LTC4259A-1
25
4259a1fa
Figure 20. Optoisolating the I
2
C Bus
APPLICATIO S I FOR ATIO
WUUU
4258A F20
V
DD
INT
SCL
SDAIN
SDAOUT
AD0
AD1
AD2
AD3
DGND
AGND
LTC4259A-1
BYP
V
DD
INT
SCL
SDAIN
SDAOUT
AD0
AD1
AD2
AD3
DGND
AGND
LTC4259A-1
V
DD
INT
SCL
SDAIN
SDAOUT
AD0
AD1
AD2
AD3
DGND
AGND
LTC4259A-1
V
DD
INT
SCL
SDAIN
SDAOUT
AD0
AD1
AD2
AD3
DGND
AGND
LTC4259A-1
V
DD
INT
SCL
SDAIN
SDAOUT
AD0
AD1
AD2
AD3
DGND
AGND
LTC4259A-1
0.1µF
0.1µF
0.1µF
0.1µF
0.1µF
+
10µF
2k
2k
0.1µF
0.1µF
0.1µF
BYP
0.1µF
BYP
0.1µF
BYP
0.1µF
BYP
0.1µF
200
200
200
200
U2
U3
U1
HCPL-063L
HCPL-063L
V
DD
CPU
SCL
SDA
SMBALERT
GND CPU
U1: FAIRCHILD NC7WZ17
U2, U3: AGILENT HCPL-063L
TO
CONTROLLER
ISOLATED
3.3V
ISOLATED
GND
0100000
0100001
0100010
0101110
0101111
I
2
C ADDRESS
LTC4259A-1
26
4259a1fa
direction. A STOP condition is not used to set up a
REPEATED START condition, for this would clear any data
already latched in. When the master has finished commu-
nicating with the slave, it issues a STOP condition. A STOP
condition is generated by transitioning SDA from low to
high while SCL is high. The bus is then free for communi-
cation with another SMBus or I
2
C device.
Acknowledge
The Acknowledge signal is used for handshaking between
the master and the slave. An Acknowledge (active LOW)
generated by the slave lets the master know that the latest
byte of information was received. The corresponding SCL
clock pulse is always generated by the master. The master
releases the SDA line (HIGH) during the Acknowledge
clock pulse. The slave must pull down the SDA line during
the Acknowledge clock pulse so that it remains a stable
LOW during the HIGH period of this clock pulse. When the
master is reading from a slave device, it is the master’s
responsibility to acknowledge receipt of the data byte in
the bit that follows unless the transaction is complete. In
that case the master will decline to acknowledge and issue
the STOP condition to terminate the communication.
Write Byte Protocol
The master initiates communication to the LTC4259A-1
with a START condition and a 7-bit bus address followed
by the Write Bit (Wr) = 0. If the LTC4259A-1 recognizes its
own address, it acknowledges and the master delivers the
command byte, signifying to which internal LTC4259A-1
register the master wishes to write. The LTC4259A-1 ac-
knowl
edges
and latches the lower five bits of the com-
mand byte into its Register Address register. Only the lower
five bits of the command byte are checked by the LTC4259A-
1; the upper three bits are ignored. The master then deliv-
ers the data byte. The LTC4259A-1 acknowledges once
more and latches the data into the appropriate control
register. Finally, the master terminates the communica-
tion with a STOP condition. Upon reception of the STOP
condition, the Register Address register is cleared (see
Figure 7).
Read Byte Protocol
The master initiates communication from the LTC4259A-
1 with a START condition and the same 7-bit bus address
followed by the Write Bit (Wr) = 0. If the LTC4259A-1
recognizes its own address, it acknowledges and the
master delivers the command byte, signifying which
internal LTC4259A-1 register it wishes to read from. The
LTC4259A-1 acknowledges and latches the lower five bits
of the command byte into its Register Address register. At
this time the master sends a REPEATED START condition
and the same 7-bit bus address followed by the Read Bit
(Rd) = 1. The LTC4259A-1 acknowledges and sends the
contents of the requested register. Finally, the master
declines to acknowledge and terminates communication
with a STOP condition. Upon reception of the STOP
condition, the Register Address register is cleared (see
Figure 8).
Receive Byte Protocol
Since the LTC4259A-1 clears the Register Address regis-
ter on each STOP condition, the interrupt register (register
0) may be read with the Receive Byte Protocol as well as
with the Read Byte Protocol. In this protocol, the master
initiates communication with the LTC4259A-1 with a
START condition and a 7-bit bus address followed by the
Read Bit (Rd) = 1. The LTC4259A-1 acknowledges and
sends the contents of the interrupt register. The master
then declines to acknowledge and terminates communi-
cation with a STOP condition (see Figure 9).
Alert Response Address and the INT Pin
In a system where several LTC4259A-1s share a common
INT line, the master can use the Alert Response Address
(ARA) to determine which LTC4259A-1 initiated the
interrupt.
The master initiates the ARA procedure with a START
condition and the 7-bit ARA bus address (0001100)b
followed by the Read Bit (Rd) = 1. If an LTC4259A-1 is
asserting the INT pin, it acknowledges and sends its 7-bit
bus address (010A
3
A
2
A
1
A
0
)b and a 1 (see Figure 10).
While it is sending its address, it monitors the SDAIN pin
to see if another device is sending an address at the same
APPLICATIO S I FOR ATIO
WUUU
LTC4259A-1
27
4259a1fa
time using standard I
2
C bus arbitration. If the LTC4259A-
1 is sending a 1 and reads a 0 on the SDAIN pin on the
rising edge of SCL, it assumes another device with a lower
address is sending and the LTC4259A-1 immediately
aborts its transfer and waits for the next ARA cycle to try
again. If transfer is successfully completed, the LTC4259A-
1 will stop pulling down the INT pin. When the INT pin is
released in this way or if a 1 is written into the Clear
Interrupt pin bit (bit 6 of register 1Ah), the condition
causing the LTC4259A-1 to pull the INT pin down must be
removed before the LTC4259A-1 will be able to pull INT
down again. This can be done by reading and clearing the
event registers or by writing a 1 into the Clear All Interrupts
bit (bit 7 of register 1Ah). The state of the INT pin can only
change between I
2
C transactions, so an interrupt is cleared
or new interrupts are generated after a transaction com-
pletes and before new I
2
C bus communication com-
mences. Periodic polling of the alert response address can
be used instead of the INT pin if desired. If any device
acknowledges the alert response address, then the INT
line, if connected, would have been low.
System Software Strategy
Control of the LTC4259A-1 hinges on one decision, the
LTC4259A-1’s operating mode. The three choices are
described under Operating Modes. In Auto mode the
LTC4259A-1 can operate autonomously without direc-
tion from a host controller. Because LTC4259A-1s run-
ning in Auto mode will power every valid PD connected to
them, the PSE must have 15.4W/port available. To reduce
the power requirements of the –48V supply, PSE systems
can track power usage, only turning on ports when
sufficient power is available. The IEEE describes this as a
power allocation algorithm and places two limitations: the
PSE shall not power a PD unless it can supply the
guaranteed power for that PD’s class (see Table 2) and
power allocation may not be based solely on a history of
each PD’s power consumption. In order for a PSE to
implement power allocation, the PSE’s processor/con-
troller must control whether ports are powered—the
LTC4259A-1 cannot be allowed to operate in Auto mode.
Semiauto mode fits the bill as the LTC4259A-1 automati-
cally detects and classifies PDs, then makes this informa-
tion available to the host controller, which decides to
APPLICATIO S I FOR ATIO
WUUU
apply power or not. Operating the LTC4259A-1 in Manual
mode also lets the controller decide whether to power the
ports but the controller must also control detection and
classification. If the host controller operates near the limit
of its computing resources, it may not be able to guide a
Manual mode LTC4259A-1 through detect, class and port
turn-on in less than the IEEE mandated maximum of
950ms.
In a typical PSE, the LTC4259A-1s will operate in Semiauto
mode as this allows the controller to decide to power a
port without unduly burdening the controller. With an
interrupt mask of F4h, the LTC4259A-1 will signal to the
host after it has successfully detected and classed a PD,
at which point the host can decide whether enough power
is available and command the LTC4259A-1 to turn that
port on. Similarly, the LTC4259A-1 will generate inter-
rupts when a port’s power is turned off. By reading the
LTC4259A-1’s interrupt register, the host can determine
if a port was turned off due to overcurrent (t
START
or t
ICUT
faults) or because the PD was removed (Disconnect
event). The host then updates the amount of available
power to reflect the power no longer consumed by the
disconnected PD. Setting the MSB of the interrupt mask
causes the LTC4259A-1 to communicate fault conditions
caused by failures within the PSE, so the host does not
need to poll to check that the LTC4259A-1s are operating
properly. This interrupt driven system architecture pro-
vides the controller with the final say on powering ports
at the same time, minimizing the controller’s computation
requirements because interrupts are only generated when
a PD is detected or on a fault condition.
The LTC4259A-1 can also be used to power older powered
Ethernet devices that are not 802.3af compliant and may
be detected with other methods. Although the LTC4259A-
1 does not implement these older detection methods
automatically, if software or external circuitry can detect
the noncompliant devices, the host controller may com-
mand the LTC4259A-1 to power the port, bypassing IEEE
compliant detection and classification and sending power
to the noncompliant device.

LTC4259AIGW-1#TRPBF

Mfr. #:
Manufacturer:
Analog Devices Inc.
Description:
Power Switch ICs - POE / LAN 4x IEEE 802.3af Pwr over E Cntr w/ AC Di
Lifecycle:
New from this manufacturer.
Delivery:
DHL FedEx Ups TNT EMS
Payment:
T/T Paypal Visa MoneyGram Western Union

Products related to this Datasheet