Ever found yourself in the situation where you have a Power over Ethernet (PoE) Powered Device (PD) device and Power Sourcing Equipment (PSE) but the two don't seem to be working together? Well you may have fallen down the not-so-well-documented rabbit hole of Mode A vs. Mode B.
Taking a closer look 802.3af PoE, you'll find two different sub-standards of 802.3af associated with 10/100 TX Ethernet, namely Mode A and Mode B. The main difference is best illustrated with the following diagrams.
The diagram above shows a PSE supplying power on top of the 10/100 TX data on pins 1,2,3,and 6. PSE providing "phantom" power on the data pairs is known as Mode A, or Endspan.
This diagram above shows a PSE supplying power on pins 4,5, 7, and 8, the "spare" pairs in an Ethernet cable running 10/100 TX. PSE providing power on the spare pairs is known as Mode B, or Midspan. Most "injector" devices (as opposed to full switches) are Mode B.
And here come the "Gotchas." While PDs must support BOTH Mode A and Mode B to be compliant with the 802.3af standard, there is no such requirement for PSEs. There are a few different ways this can lead to mistakes.
1) You might assume your device is 802.3af compliant because it is listed as PoE-capable, but this is not always the case. A recent customer had a PoE-capable Ethernet/IP Rotary Shaft Encoder, but closer datasheet inspection revealed it was not 802.3af compliant. As it turned out, this was only a 4-wire device and therefore incompatible with the customer's existing Mode B PSE. We supplied a Mode A PSE and resolved the issue.
2) In addition to no requirement for PSEs to perform BOTH Mode A and Mode B, many vendors' documentation doesn't state which mode their device utilizes. This makes it challenging to locate the proper equipment for scenarios like the one above.
3) You must confirm specific model numbers and operation. We came across one vendor's PoE switch whose generic datasheet listed Mode A and Mode B and offered two part numbers ending in A and B with PoE Mode as the only difference between the two. Only upon close inspection did we notice the A part number utilized Mode B and the B part number utilized Mode A. Crazy enough?
With the introduction of new Ethernet-enabled devices expanding geometrically, the need to power these devices from standard AC power outlets has become a limiting factor. IP telephones, wireless access points, IP cameras and device servers are examples of devices limited by the need to have an AC power outlet nearby to plug in a DC power adapter. At best, power supply installation and wiring adds labor and results in the mess of extra wiring; worst case, the lack of nearby AC power means devices cannot be installed where they are needed.
In response to this need, IEEE developed IEEE802.3af to standardize a system of supplying low voltage power to networked devices via the communications line. Released in mid-2003, the standard was entitled: Data Terminal Equipment (DTE) Power via Media Dependant Interface (MDI). It is more commonly referred to as Power over Ethernet (POE). PoE supplies power to network devices over the same standard Category 5e (CAT-5) cabling that carries the data. Devices can be installed wherever structured Ethernet wiring is located, without the need for AC power outlets nearby.
The benefits of PoE include increased mobility for end devices, added safety (no AC power involved), simplicity of installation, reliability, security and cost savings. These advantages have led to the development of a variety of new PoE-enabled products (including B&B Electronics’ PES1A and PESV1A RS-232 to Ethernet Converters).
In the PoE scheme, two different types of devices are involved: Power Sourcing Equipment (PSE) and Powered Devices (PD).
PDs are PoE-enabled network end devices equipped to accept low voltage power transmitted over structured Ethernet CAT-5 cabling.
PSEs provide the DC power to PDs. PSEs source up to 12 watts at 48 volts to each PD. A PSE may be an endspan device or a midspan device. An endspan device typically is a network switch enabled to provide PoE power on each port. A midspan device is connected in-line to each end device and adds power to the line.
Power is carried over the cabling using two techniques: Alternative A and Alternative B
Alternative A implements a simplex, or ‘phantom feeding’ method for delivering power to the end device. Power is carried on the same conductors as data. CAT-5 cabling for standard 10BaseT and 100Base-TX Ethernet uses two data/signal pairs connected to pins 1 and 2 and pins 3 and 6 on RJ-45 connectors. Power sourcing equipment superimposes power onto these pins via the center tap of internal signal coupling transformers. In the powered device (PD) the power is derived from these lines using the reverse technique.
Figure 1: Power over Ethernet Alternative A
Alternative B carries power over spare wire pairs in the cable. The power sourcing equipment applies positive voltage to pins 4 and 5. Negative voltage is applied to pins 7 and 8.
Figure 2: Power over Ethernet Alternative B
PDs are designed to accept PoE in either format (Alternative A or B) so that they can receive power regardless of which Alternative is implemented in the PSE. When operating in Alternative A mode PDs automatically adjust for polarity of the power supply voltage. This ensures that the device will operate even if a crossover cable is being used. (This sometimes happens if a Media Dependant Interface (MDI), such as a PC network card, is connected to the PD, which also is wired as a MDI.)
Figure 3: A Powered Device
PDs operate from 48 VDC (approximately) supplied by the PSE, which means they are classified as a Safety Extra Low Voltage (SELV) device. The maximum current supplied by the PSE is about 350 mA. With some loss in the CAT-5 cable, the PD is limited to about 12 watts of power. Maximum allowable cable resistance is 20 ohms.
The two types of PSE (endspan and midspan) are used in different situations. Endspan PSEs typically are used in new installations where a new switch or router is required. Midspan devices more often are added to an existing network to add PoE capabilities. Midspan devices are connected in-line, between the existing switch and the PD. They can be obtained as a single or multiple line units.
On initial power up Power Sourcing Equipment is designed to detect, and only supply power to the network device if it is identified as a PoE-enabled device. When first connected the PSE initiates a detection process that involves low probing voltages to sense the type of PD connected while avoiding damage to non-PoE network devices.
Some Power Sourcing Equipment may also detect the power class of the PD. PDs are assigned to one of the following five classes, depending on the amount of power they require.
PD Power (W)
0.44 to 12.95
0.44 to 3.84
3.84 to 6.49
6.49 to 12.95
If a PD does not support classification it is assigned a default classification of 0, indicating the PSE should supply the maximum needed power. Classification of connected PDs allows the PSE to optimize its power distribution.
PSEs are able to detect several undesirable load conditions, including shorted communications cables, disconnected PDs and the connection of non-PoE devices. When a fault is detected the PSE will disconnect power from the communications line before damage can occur.