A look at Microchip’s new dsPIC33A digital signal controller

Discussing the recently released dsPIC33A with the corporate vice president of digital signal controllers at Microchip. The post A look at Microchip’s new dsPIC33A digital signal controller appeared first on EDN.

A look at Microchip’s new dsPIC33A digital signal controller

INCREASE YOUR SALES WITH NGN1,000 TODAY!

Advertise on doacWeb

WhatsApp: 09031633831

To reach more people from NGN1,000 now!

INCREASE YOUR SALES WITH NGN1,000 TODAY!

Advertise on doacWeb

WhatsApp: 09031633831

To reach more people from NGN1,000 now!

INCREASE YOUR SALES WITH NGN1,000 TODAY!

Advertise on doacWeb

WhatsApp: 09031633831

To reach more people from NGN1,000 now!

Microchip recently unveiled the newest iteration of its enduring dsPIC family: the dsPIC33A. In an interview with Joe Thomsen, corporate vice president of digital signal controllers (DSCs) and Alexis Alcott, associate director of marketing at Microchip, the design and application scope of the new 32-bit MCU were elucidated. 

Performance upgrades

The dsPIC was intended for real-time control to fundamentally: 

  1. Sense the state of a machine
  2. Perform calculations, and
  3. Make adjustments to the machine via a feedback loop

And while this functionality is not new to the 20-year old dsPIC family, the large leaps in performance are (Figure 1). 

Figure 1: The dsPIC33 DSC core evolution from the dsPIC30F to the dsPIC33A. Source: Microchip 

The dsPIC33A upgraded from its previous 16-bit core to a 32-bit CPU operating at 200 MHz for high performance. “All data paths have doubled, instruction set architecture doubled in size, the map engines all doubled, and by doing just that, you effectively get almost a 2x increase in performance” states Joe Thomsen. An additional double precision floating point unit (DP-FPU) addresses the previous generation’s limitations with fixed point math where users had to manually convert all floating point numbers, leaving potential room for human error. “If I’m using MATLAB, or any high level language that does math, you really want to be using a floating-point unit to use floating point math directly for ease of use.” The DSP engine advanced from 40-bit to 72-bit accumulators for better accuracy and 32-bit working registers “so when there is an interrupt, the user can switch a register set instead of having to push the state of the machine onto the stack at the beginning of an interrupt and pop it off the stack at the end of the interrupt”. 

Two high speed 12-bit ADCs jump from 3.5 megasamples per second (MSPS) to 40 MSPS for more efficient real-time control, allowing the embedded system to more quickly and accurately sense and respond. The DSC includes security with many hooks in hardware that will limit access including the potential for memory partitioning, secure debug, and immutable root of trust (RoT), etc. A slew of new peripherals have been included such as high resolution PWM modules, quadrature encoder interface (QEI) for motor control, as well as revamped comparators and op-amps with 100 MHz gain-bandwidth product (GBW). 

Several communication peripherals have also been added including I3C, ethernet T1S, and bidirectional synchronous serial interface (BiSS)—a protocol to implement a real-time interface. Finally, hardware trace has been introduced to remove the need for hardware breakpoints in the processing of debugging a control loop. Thomsen remarks, “If you’re spinning a motor in an EV that is taking ~300 kW of power, you don’t want to be stopping it and not controlling it anymore.” Hardware trace allows customers to monitor real-time variables while the motor is running without disturbing the program.

Figure 2: The dsPIC33A platform block diagram with key features and enhancements. Source: Microchip

Application 

Motor control

However, what is the justification for this massive leapfrog in performance? Thomsen comments on this point by zooming into the two major application spaces of the dsPIC33A: power conversion and motor control. “Algorithms keep on getting more complex, customers are expecting to be able to run three, four, or even five motors from the same microcontroller and want to integrate other functions in cost-driven power conversion.” Applications for motor control include industrial fans, pumps, robotic arms, and autonomous guided vehicles (AGVs) on factory floors that demand more automation (Figure 3). In the automotive space, subsystems that were traditionally hydraulic or mechanical are now leveraging electric motors, increasing the number of motors per vehicle by an order of magnitude. Automotive use cases are generally not limited to commuter cars or even commercial vehicles but also e-mobility with e-scooters, e-bikes, and e-motorcycles. 

Power conversion

Regarding power conversion Joe notes, “The industry has shifted towards wide bandgap devices such SiC and GaN, which has enabled our customer to have extremely fast control loops that allow for incredibly efficient algorithms, but they need the performance for it.” In the pursuit of more sustainable solutions, electronic devices across numerous industries demand power density and energy efficiency, ranging from energy-efficient appliances to server power supplies in data centers. The applications for efficient power conversion and control extend beyond these industries: “LED lighting also necessitates responsive, flexible PWM controls to adjust brightness and color.” 

Figure 3: The dsPIC33A application spaces. Source: Microchip

dsPIC33 support

“As we did the dsPIC33A, our goal was to allow you to sense the state of the machine, do calculations based on it, and adjust the machine as fast as possible, because the faster you can do that, then the more cycles of learning you have with the machine,” Thomsen adds. However, accomplishing this fine-tuning within an embedded system can demand a substantial level of expertise. It is in this context that Microchip’s established ecosystem plays a pivotal role, accelerating the process through the provision of evaluation boards, reference designs accompanied by source code, software tools, and application firmware. Moreover, the dsPIC33A ensures compatibility with the legacy code and ecosystem of preceding dsPIC33 generations.

According to Thomsen, “the whole world is going model-based and so we’re working hard to make sure that you can model all of our parts,” where the goal is to equal or equivalent results when the system is put into actual hardware, “this way, you’re much more a system architect and the software is generated by the tools.” The seamless integration of model-based designs is particularly advantageous in high-power applications where traditional testing methods, such as those required for hundred-kilowatt fast chargers, pose significant cost and safety risks. Moreover, standards and safety certifications play a crucial role in various application domains, including automotive and data centers. The dsPIC33A incorporates an array of standards to cater to evolving regulatory requirements, encompassing functional safety, cybersecurity, and compliance with NIST standards.

Aalyia Shaukat, associate editor at EDN, has worked in the design publishing industry for seven years. She holds a Bachelor’s degree in electrical engineering, and has published works in major EE journals.

Related Content

The post A look at Microchip’s new dsPIC33A digital signal controller appeared first on EDN.

What's Your Reaction?

like

dislike

love

funny

angry

sad

wow