The long-awaited availability of IEEE 802.3 “Physical Layer and Management Parameters for Power over Ethernet over 4 pairs” (also known as IEEE 802.3bt-2018 and PoE++) is official, with the standard approved in December, and formally published in January of this year. This was not a surprise, as the standard has been in the works for several years. Component and box-level vendors have been working with its pre-formal versions to develop supporting ICs and switches, such as the  PDS-408G PoE switch from the Microsemi subsidiary of Microchip Technology (Figure 1), and they can now even mark them as “compliant” (assuming they have been following it closely and do a final approval check, of course).

Figure 1 This PDS-408G PoE switch from Microsemi was released soon after the official approval of the IEEE 802.3 upgrade standard for PoE++. (Source: Microchip Technology)

I won’t go through the specifications and capabilities of this latest PoE version; there are plenty of sources available online and in the references below. In brief, under the right circumstances, and with the correct Ethernet cabling, it can deliver up to 100 W to a load. That’s serious power, enough to run a large variety of devices including LED-based area lighting.

Who wants PoE++? (I prefer the informal, descriptive name rather than the formal, “faceless” IEEE designation; that’s just a personal idiosyncrasy.) Apparently, architects and office lighting installers are among the first to try it. There’s been quite a lot of coverage of the standard in the excellent publication Cabling Installation & Maintenance, giving an in-the-field, reality-based view from those who actually have to install PoE++ systems, rather than just IC or box designers. Their recent article “The IEEE 802.3bt standard’s impact on the expanding PoE marketplace” provided a good introduction, while the brief case study “Cabling Innovators Spotlight” shows that PoE++ is viable for office lighting using a very real example (they have had others, as well.) So PoE++ is very much here; real, and in the field.

Note that PoE++ with its higher-power rating is not a simple drop-in upgrade to existing Ethernet cabling. First, you have to use the new Category 6A Ethernet cabling and there are classes 1 through 8 of available power it can deliver. The new cabling is identified by certification logos (Figure 2) provided by the Ethernet Alliance (an independent, vendor-neutral organization focused on expansion of Ethernet technology and “operates as a marketing organization for IEEE standards”).

Figure 2 The Ethernet Alliance has developed package logos used to indicate the level of compliance for IEEE 802.3 Category-6 cabling (Source: Ethernet Alliance, via Cabling Installation & Maintenance)

Despite their efforts, I worry that there will be counterfeit Cat-6A cables with those logos, promising performance they really can’t provide. This happens with Cat-X cabling quite often according to Cabling Installation & Maintenance and other sources, and it’s very hard to validate or confirm cabling on-site; the deficiencies are usually apparent only after cabling is installed, in use, or several months have passed.

Even with the genuine cable, there are serious thermal issues, and this may be especially troublesome as we don’t associate data cables with dissipation. There will be IR drop and I2R loss, and thus self heating. That’s where it gets tricky, as the PoE++ Cat-6A cabling will often be packed within tight bundling of other PoE++ and even AC-power cables in a raceway or enclosed duct. As a result, the actual power rating will be far less than 100 W since significant temperature rise can occur, enough so that the National Electrical Code (NEC) is addressing the issue in terms of allowable packing, overheating, and possible flaming.

I see the potential benefits of PoE++, as it allows for more and better power control without the need to run high-voltage AC lines for both power and its control, and may eliminate or reduce the need for licensed electricians to run the cable. But that’s only part of the cabling-install story. My concern is that PoE++ is transforming a simple loop consisting of source, on/off switch, copper cable, and load into yet another protocol-driven system where a complex sequence of queries, responses, and validations are needed to get it working the first time, and then keep it working.

There’s a lot to be said in favor of the simple-to-connect, simple-to-troubleshoot direct-connect approach. When you start adding power delivery and management to data-first cables (think of PoE++ or USB-C, and even RS-232/485) or data capability to a primarily power-delivery cable (broadband over powerline – BPL) things can get tricky quickly in terms of start-up and test. Certainly, adding low-speed data capability to the power line makes sense for smart meters, but I’m not so sure about BPL; it takes a lot of active and passive components to merge them while maintaining safety. PoE++ is somewhere in-between, as the voltage levels are low enough that immediate safety is not as issue, but you still need active, passive, and magnetic components at each end to merge, then separate the data from the power (Figure 3). Of course, there are many “bootstrapped” non-standard schemes which engineers devise to be used just with their application – that’s another story completely.

Figure 3 This highly simplified block diagram can only begin to show what is needed to merge and then separate power and data for PoE installations; the reality is far more complicated and needs a lengthy BOM with many active and passive components. (Source: Kintronics)

So, where will PoE++ go? Will it become widespread and commonplace? Or will its use be limited to highly-specialized applications such as commercial building lighting and control? Will it become another “could have been a contender” that just didn’t catch on as anticipated? Check back in five years or so, it will be interesting to see. After all, perhaps the developers of USB had a vision of what their low-speed, low-cost interconnect between a PC and its keyboard/mouse would become for data speed and power delivery in just 20 years (USB was introduced in the mid-1990s), but I suspect most design engineers did not.

What’s your view on the desirability and benefits of PoE++ (and PoE+ and basic PoE)? Will the inevitable pain which accompanies any major new option be worth the gain, from your perspective?

Bill Schweber is an EE who has written three textbooks, hundreds of technical articles, opinion columns, and product features.

References

Related articles: