The user can connect multiple devices in parallel in what is known as a "stack"; for example, to increase output current capability or reduce device junction temperature. A stack comprises one primary device and one or more secondary devices. During initialization, each device monitors the SYNC_OUT pin to determine if must operate as a primary device or a secondary device:

• If there is a 47kΩ resistor between the SYNC_OUT pin and ground, the device operates as a secondary device.
• If the SYNC_OUT pin is high impedance, the device operates as a primary device.

Figure 8-14 shows the recommended interconnections in a stack of two TPS6287x devices.

Figure 8-14 Two TPS6287x Devices in a Stacked Configuration

The key points to note are:

• All the devices in the stack share a common enable signal, which must be pulled up with a resistance of at least 15kΩ.
• All the devices in the stack share a common power-good signal.
• All the devices in the stack share a common compensation signal.

• All secondary devices must connect a 47kΩ resistor between the SYNC_OUT pin and ground.

• The remote sense pins (VOSNS and GOSNS) of each device must be connected (do not leave these pins floating).
• Each device must be configured for the same switching frequency.
• The primary device must be configured for forced PWM operation (secondary devices are automatically configured for forced PWM operation).
• A stacked configuration can support synchronization to an external clock or spread-spectrum clocking.
• Only the VSEL pin of the primary device is used to set the default output voltage. The VSEL pin of secondary devices is not used and must be connected to ground.
• The SDA and SCL pins of secondary devices are not used and must be connected to ground.
• A stacked configuration uses a daisy-chained clocking signal, in which each device switches with a phase offset of approximately 140° relative to the adjacent devices in the daisy-chain. To daisy-chain the clocking signal, connect the SYNC_OUT pin of the primary device to the MODE/SYNC pin of the first secondary device. Connect the SYNC_OUT pin of the first secondary device to the MODE/SYNC pin of the second secondary device. Continue this connection scheme for all devices in the stack, to daisy-chain them together.
• Hiccup overcurrent protection must not be used in a stacked configuration.

In a stacked configuration, the common enable signal also acts as a SYSTEM_READY signal (see Section 8.3.3). Each device in the stack can pull the EN pin low during device start-up or when a fault occurs. Thus, the stack is only enabled when all devices have completed the start-up sequence and are fault-free. A fault in any one device disables the whole stack for as long as the fault condition exists.

During start-up, the primary converter pulls the COMP pin low for as long as the enable signal (SYSTEM_READY) is low. When the enable signal goes high, the primary device actively controls the COMP pin and all converters in the stack follow the COMP voltage. During start-up, each device in the stack pulls the PG pin low while it initializes. When initialization is complete, each secondary device in the stack sets the PG pin to a high impedance and the primary device alone controls the state of the PG signal. The PG pin goes high when the stack has completed the start-up ramp and the output voltage is within specification. The secondary converters in the stack detect the rising edge of the power-good signal and switch from DCM operation to CCM operation. After the stack has successfully started up, the primary device controls the power-good signal in the normal way. In a stacked configuration, there are some faults that only affect individual devices, and other faults that affect all devices. For example, if one device enters current limit, only that device is affected. But a thermal shutdown or undervoltage lockout event in one device disables all devices through the shared enable (SYSTEM_READY) signal.

Functionality During Stacked Operation

Some device features are not available during stacked operation, or are only available in the primary converter. Table 8-8 summarizes the available functionality during stacked operation.

Fault Handling During Stacked Operation

In a stacked configuration, there are some faults that only affect individual devices and other faults that affect all devices. For example, if one device enters current limit, only that device is affected. A thermal shutdown or undervoltage lockout event in one device disables all devices through the shared enable (SYSTEM_READY) signal. Table 8-9 summarizes the fault handling of the TPS6287x devices during stacked operation.