User’s Manual: SW-10200
38
C.2.2. Classification
Once a PD is detected, the PSE may optionally perform classification to determine the maximal power a PD is to
consume. The PSE induces 15.5-20.5 VDC, limited to 100 mA, for a period of 10 to 75 ms responded by a certain current
consumption by the PD, indicating its power class.
The PD is assigned to one of 5 classes: 0 (default class) indicates that full 15.4 watts should be provided, 1-3 indicate
various required power levels and 4 is reserved for future use. PDs that do not support classification are assigned class 0.
Special care must be employed in the definition of class thresholds, as classification may be affected by cable losses.
Classifying a PD according to its power consumption ideally helps a PoE system in optimizing its power distribution, so
that efficient power management based on classification results may reduce total system costs.
C.2.3. Start-up
Once line detection and optional classification stages are completed, the PSE must switch from low voltage to its full
voltage capacity (44-57 Volts) over a minimal amount of time (above 15 microseconds). A gradual startup is required to
avoid a sudden rise in voltage (reaching high frequencies) would introduce noise on the data lines. Once provision of power
is initiated, it i
s common for an inrush current to be experienced at the PSE port, due to the PD’s input capacitance. A PD
must be designed to cease inrush current consumption (of over 350 mA) within 50 ms of power provision startup.
C.2.4. Operation
During normal operation, the PSE provides 44-57 VDC, able to support a minimum of 15.4 watts power.
C.2.5. Power Overloads
The IEEE 802.3af standard defines the handling of overload conditions. In the event of an overload (a PD drawing a
higher power level than the allowed 12.95 Watts), or an outright short circuit caused by a failure in cabling or in the PD, the
PSE must shut down power within 50 to 75 milliseconds, while limiting current drain during this period to protect the cabling
infrastructure. Immediate voltage drop is avoided to prevent shutdown due to random fluctuations.
C.3. Power Disconnection Scenarios
The IEEE 802.3af standard requires that devices powered over Ethernet be disconnected safely (i.e. power needs be
shut down within a short period of time following disconnection of a PD from an active port). When a PD is disconnected,
there is a danger that it will be replaced by a non-PoE-ready device while power is still on to avoid damaging the device and
risking injury.
The standard defines two means of disconnection, DC Disconnect and AC Disconnect, both of which provide the same
functionality - the PSE shutdowns power to a disconnected port within 300 to 400ms. The upper boundary is a physical
human limit for disconnecting one PD and reconnecting another.
DC Disconnect
DC Disconnect detection involves measurement of current. A disconnected PD stops consuming current, which can be
inspected by the PSE. The PSE must therefore disconnect power within 300 to 400 ms from the current flow stop. The lower
time boundary is important to prevent shutdown due to random fluctuations.
AC Disconnect
This method is based on the fact that when a valid PD is connected to a port, the AC impedance measured on its
terminals is significantly lower than in the case of an open port (disconnected PD).
AC Disconnect detection involves the induction of low AC signal in addition to the 48 VDC operating voltage. The
returned AC signal amplitude is monitored by the PSE at the port terminals. During normal operation, the PD's relatively low
impedance lowers the returned AC signal while a sudden disconnection of this PD will cause a surge to the full AC signal
level and will indicate PD disconnection.
