SFFS002 February   2023 TCA6408A-Q1

 

  1.   Trademarks
  2. 1Overview
  3. 2Functional Safety Failure In Time (FIT) Rates
  4. 3Failure Mode Distribution (FMD)
  5. 4Pin Failure Mode Analysis (Pin FMA)

Pin Failure Mode Analysis (Pin FMA)

This section provides a Failure Mode Analysis (FMA) for the pins of the TCA6408A-Q1. The failure modes covered in this document include the typical pin-by-pin failure scenarios:

  • Pin short-circuited to Ground (see Table 4-2)
  • Pin open-circuited (see Table 4-3)
  • Pin short-circuited to an adjacent pin (see Table 4-4)
  • Pin short-circuited to supply (see Table 4-5)

Table 4-2 through Table 4-5 also indicate how these pin conditions can affect the device as per the failure effects classification in Table 4-1.

Table 4-1 TI Classification of Failure Effects
ClassFailure Effects
APotential device damage that affects functionality
BNo device damage, but loss of functionality
CNo device damage, but performance degradation
DNo device damage, no impact to functionality or performance

Figure 4-1 shows the TCA6408A-Q1 pin diagram. For a detailed description of the device pins please refer to the Pin Configuration and Functions section in the TCA6408A-Q1 data sheet.

Figure 4-1 Pin Diagram

Following are the assumptions of use and the device configuration assumed for the pin FMA in this section:

  • None

Table 4-2 Pin FMA for Device Pins Short-Circuited to Ground
Pin NamePin No.Description of Potential Failure Effect(s)Failure Effect Class
Vcci1System short to GND. No damage to device expected.B
ADDR2Potential functionality issue if ADDR is referenced to Vcc through a pull up resistor. On a system level, if ADDR is tied to Vcc through a resistor, then additional leakage current will occur through the resistor. On a system level, the I2C address of the device would change and the controller would not be able to communicate with the device. There would also be potential for signal integrity as there may be an address conflict on the bus.

If ADDR is tied to GND via pull down resistor, then functionality is okay if ADDR were shorted to GND.

B
#RESET3Functionality issue as the device would be held in the RESET state. On a system level, the controller would not be able to communicate with the device. The device would NACK its own address. B
P04Worst case situation: Pin is configured to be an output HIGH. A short to GND would cause high current draw from the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOH parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to GND, the pin never changes states and would not assert the INT function.

A
P15Worst case situation: Pin is configured to be an output HIGH. A short to GND would cause high current draw from the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOH parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to GND, the pin never changes states and would not assert the INT function.

A
P26Worst case situation: Pin is configured to be an output HIGH. A short to GND would cause high current draw from the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOH parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to GND, the pin never changes states and would not assert the INT function.

A
P37Worst case situation: Pin is configured to be an output HIGH. A short to GND would cause high current draw from the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOH parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to GND, the pin never changes states and would not assert the INT function.

A
GND8No expected issue.D
P49Worst case situation: Pin is configured to be an output HIGH. A short to GND would cause high current draw from the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOH parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to GND, the pin never changes states and would not assert the INT function.

A
P510Worst case situation: Pin is configured to be an output HIGH. A short to GND would cause high current draw from the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOH parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to GND, the pin never changes states and would not assert the INT function.

A
P611Worst case situation: Pin is configured to be an output HIGH. A short to GND would cause high current draw from the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOH parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to GND, the pin never changes states and would not assert the INT function.

A
P7

12

Worst case situation: Pin is configured to be an output HIGH. A short to GND would cause high current draw from the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOH parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to GND, the pin will never change states and would not assert the INT function.

A
#INT

13

Functionality lost as #INT is always be asserted. If controller is monitoring the #INT pin, the controller would always see #INT asserted (which may be false).

Device damage is not expected though leakage current would be seen at a system level due to a path from Vcc to GND through the pull up resistor tied to #INT.

B
SCL14Functionality lost as SCL is stuck as a logic low. On a system level, this would lock the I2C bus. I2C communication would no longer work.B
SDA15Functionality lost as SDA remains a logic low. On a system level, this would lock the I2C bus. I2C communication would no longer work. B
Vccp16System short to GND. No damage to device expected.

Potential chance for SDA to be locked low. Please check datasheet's power supply recommendation section.

B
Table 4-3 Pin FMA for Device Pins Open-Circuited
Pin NamePin No.Description of Potential Failure Effect(s)Failure Effect Class
Vcci1Functionality lost. Device will not respond to its address, thus is not able to be configured or communicated with.B
ADDR2Functionality may be lost. Device address may float meaning the I2C address of the device float to an unexpected value and the controller would not be able to communicate with the device. There would also be potential for signal integrity as there may be an address conflict on the bus if it floated to the opposite logic level of the external resistor reference.B
#RESET3Functionality may be lost. If #RESET floats low, the device would be held in the RESET state. In this state, the device would not ACK its address would set all registers back to the default state.B
P04Functionality lost.

When pin is configured as an INPUT, any input changes would not be registered. If the input floats, the INT could falsely assert. Any input reads could provide incorrect values.

When the pin is configured as an OUTPUT, the output state would not change at the pin. On a system level, the bias voltage at the pin would just float.

B
P15Functionality lost.

When pin is configured as an INPUT, any input changes would not be registered. If the input floats, the INT could falsely assert. Any input reads could provide incorrect values.

When the pin is configured as an OUTPUT, the output state would not change at the pin. On a system level, the bias voltage at the pin would just float.

B
P26Functionality lost.

When pin is configured as an INPUT, any input changes would not be registered. If the input floats, the INT could falsely assert. Any input reads could provide incorrect values.

When the pin is configured as an OUTPUT, the output state would not change at the pin. On a system level, the bias voltage at the pin would just float.

B
P37Functionality lost.

When pin is configured as an INPUT, any input changes would not be registered. If the input floats, the INT could falsely assert. Any input reads could provide incorrect values.

When the pin is configured as an OUTPUT, the output state would not change at the pin. On a system level, the bias voltage at the pin would just float.

B
GND8Damage could occur since the GND is used to bias substrates of FETs, power at pins while GND is floating may result in current to unexpected paths in device. A
P49Functionality lost.

When pin is configured as an INPUT, any input changes would not be registered. If the input floats, the INT could falsely assert. Any input reads could provide incorrect values.

When the pin is configured as an OUTPUT, the output state would not change at the pin. On a system level, the bias voltage at the pin would just float.

B
P510Functionality lost.

When pin is configured as an INPUT, any input changes would not be registered. If the input floats, the INT could falsely assert. Any input reads could provide incorrect values.

When the pin is configured as an OUTPUT, the output state would not change at the pin. On a system level, the bias voltage at the pin would just float.

B
P611Functionality lost.

When pin is configured as an INPUT, any input changes would not be registered. If the input floats, the INT could falsely assert. Any input reads could provide incorrect values.

When the pin is configured as an OUTPUT, the output state would not change at the pin. On a system level, the bias voltage at the pin would just float.

B
P712Functionality lost.

When pin is configured as an INPUT, any input changes would not be registered. If the input floats, the INT could falsely assert. Any input reads could provide incorrect values.

When the pin is configured as an OUTPUT, the output state would not change at the pin. On a system level, the bias voltage at the pin would just float.

B
#INT13Functionality lost. #INT never asserts. On a system level, if controller is using #INT to determine input changes, it is missed.B
SCL14Functionality lost.

Device state machine may not see SCL toggle. Device would NACK its own address and would not be configurable.

B
SDA15Functionality lost.

Device state machine may not see SDA toggle. Device would NACK its own address and would not be configurable.

B
Vccp16Functionality lost. Device does not respond to its address, thus is not able to be configured or communicated with.B
Table 4-4 Pin FMA for Device Pins Short-Circuited to Adjacent Pin
Pin NamePin No.Shorted toDescription of Potential Failure Effect(s)Failure Effect Class
Vcci1ADDRPotential functionality issue if ADDR is referenced to GND through a pull-down resistor. On a system level, if ADDR is tied to GND through a resistor, then additional leakage current will occur through the resistor. On a system level, the I2C address of the device would change and the controller would not be able to communicate with the device. There would also be potential for signal integrity as there may be an address conflict on the bus.

If ADDR is tied to VCC via pull up resistor, then functionality is okay if ADDR were shorted to Vcc.

B
ADDR2#RESET#RESET is typically tied to a resistor referencing Vcc.

Potential functionality issue if ADDR is referenced to GND through a pull-down resistor. Both the ADDR and #RESET pins will form a voltage divider between the pull up and pull down resistors. There is a potential for the device to float between the reset state and normal operation state. Any noise which couples to this node would also toggle the reset state. On a system level, the device may intermittently reset or could remain in reset depending on the node's voltage. Another possibility is the device is not in reset but the device's configurable address could change from the intended logic 0 to a logic 1 due to the voltage divider. On a system level, this could result in the device NACK'ing to the expected address or cause signal integrity issues due to a potential in conflicting I2C addresses. If the latter case is true, writes to the device could unintentionally set the p-port to be outputs or inputs when they were not intended to be.

If ADDR is tied to VCC via pull up resistor, then funtionality-wise, the device should operate correctly if ADDR were shorted to #RESET.

B
#RESET3P0If P0 is an INPUT (by default) and referenced with an external pull-up resistor then no functionality issues are expected.

If P0 is an INPUT (by default) and referenced with an external pull-down resistor then P0 and #RESET form a voltage divider, and the voltage at these pins settle somewhere mid rail if pull-up and pull-down resistors are equal in value. #RESET and P0 may be in an unknown state because they are not above/below VIH /VIL levels. Noise coupling onto these pins may toggle reset. Assume functionality is lost.

If P0 is configured to be an OUTPUT HIGH, device would NOT be reset and functionality should be okay. Some additional leakage may be present.

If P0 is configured to be an OUTPUT LOW, device would be reset and then all GPIOs would return to their default power up state of an input. Functionality would be lost if P0 is ever configured to be an output LOW.

B
P04P1If P0 and P1 are OUTPUTs but logic levels are not the same (one is OUTPUT LOW and other is OUTPUT HIGH), then contention between the two will occur. If IoH exceeds 50mA then damage may occur but may not be instantaneous.

Failures overtime may occur.

If P0 and P1 are OUTPUTs and logic levels are the same (both are OUTPUT HIGH or OUTPUT LOW), damage is not expected but some leakage current may occur.

If P0 and P1 are both INPUTs, damage is not expected.

If P0 and P1 are configured such that the pair is an OUTPUT and an INPUT, then no damage is expected.

A
P15P2

If P1 and P2 are OUTPUTs but logic levels are not the same (one is OUTPUT LOW and other is OUTPUT HIGH), then contention between the two will occur. If IoH exceeds 50mA then damage may occur but may not be instantaneous.

Failures overtime may occur.

If P1 and P2 are OUTPUTs and logic levels are the same (both are OUTPUT HIGH or OUTPUT LOW), damage is not expected but some leakage current may occur.

If P1 and P2 are both INPUTs, damage is not expected.

If P1 and P2 are configured such that the pair is an OUTPUT and an INPUT, then no damage is expected.

A
P26P3

If P2 and P3 are OUTPUTs but logic levels are not the same (one is OUTPUT LOW and other is OUTPUT HIGH), then contention between the two will occur. If IoH exceeds 50mA then damage may occur but may not be instantaneous.

Failures overtime may occur.

If P2 and P3 are OUTPUTs and logic levels are the same (both are OUTPUT HIGH or OUTPUT LOW), damage is not expected but some leakage current may occur.

If P2 and P3 are both INPUTs, damage is not expected.

If P2 and P3 are configured such that the pair is an OUTPUT and an INPUT, then no damage is expected.

A
P37GND

If P3 is set to an OUTPUT LOW, no functionality or damage is expected.

If P3 is an INPUT, no damage is expect but functionality will be lost as P3 will likely not see a logic HIGH and set the INT.

If P3 is an OUTPUT HIGH, damage is likely to occur as a large IoH current will flow from the pin and likely exceed the VoH limit.

A
P49P5

If P4 and P5 are OUTPUTs but logic levels are not the same (one is OUTPUT LOW and other is OUTPUT HIGH), then contention between the two will occur. If IoH exceeds 50mA then damage may occur but may not be instantaneous.

Failures overtime may occur.

If P4 and P5 are OUTPUTs and logic levels are the same (both are OUTPUT HIGH or OUTPUT LOW), damage is not expected but some leakage current may occur.

If P4 and P5 are both INPUTs, damage is not expected.

If P4 and P5 are configured such that the pair is an OUTPUT and an INPUT, then no damage is expected.

A
P510P6

If P5 and P6 are OUTPUTs but logic levels are not the same (one is OUTPUT LOW and other is OUTPUT HIGH), then contention between the two will occur. If IoH exceeds 50mA then damage may occur but may not be instantaneous.

Failures overtime may occur.

If P5 and P6 are OUTPUTs and logic levels are the same (both are OUTPUT HIGH or OUTPUT LOW), damage is not expected but some leakage current may occur.

If P5 and P6 are both INPUTs, damage is not expected.

If P5 and P6 are configured such that the pair is an OUTPUT and an INPUT, then no damage is expected.

A
P611P7If P6 and P7 are OUTPUTs but logic levels are not the same (one is OUTPUT LOW and other is OUTPUT HIGH), then contention between the two will occur. If IoH exceeds 50 mA, then damage may occur but may not be instantaneous.

Failures overtime may occur.

If P6 and P7 are OUTPUTs and logic levels are the same (both are OUTPUT HIGH or OUTPUT LOW), damage is not expected but some leakage current may occur.

If P6 and P7 are both INPUTs, damage is not expected.

If P6 and P7 are configured such that the pair is an OUTPUT and an INPUT, then no damage is expected.

A
P712#INTIf P7 is an INPUT then P7 will mirror #INT status.

If P7 is an OUTPUT LOW, the #INT will get stuck low. On a system level, the controller would not be able to deassert #INT by reading the input port. If sw polls INT, then the controller is stuck reading the device.

If P7 is an OUTPUT HIGH, when #INT is asserted, contention will occur between the P7 and the #INT's NFET. If the IOL exceeds 25 mA, damage may occur.

A
#INT13SCLFrom a device standpoint, no functionality loss or damage expected from this.

From a system level, this may cause problems to the controller if it is monitoring #INT.

The worst case issue is if #INT is asserted, the I2C bus would get stuck low unless #INT deasserts which may not happen depending on how the input that changed is set up.

B

SCL14SDAI2C communication is lost both to the device and to the system's I2C bus. Functionality will be lost but no damage expected.B
SDA15VccpDevice is likely to be damaged during ACKs and read transaction due to large excessive current through pin. If IoL exceeds 6 mA at 85°C or less, device may be damaged.

Damage may not be instantaneous, but may occur over time.

VOL from device may also be too large for the controller to accept as a valid low during ACKs.

A
Table 4-5 Pin FMA for Device Pins Short-Circuited to supply
Pin NamePin No.Description of Potential Failure Effect(s)Failure Effect Class
Vcci1If supply voltage is the same voltage as Vcci then no issue is expected.

If supply voltage is not the same voltage as Vcci and is within absolute maximum specification then no damage is expected to the device. On a system level, high current may occur due to a short between two voltage rails.

If supply voltage is larger than the absolute maximum specification for the device, damage could occur.

A
ADDR2Potential funtionality issue if ADDR is referenced to GND through a pull-down resistor. On a system level, if ADDR is tied to GND through a resistor, then additional leakage current occurs through the resistor. On a system level, the I2C address of the device would change and the controller would not be able to communicate with the device. There would also be potential for signal integrity as there may be an address conflict on the bus.

If ADDR is tied to VCC via pull up resistor, then functionality is okay if ADDR were shorted to Vcc.

B
#RESET3Device would maintain functionality though from a system level, the controller would not be able to drive the #RESET low to reset the device. Any attempts to do so could damage the controller due to high current from Vcc to GND from the low side driver of the controller.D
P04Worst case situation: Pin is configured to be an output LOW. A short to supply would cause high current draw to the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOL parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to supply, the pin never changes states and would not assert the INT function.

A
P15Worst case situation: Pin is configured to be an output LOW. A short to supply would cause high current draw to the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOL parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to supply, the pin never changes states and would not assert the INT function.

A
P26Worst case situation: Pin is configured to be an output LOW. A short to supply would cause high current draw to the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOL parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to supply, the pin never changes states and would not assert the INT function.

A
P37Worst case situation: Pin is configured to be an output LOW. A short to supply would cause high current draw to the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOL parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to supply, the pin never changes states and would not assert the INT function.

A
GND8Short between GND and Supply. Damage to device is not expected but may cause damage on a system level.

Functionality is not guaranteed under these conditions.

B
P49Worst case situation: Pin is configured to be an output LOW. A short to supply would cause high current draw to the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOL parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to supply, the pin will never change states and would not assert the INT function.

A
P510Worst case situation: Pin is configured to be an output LOW. A short to supply would cause high current draw to the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOL parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to supply, the pin will never change states and would not assert the INT function.

A
P611Worst case situation: Pin is configured to be an output LOW. A short to supply would cause high current draw to the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOL parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to supply, the pin never changes states and would not assert the INT function.

A
P712Worst case situation: Pin is configured to be an output LOW. A short to supply would cause high current draw to the pin due to a path from Vcc to GND. Potential damage to device may occur due to violation of absolute maximum IOL parameter.

Functionality may be affected if the pin is configured as an input and the controller is monitoring the INT. In the case of a short to supply, the pin never changes states and would not assert the INT function.

A
#INT13

Damage to device may occur as the IoL through the pin may exceed the absolute maximum specification for the device if #INT asserts.

A
SCL14

Damage to device is not expected because the SCL pin of the device does not drive low (does not clock stretch). On a system level, functionality could be lost due to large VoL shifts because of the large IoL on the bus from the short to supply. The controller's VoL may exceed the ViL of the device. The large IoL may also damage the controller's SCL driver if the IoL is larger than what the controller is spec'd for.

B
SDA15Damage to device may occur as the IoL through the pin may exceed the absolute maximum specification for the device when the device's SDA ACKs and when it sends data back

To the controller.

A
Vccp16If supply voltage is the same voltage as Vccp then no issue is expected.

If supply voltage is not the same voltage as Vcci and is within absolute maximum specification, then no damage is expected to the device. On a system level, high current may occur due to a short between two voltage rails.

If supply voltage is larger than the absolute maximum specification for the device, damage could occur.

A