Hi
What is the reset sequence of the Thor T5000? Does the SYS_RESET_N signal get pulled low after a hot reset (our tests show it does not)? What signals are pulled low during a hot reset?
- When I pull L on SYS-REET-N, the system will shut down directly. The measurement shows that SYS-REET-N will always be in the ‘L’ state. Is this behavior normal?
- If the above behavior is normal, how can we implement button control system reboot? Is the Thor Dev kit implemented using MCU?
It sounds like you asserted reset by setting PERIPHERAL_RESET_N low since you did not see SYS_RESET_N go low.
Signal descriptions from Table 5-1 of the Jetson Thor Series Modules Interface Design Guide (DG):
Figure 5-1 shows the circuits on the module:
Asserting either SYS_RESET_N or PERIPHERAL_RESET_N will result in resetting the SoC and boot devices. The typical source for SYS_RESET_N is the power sequencer. As mentioned in 5.1.1 of the DG:
SYS_RESET_N is driven by the power sequencer on the module during power-on. It
does not need to be controlled by the carrier board. If the carrier board supplies
required for powering on require additional time, the PERIPHERAL_RESET_N signal can
be held low. This will keep the SoC and other boot devices in reset.
SYS_RESET_N alone should not result in a full system shutdown. SYS_RESET_N only resets the SoC and boot devices as described above in the DG. Only setting MODULE_POWER_ON low should shut down. If MODULE_POWER_ON remains high and SYS_RESET_N is driven low by the carrier to issue a reset, the pull up on the SYS_RESET_N should pull the signal high after the carrier stops driving low since the power sequencer will not re-sequence the power rails it controls. The power sequencer would only keep SYS_RESET_N low if it is in the middle of executing its power up sequence. So seeing SYS_RESET_N stuck low when MODULE_POWER_ON is high is unexpected.
Or, in the SW shut-down sequence, how long does it take for MODULE-POWER-ON to be pulled L after SYS-REET-N is pulled L (the above issue may be caused by SYS-REET-N and MODULE-POWER-ON being pulled L almost simultaneously)
MODULE_POWER_ON is an input to the Thor module, driven by the carrier board, so the MODULE_POWER_ON timing depends on your implementation.
The previously described statement ‘If MODULE.POWER_SON remains high and SYS-REET-N is driven low by the carrier to issue a reset, the pull up on the SYS-REET-N should pull the signal high after the carrier stops driving low since the power sequencer will not re sequence the power rails it controls’ is not specified in the SW shut-down sequence of the Design Guide.
- In the Sequence of SW shutdown, should the time for SYS-REET-N pulled down ->MODULE-POWER-ON pulled down need to be delayed?
- May I ask if there will be an action of pulling L and then pulling H when rebooting under the system?
3，Will SYS-REST-N pull L when resetting in the system, and what is the reset sequence of Thor？
Yes there needs to be a delay. When the software shutdown is initiated, an I2C write to the power sequencer starts the power off sequence. As shown in Figure 5-4 of the Design Guide, the first action will be the power sequencer setting SYS_RESET_N low. As part of the power off sequence, hardware on the module will eventually set CARRIER_POWER_ON low. Only after this can the MODULE_POWER_ON signal go low, as noted by the blue text.
During a reboot, the power sequencer will set the SYS_RESET_N low and then drive it high.
If the carrier asserts SYS_RESET_N low, that will immediately reset the SoC and boot devices but not affect the power sequencer, as mentioned in the signal descriptions in Table 5-1.
