Depending on the power supply source, TX 1 may not start.

Hello. nVIDIA staff

TX 1 may not start depending on the power supply source, which is troubling.

It occurred during product start-up test.
For the power supply system 1, start up 100%.
In the case of the power supply system 2, only about 50% is normally activated.

· It is a custom board.
· It adopts Auto-Power_ON mode.
· The power supply of the carrier board is 100% normal.
· The power supply of the carrier board is on, and RESET is negated.
· There is no log output on UART 0.

Is there something the TX 1 does not move while the power management IC is running?
What is the cause?
Please let me know if there are regulations that I must not protect.

that’s all.

Hi, what’s the difference between system 1 and system 2 ? You mean no problem with power supply system 1?


Power system 1: AC-DC 12 V (VDD_IN: 12 V) OK
Power system 2: PoE + (VDD_IN: 10.5 V) NG

In both cases the voltage is stable and does not look like there is a problem.

that’s all.

additional information.

If RESET_IN is asserted ‘L’ while TX 1 does not output logs when the power supply of carrier board is turned on, TX 1 starts normally.

I think there is a critical timing or voltage rise somewhere, but I do not know what I do not like.

that’s all.

Suggest to measure the signals to check if they are same as power up sequence of OEM DG, basically if the VDD_IN input is stable, there should be no such issue if no other difference between your two systems.


Thanks for your advice.
I will try to measure it again.
However, what is stated in OEM_DG is the measured value at P2597, not the standard of Jetson_TX1.
Should not we show standard values even if we review it?

P2597 is the carrier board of Jetson TX1.

There is no typ. In the standard value.
min or max

Table 7 list the typ value of each phase, min or max is not must for most phases.


I tried to observe the waveform.

t2-3: 105 [ms]] (VDD_IN ramp to VIN_PWR_BAD #: ‘H’)
t2-4: 330 [ms] (VDD_IN ramp to CHARGER_PRSNT #: ‘L’)
t4-6: 8 [ms] (CHARGER_PRSNT #: ‘L’ to CARRIER_POWER_ON: ‘H’)
t6 - 8: 150 [ms] (CARRIER_POWER_ON: ‘H’ to RESET_OUT #: ‘H’)

However, Tegra_X1 may not start.

A voltage detection IC is connected to RESET_OUT #.
Since the rise of 1.8 V is fast, the delay time setting is set to 5 [ms].
Therefore, the above 150 [ms] is almost the assertion time of TX 1.

Started when delay time setting is extended. (100% start up so far)
Is there a minimum time to assert RESET_OUT # on the board side?

Please check the note 6 of table 7, the min time of RESET_OUT# is same RESET_IN#.

“6. User Dependent if RESET_IN# connected to button. Otherwise, carrier board dependent. Not shown in Power up
sequence figure.”

Is this description?
Where is it written that RESET_IN # and RESET_OUT # are the same?

the min time of RESET_OUT# is same RESET_IN#

Do you say that RESET_OUT # requires more than 50 [ms]?
After setting CARRIER_PWR_ON: ‘H’ for RESET_OUT #, TX1 asserts as much as 150 [ms].

RESET_OUT# is connected to RESET_IN# thru a diode. The min assertion time is 50 ms.

What’s this delay time? “Since the rise of 1.8 V is fast, the delay time setting is set to 5 [ms].”

OEM_DG is not written as asserting on the carrier board side as stipulated by RESET_OUT #.
The minimum value of RESET_OUT # is also not described.
Only the regulation for 1.8 V is written.

Observe that RESET_OUT # (min) = RESET_IN # (min).
Is it better than 50 [ms]?

Again, what’s this 5ms delay time? “Since the rise of 1.8 V is fast, the delay time setting is set to 5 [ms].”

Hi BlueYatagarasu,

Have you clarified the cause and got the device working normally?
Any further support required?


The TX1 takes a hefty bit of current to power on, if you use a supply that’s not up to the task you might see the fan twitch but nothing else. We’ve seen that quite a few times during testing if we try to use too small a regulator.