Evolving Power over Ethernet Expands Power Distribution


What you’ll learn:

  • Advantages of Power over Ethernet in the context of industrial Ethernet.
  • PoE standards continue to evolve to keep pace with increasing power demands.
  • Driver for standards updating is the potential movement toward Gigabit Ethernet.

 

Industrial Ethernet has disrupted factory automation through its improvement in speed and latency performance. Engineers achieve these conditions by employing 100-Mb/s fast Ethernet in switch mode with full-duplex capability. This design unlocks point-to-point device connectivity, creating an Ethernet solution to help manage heavy data usage of automated industrial applications.

As a component of industrial Ethernet, Power over Ethernet (PoE) integrates the transmission of electrical power and data in a coupled cable. The cables deliver dc power without the need to adapt ac, enabling centralized control.

Though lower voltage partially offsets the increased efficiency and ease of deployment of PoE, standards written to address the energy efficiency augment the complexity reduced by combining power transmission with data. Standards continue to evolve to accommodate higher power-distribution capability, extending PoE for a broader range of equipment.

About Industrial Ethernet

To understand why professional associations such as IEEE write standards to expand power-distribution capability, it’s worth reviewing the history of Ethernet. The initial suite of Ethernet speeds transmitted data in the following forms: 10 Mb/s for original Ethernet, 100 Mb/s for Fast Ethernet, and 1 Gb/s for Gigabit Ethernet.

This first incarnation of Ethernet provided a way to transmit a signal between devices over a shared medium, but it wasn’t useful for industrial applications due to limited capacity. However, with the improvement of using Fast Ethernet (100 Mb/s) in switch mode with full-duplex capability, engineers could leverage a point-to-point link between devices, allowing Ethernet to be used for most industrial applications.

All industrial Ethernet protocols require some level of determinism, which traditionally has been addressed by using a unique software protocol stack. With these protocol opportunities, several other design trends are driving standards updates or their creation.

Speed and Latency

Beyond combining Fast Ethernet with full-duplex capability, effective industrial Ethernet solutions require application-specific design features. For example, differing applications carry unique requirements for data-response time. Healthcare may require response times for brain-computer interface applications where the user’s mind sends signals via a sensor to a processor, then back to the body for response.

In the same field, providers may need only millisecond response times for remote patient monitoring. The volume, speed, and latency requirements of data transferred and analyzed dictate the system design. Multiple design approaches can achieve a communication protocol to support differing latency requirements.

Embedded Communication Protocol

With increased integration between system actions, the communication function has become deeply embedded due to pressure to reduce system cost, form factor, and power budget. In the past, customers bought an off-the-shelf communication module costing hundreds of dollars, which could then be added to the drive module. This type of module is neither cost-effective nor appropriate for smaller form-factor drive designs.

Another alternative was to include a separate application-specific standard product (ASSP) dedicated to communication functions. This ASSP would be overdesigned to support many protocols because different customers would use different industrial Ethernet standards. However, as vendors seek to integrate all of their digital drive functions onto a single piece of silicon, the communication protocol is by necessity becoming a small function implemented as part of the entire “drive-on-chip” design.

Movement Toward Gigabit Ethernet

Another driver for standards updates is the potential movement toward Gigabit Ethernet. Since almost all FPGAs can support Gigabit Ethernet, even if the standards start to move toward the higher 1-Gb/s speeds, a well-thought-through system design would need just a new FPGA programming file to support any such standard evolution.

Implementing an industrial Ethernet protocol as a deeply embedded function within a programmable fabric provides engineers with the flexibility of supporting multiple protocols with the same hardware. It also aligns the system with all of the benefits that flow from a highly integrated design: optimized power, cost, and form factor.

The Proliferation of Industrial Ethernet Standards

With the advances and interdependencies stemming from industrial Ethernet, there are many variants of industrial Ethernet protocols. The evolution of how data is used, processed, transmitted, and stored creates the need for professional organizations to define a set of best practices around industrial Ethernet usage. The standards would need to be written inclusively to ensure they cover all competitive models of a specific system component.

For example, a drive-system vendor must be able to support about six to eight standards to sell its products worldwide to different plants. If the vendor plans to sell the drives in both Asia and Europe and operate agnostically to Ethernet technology, there are three principal options:

  • Design, develop, and maintain multiple sets of drive designs.
  • Include an ASSP that supports multiple protocols—and hope the protocols don’t change.
  • Use a programmable platform.

Historically, when the industrial Ethernet standards used a standard MAC/switch, it was easy to dedicate a microprocessor unit (MPU) for communications. To support a new standard, system designers could swap out the protocol stack (software). However, many of the more recent standards require a specialized MAC implementation.

Standardizing the communication protocol implementation on a standard MPU with a standard Ethernet MAC and switches falls short when dealing with such newer standards. Some MPU vendors have developed techniques like developing custom microcode for a proprietary embedded processor to emulate a non-standard MAC. But these are esoteric methodologies with unknown pitfalls.

Protocols that require a specialized MAC implementation are usually best applied with a custom hardware approach using either an ASIC or an FPGA, depending on the volume and desired price point. Also, there’s always a possibility that the MAC design may change as these standards evolve. To future-proof the system design against such an eventuality, a programmable hardware approach is the safest path.

PoE is a crucial enabler to the future of industrial Ethernet and has its own set of standards to deliver increasing amounts of power to connected devices. The connection between data transmission and power is a critical one, necessitating PoE’s own set of standards that must evolve with industrial Ethernet.

Why Standards Need to Evolve

Because low-voltage operation hinders PoE by lower-voltage operation, professional organizations modify standards to promote higher-power-distribution capability for industrial Ethernet solutions using PoE. The infrastructure for PoE already exists in the data-transmission pathways, so making it easier to distribute power with PoE enables its application to an increasing amount of equipment. The relevant standards are summarized below (multiple classes exist for a given device type):

  • IEEE 802.3af-2003: The original PoE standard specified up to 15.4 W dc power to be transmitted through the 2-twisted-pair cable. Line losses were allowed to consume no more than 16% of the power before reaching the powered device, delivering 12.95 W dc (type 1 device).
  • IEEE 802.3at (2009): The 2009 improvement to the PoE standard pushed the limit to 30 W dc (type 2 device) with delivered power of 25.5 W dc over the pair.
  • IEEE 802.3bt (2018): Applications continued to evolve to push power demand higher. As a result, the latest revision of the standard in 2018 elevated the power delivered to powered devices from 51 W dc (type 3) to 71.3 W dc (type 4) with less than 21% line losses required.

The increasing power specified in the standards underscores the point that PoE is an increasingly desirable method to distribute power to connected devices. Engineers achieve higher power levels by passing current through all four twisted pairs of the network cable for the higher classes of type 4 devices.

Expanding power-distribution capability enables the IoT by removing the IT and Operations disconnect within facilities management—a historical challenge—by improving integrating connectivity of technology with equipment. Rather than run electrical and Ethernet, only one cable bundle is needed. This integration dramatically simplified installation and routing paths while decreasing installation costs.

With the propagation of 5G addressing bandwidth constraints, system performance becomes limited by how power is delivered to connected devices. Stated differently, supplying on-demand, reliable power (when needed) is the critical function of the facility. More delivered power unlocks additional features, improving the user experience with a connected device. With the new PoE standard, more devices over a network can integrate power with data transmission.

Conclusion and Applications

Like many other communications functions, industrial Ethernet implementation is moving from the module to the device to becoming a deeply embedded function. This evolution is standard for many tasks as system vendors work hard to optimize their design for cost, power, form factor, and more.

There are numerous advantages of PoE in the context of industrial Ethernet. PoE is cost-efficient due to a single cable handling both power and data transmission. This benefit leads to fewer connections, and it facilitates centralized control better than traditional ac-dc conversion while offering near-infinite flexibility in cable-routing patterns. PoE is efficient, delivering on-demand power to the active powered device in only the amount it needs to operate.

While enabling the IoT and propelling innovation in industrial Ethernet, PoE standards continue to evolve to keep pace with increasing power demands. Higher degrees of integration of devices and facilities equipment, coupled with increased data generation and analytic rates, drive demand for reliable delivered power. IEEE 802.3 has continued to evolve over the years to provide added power for data transmission.

The surge in delivered power allows more types of equipment and devices to realize the advantages of PoE. Original PoE supported static surveillance cameras, wireless access points, and [voice-only] phones. The 2009 revision allowed alarm systems and video phones to utilize PoE. The 2018 update brought in video conferencing, laptop computers, flatscreen TVs, and many of the devices used by the IoT.

The rate at which that power can be delivered to connected devices determines the speed that data can be transmitted and analyzed over industrial Ethernet. PoE enables reliable high-power loads to reach connected devices in a manner that improves facilities’ installation processes and envelops a wider range of equipment.



Source link