Why the firmware doesn't works on the carrier board of jetson agx

My company makes a carrier board of Jetson AGX,We replicate the circuit on the latest P2822 carrier board exactly.And I download the firmware on https://developer.nvidia.com/embedded/downloads#?search=firmware,I flashed the firmware to the MCU,but the led turns on for less then one second and turns off everytime when I pressed the power button.Is anyone can tell me the reason?

Hi, which MCU and LED do you mean?

the MCU is efm8sb10f8a-qfn20,and the led is the power indicator led.可以用中文交流么?

It should not be firmware issue as that has been validated no problem by other NV and other users. You might need to check the power part design comparing to that of P2822 to find out why the power LED has such behavior first.

ok,thanks for your reply.
And I have another question to ask.In the same carrier board,There is a USB PD Controller circuit. The control chip is CYPD4226-40LQXIT, I noticed that there is also the firmware of this pd controller on the NVIDIA website, however, I can’t flash it with the flashing tool. What I want to ask is, is this because our Power Supervisor chip is not done well? Is this a requirement that the carrier board and core board must be powered on before this PD controller chip can be programmed?

As you can see in DG: Designs that intend to follow the NVIDIA carrier board design and include the Type C PD Controller (CYPD4226 - U513 on NVIDIA carrier board) need to replicate the circuitry on the latest P2822 carrier board exactly. NVIDIA will provide the binary and the customer should get the flashing instructions from Cypress.

The chip should be powered up correctly at least so that it can be flashed. Not sure about if it is related to your design of “Power Supervisor chip”, better to compare to reference first. In addition, there is a checklist sheet attached in DG doc that might be helpful to your comparing.

thanks for your help.It’s very useful.

When I pressed the power on button of the carrier board,There is a force_shutdown signal from jetson agx xavier L52 to the efm8sb10f8g’s P1.0,so the power on indice led turned on and turned off less than a second.Is it normal about the force_shutdown signal from jetson AGX xavier?

Do you have a devkit in hand with which you can compare the behavior of force_shutdown? You can also check the power on sequence in Design Guide for correct signals process.

yes,I do have the devkit.but when I plugged the core board into the carrier board(devkit),The test pin is occluded.As I check the power on sequence in Design Guide,How can I test the sequence?I don’t know which instrument to use.I readed the source code of the firmware,I noticed there is three case which cause the power off behavier.I quote the annotation in the source code:" PWR_BTN_N asserted for more than 10s, FORCE_SHUTDOWN_N is asserted or PWR_GOOD drops, power off",It‘s obvious that we didn’t pressed the pwr_btn_n for more than 10s,and I cut off the force_shutdown circuit so the efm8sb10f8g’s pin 1.0 can not be disturbed anymore.It seems that the only reason is pwr_good drops,but how can I confirm that?

Have you done above steps? You need to confirm there is no obvious design issue first, and then the discussion of issue would be meaningful.

Yes,We confirm there is no obvious design issue.As we replicated the power on circuit exactly. We tested all the pins of efm8sb10f8g on the carrier board.But Limited by the shape of the devkit,we can’t compare the signals between devkit and our carrier board.About the checklist you mentioned above,Is it in the efm8sb10f8g DG,or other DG file?

The checklist is attached in Xavier DG doc.