent factory automation, can be supplied as a paid stack through Simma Software, a third-party company of TI.