Hi all,
i use FPGA to control the TX2 modules’s powering-up in my custom carrier board,
FPGA will do the following:
A. assert VIN_PWR_BAD_L for 500ms, then deassert this signal to notiy the PMIC power is ready.
B. after the deassertion of VIN_PWR_BAD_L, pulsing the POWER_BTN_L low for about 300ms (to simulate a on key press).
C. all the other power-up specific pins (like RESET_OUT_L, RESET_IN_L,etc ) are kept disconnected with my custom board
Note that there is a little difference between my custom carrier board and the NVIDIA’s developer’s TX2 kit
i.e. :
some pins of TX2 modules are connected to other ICs on my carrier board, like:
FAN_PWM ( pulled up by VDD_1V8 from my carrier board)
HDMI_DDC_SCL_POL, HDMI_DDC_SDA_POL, HDMI_HPD ( connected to a HDMI--->VGA converter IC)
CAN_WAKE, UART1/2 TX/RX (connected to CAN & RS422 transceiver)
...
however, these ICs are already powered up before TX2,
that is to say:
[b]the above pins (FAN_PWM, HDMI_DDC_SCL_POL, HDMI_DDC_SDA_POL, HDMI_HPD,CAN_WAKE, UART1/2 TX/RX, ... )
may have sensed voltages ( 3.3V/1.8V )[/b]
the OEM design guide says that TX2 must be powered up before
the carrier board…
BUT, i THINK this should not affect the normal power-up of TX2 module, since
FPGA has already driven VIN_PWR_BAD_L for the initial 500ms ( which “lies” to TX2 that power is not ready )
Is my understanding correct ???
Appendix:
As FPGA completes the Power up sequence, as described earlier.
we found that TX2’s UART-0 can only work in “echo” mode,
i.e., it simply bounces your input:
in SecureCRT (a UART utility run in windows ), if you enter ‘A’, TX2’s UART-0 will
feedback ‘A’ and it will be displayed in the console of SecureCRT
<b>but TX2 itself does not dump any log</b>
Shall carrier board be powered up later than the TX2 module in a strict manner ??
What if this rule is violated ?
will it actually affect the TX2’s boot up process or
just in case of potential in-rush current to TX2’s pin (only for better electrical protection)