Posted on 01 June 2019

Evolution of Power over Ethernet

Free Bodo's Power Magazines!




The four-pair technique has advantage

The IEEE 802.3af standard for Power over Ethernet (PoE) introduced a new facet to Ethernet networking, delivering DC power in tandem with 10/100/1000 Mbps data. PoE brought with it a unique set of problems and new ways of thinking that were unfamiliar to many engineers with experience designing Ethernet equipment.

By Clay Stanford, Design Manager - PoE is now commonly used for VoIP phones, wireless access points and security cameras. As PoE evolves, there is a need for enhancements to the standard to enable emerging applications.


By way of review, the PoE link allows a Powered Device (PD) to draw up to 12.95W from the Power Sourcing Equipment (PSE). The PoE link or port is controlled by the PSE, which identifies PDs via detection and classification before powering and monitoring the port (ICUT, ILIM and disconnect). Much of the burden of PoE rests on the PSE, which must perform detection and disconnect flawlessly to avoid damaging legacy devices. If the PSE does not adequately perform classification, power delivery and monitoring, intermittent failures and instabilities may result. The PSE cannot control everything. When it applies power, it trusts the PD to follow the standard, turn on without oscillating and avoid drawing more power than requested. Because both types of devices must cooperate, PD and PSE designers should consider issues from the perspective of both devices.

Higher Power for Emerging Applications

Thirteen watts is adequate for basic IP phones, but motorized cameras, multi-radio access points, and devices with large color screens are seriously constrained. The IEEE is currently working a higher power standard, dubbed PoE+ (officially IEEE 802.3at) that will coexist with 802.3af devices available today. The ultimate power levels defined by the new standard have not been established, but as of today, it is likely we will see 30-watt 2-pair systems and 60-watt 4-pair systems. The IEEE 802.3at committee has taken on a daunting task to define a world standard that will define a safe, higher-power standard that is backward compatible and interoperable with currently deployed 802.3af equipment. Because of the complexities of writing this specification, we are not likely to see a final specification for 1 ½ to 2 years from now. Although typical CAT 5 cabling includes four twisted pairs, the 802.3af standard only allows two pairs to carry current at any given time. One option is to allow additional current down the third and fourth pairs, doubling the available power. A second option is to raise the current limit, allowing more power down the same two pairs. Each of these techniques has appeared in proprietary PoE systems. However, each has drawbacks, complicating the choice between them.

Implementing Pre-Standard High Power PSEs

In the interim, some applications require high power and cannot wait for the new standard. There are several ways to do this. The following circuits build on the basic 802.3afcomplaint PSE circuit using the LTC4258, shown in Figure 1a, and the basic PD circuit using the LTC4257 shown in Figure 1b. The LTC4259 can be substituted in the PSE circuit if the application requires AC Disconnect. The LTC4267 can be substituted in the PD circuit if the application requires an integrated switching regulator.

 Basic 802.3af-compliant PSE circuit using LTC4258

The following circuit examples demonstrate several ways that high power operation can be achieved. Note that in some of the following PSE circuits, Channel 4 is used to illustrate the circuit changes, but any channel can be used.

Basic 802.3af-compliant PD circuit using LTC4257

Two-Pair, High Current

The power level can be increased in the PSE simply by changing the value of the sense resistor (RS1 – RS4 in Figure 2a). RSn is set to 0.5 for 802.3af compliance (375mA ICUT, 425mA ILIM). Cutting the value of RSn to 0.25, for example, doubles all the current limits (750mA ICUT, 850mA ILIM). This doubles the power to the PD when a short cable is used. Increased losses in longer cables limit the power delivered to the PD to somewhat less than 2x.

Dual current High Power- 802.3af PSE

Note that the LTC4258 also uses the sense resistor to detect DC disconnect. Cutting the resistor value to 0.25 will double the DC disconnect threshold, rendering it technically non-compliant. Other 802.3af parameters are not affected: detection and classification will remain compliant, and the AC disconnect threshold (LTC4259 only) is not affected by the sense resistor change. Because the raised DC threshold runs a small risk of disconnecting a very low power 802.3af PD, AC disconnect is recommended for interoperability with 802.3af PDs.

Two other components (per channel) need to be modified to handle the extra current. MOSFET Q4 will typically need to be replaced with a larger device to tolerate the higher power during current limit. In this application, IRF530-type devices in D2PAK packaging are adequate. The PoE data magnetics also need to be specified to carry higher current. Several magnetics vendors have newly released parts with adequate current capability.

By adding two new components, we can switch between 802.3af-compliant operation and high power. In this case, RS4 is set to the original 0.5 value and RS4B is chosen so that RS4 || RS4B gives the desired higher current level. Setting RS4B to 0.5 (the same value as RS4) sets the high power mode to twice the power level of 802.3af.

When Q4B is off, the port operates in 802.3af-compliant mode. Turning on Q4B switches the port to high current mode. This transition can be made at any time: before detection/classification, after detection/classification, but before port power-up, or after power is applied. Note that Q4B can be a low-voltage MOSFET, since only the drain of Q4 sees the high port voltage. Q4B should be selected for very low on-resistance to prevent inaccuracies in the higher current limit. The IRLML2502 is an example of a suitable device in a SOT-23 package.

The PD modification (Figure 2b) is slightly more complicated since the internal MOSFET is pre-configured for 375mA current limit operation. However, adding an external pass device controlled by the PWRGD pin allows high current operation while maintaining full 802.3af detection and classification signatures and inrush current limiting.

Two-pair high power PD

Four-pair, Low Current

An alternate technique to increase the power delivered to the PD is to power all four pairs in the CAT-5 cable. Figure 4a shows a four-pair PSE circuit with standard 802.3af power available on each pair. No changes are required to the sense resistor values.

Four-pair 802.3af Power

The bigger change is to the four-pair PD circuit (Figure 4b). Two LTC4257 devices are now required, and the power supply circuitry must be smart enough to limit the current draw from each channel to stay under the 802.3af limits. It can do this by balancing the current drawn from each pair set, or by drawing power from one pair set until it approaches (but does not exceed) the ICUT limit, then beginning to draw from the other set. This circuitry can be fairly complex, and will vary from design to design.

Four-pair low-current PD

Caption Figure 4b

The four-pair technique has the advantage of using all the conductors in the cable, minimizing the total cable resistance and the resulting power lost with long cables. Using standard current levels also comes closest to full 802.3af compliance of any of the high power techniques, since either the signal pairs or the spare pairs used alone will be fully compliant. The primary drawbacks are complexity and expense. The PSE requires two channels of the controller chip per port, halving the effective port density, while the PD requires two controller channels and additional current balancing circuitry to ensure that the current drawn from either pair set does not exceed the maximum level. In addition, four-pair techniques will not work if only the signal pairs have continuity, as in some CAT-3 building installations.

Because of the cost and complexity penalties of the four pair scheme, two-pair high current is the preferred technique at medium power levels. Four-pair systems are most applicable when the PD power rises above the 35W level.

Four Pair, High Current

Combining high current circuitry with four-pair hookup allows more power down the cable than any other technique. Fourpair high current allows as much as 50W to be delivered to the PD over a 100m CAT-5 cable, more if the cable length is kept short. Although this scheme includes the drawbacks of all of the previous schemes, it is the highest power option available.

Beyond 50W, long cables rapidly approach an “impedance matched” situation where the cable dissipates more power than it delivers to the PD. If the length of the cable is kept short, the current level can be further increased, ultimately limited by the RJ45 connector, offset current in the magnetics and the temperature rise in the CAT-5 cable. Extremely high power (>50W) should only be used in systems where the entire solution is specified by the same supplier.

Classification: When to Apply High Power

Notably absent from the above circuits is a method to determine when it is appropriate to apply high power to the line. All of the techniques will successfully power standard 802.3af PDs under normal conditions. The dual-threshold circuits need to have some information from the PD to know when to switch thresholds, and the four-pair schemes need to know when it is appropriate to switch on the second set of conductors. The IEEE 802.3at committee is working to address these issues, but no schemes have yet been finalized. In the interim, ad-hoc solutions are needed to identify high power PDs.

802.3af defines an unused class (Class 4) that looks tailor-made for high power, and both the LTC4258/59 PSE chips and the LTC4257/67 PD chips support Class 4.

Unfortunately, a Class 4 PD will be powered with Class 3 current limits if it is plugged into a standard 802.3af PSE, which may cause it to cycle on and off repeatedly if it attempts to draw higher power. Class 4 can be used as a “warning” that a high power PD is connected, but it is advisable to have an additional handshake before higher power is delivered. Ideally, a high power PD should receive some sort of signal from a highpower PSE, acknowledging it is acceptable to operate in a high power mode. If no handshake is received, the PD should give some sort of indication to the user that it is plugged into the wrong kind of PSE.

The best technique to use for pre-standard high power depends on the application. At power levels up to 30W, the two-pair high current techniques provide the lowest cost and complexity, and full 802.3af compliance by using the dual-threshold circuit. If maximum power is required (50W or more), the four-pair high current circuit is the best choice.



VN:F [1.9.17_1161]
Rating: 0.0/6 (0 votes cast)

This post was written by:

- who has written 791 posts on PowerGuru - Power Electronics Information Portal.

Contact the author

Leave a Response

You must be logged in to post a comment.