Bring-up - Sequence, with no SOM?

Autonomous Machines Jetson & Embedded Systems Jetson AGX Xavier
1

Hi,

I just got my boards from the assembly house, and they are not powering-up, so I have to debug.
The power design is with SRC0GS22D.
I see that there’s VCC_SRC_FET, VCC_SRC (~14V), and 5V_AO (5V).
The current consumption is 0 in 4 of 5 boards I got. Pressing the power-switch has no effect.

One board turns on the VDD_5V with no power-button press, and the current is 90mA.

In order to debug, I got to access the power-control signals, but most of them are hidden by the SOM.
Instead of taking wires out, for debug, I would like to remove the SOM, and simulate any signal it is driving.

My understanding is that I only need to drive CARRIER_PWR_ON high.

Is this correct?

Any other suggestion for the debug?
Which signal may cause the board to power-up with no switch-press?

2

Hi, as you can see in OEM DG, carrier board give power supply to module and generate module_power_on to module after power button pressed so as to power on module, and then module will give carrier_pwr_on to carrier board to power up rails of carrier board. The AC_OK_L is used for auto power on function, for more detail info please refer to OEM DG and reference schematic.

3

Thanks for this useful information.

And for those who will try to debug the power-sequence,
I’ll suggest to try shorting carrier_pwr_on to VDD_5V
(not sure - might need to short to higher level - will try and report.)

4

Sorry, I had a typo above - shorted CARRIER_PWR_ON to 5V_AO - not VDD_5V.

Still, the power is not on.

Probing VIN_PWR_BAD_N, it is low.
(Xavier Data Sheet 0.9 shows that it should be high).

In first debug of this signal, I see that there’s TPS3808 which monitors VCC_SRC, which is low.

I’m going to check the reference-design schematic - maybe I had something wrong in my.

5

For now, I’ve left the debug of VIN_PWR_BAD_n, and finally managed to power-up by doing several bypasses:

Today, I’ve found that CARRIER_PWR_ON can be pulled high, by assembly of the option resistor R504 (on 3.3v circuitry).

  1. The D-FF in the ‘simplified button power-on circuitry’ was forcing the BUTTON_POWER_ON# to low, so I’ve disconnected the resistor on the input to the buffer, and shorted to GND.
    Now, BUTTON_POWER_ON# is high, and going-low when I press the power-switch.
  2. I have disconnected PSHOLD input of the SRC0, and shorted to 5V_AO - this bypass CARRIER_PWR_ON and VDDIN_PWR_BAD_n, so the power is keeping on after release of the power switch.

Now I have the VDD_5V on, and will go back to debug the bypasses.

6

I’ve completed the first stage of the bring-up, so here’s my suggestions to the forum:

On first boards, assemble R504 with ~1.5K to set CARRIER_PWR_ON to high level, which will enable the 3.3v and 1.8V.
Note that the ref-design got 20K, which yield a lower-level, not enabling the 1.8 supply)

You should verify having VCC_SRC_FET, then VCC_SRC, then VDD_5V /VDD_3V3 / VDD_1V8.

Once you’ve completed this debug, remove R504, so the 3.3 and 1v8 will not be powered before time.

Note: It might be needed to assemble R17, to force 0 on CVM_PRSNT1.

Hope this will help.