Jetson NX not strartup

hi.
our product with jetson NX,can not start up.We test the singal POWN_IN POWR_EN is ok.but the system cannot work.repeatly the same information.following:

ump :[2021-07-16 09:05:56.116]
0000000000011111111110111000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000[2021-07-16 09:05:56.128]
??[2021-07-16 09:05:56.155]
[0000.024] W> RATCHET: MB1 binary ratchet value 4 is too large than ratchet level 2 from HW fuses.[2021-07-16 09:05:56.163]
[0000.033] I> MB1 (prd-version: 1.5.1.3-t194-41334769-d2a21c57)[2021-07-16 09:05:56.168]
[0000.038] I> Boot-mode: Coldboot[2021-07-16 09:05:56.171]
[0000.041] I> Chip revision : A02P[2021-07-16 09:05:56.173]
[0000.044] I> Bootrom patch version : 15 (correctly patched)[2021-07-16 09:05:56.179]
[0000.049] I> ATE fuse revision : 0x200[2021-07-16 09:05:56.181]
[0000.053] I> Ram repair fuse : 0x0[2021-07-16 09:05:56.184]
[0000.056] I> Ram Code : 0x0[2021-07-16 09:05:56.187]
[0000.058] I> rst_source : 0xb[2021-07-16 09:05:56.189]
[0000.061] I> rst_level : 0x1[2021-07-16 09:05:56.194]
[0000.065] I> Boot-device: QSPI[2021-07-16 09:05:56.196]
[0000.067] I> Qspi flash params source = brbct[2021-07-16 09:05:56.202]
[0000.071] I> Qspi using bpmp-dma[2021-07-16 09:05:56.204]
[0000.074] I> Qspi clock source : pllp[2021-07-16 09:05:56.207]
[0000.078] I> QSPI Flash Size = 32 MB[2021-07-16 09:05:56.209]
[0000.081] I> Qspi initialized successfully[2021-07-16 09:05:56.217]
[0000.085] E> LOADER: Failed to verify SMD.[2021-07-16 09:05:56.219]
[0000.089] I> Primary SMD copy is invalid, try with secondary copy…[2021-07-16 09:05:56.226]
[0000.095] E> LOADER: Failed to verify SMD.[2021-07-16 09:05:56.237]
[0000.099] E> LOADER: Failed to verify SMD & SMD_b.[2021-07-16 09:05:56.237]
[0000.103] E> Error: SMD: Load failed[2021-07-16 09:05:56.238]
[0000.107] E> Load SMD failed[2021-07-16 09:05:56.241]
[0000.109] E> LOADER: Failed to get slot for boot chain from SMD.[2021-07-16 09:05:56.247]
[0000.115] E> LOADER: Failed to get storage info for binary 0 from loader.[2021-07-16 09:05:56.250]
[0000.121] E> LOADER: Failed to get info for binary 0 from loader.[2021-07-16 09:05:56.256]
[0000.127] C> LOADER: Could not read binary 0.[2021-07-16 09:05:56.261]
[0000.131] C> Fail to load mb1-bct bin[2021-07-16 09:05:56.264]
[0000.134] E> Task 24 failed (err: 0x1d1d1318)[2021-07-16 09:05:56.266]
[0000.138] E> Top caller module: LOADER, error module: LOADER, reason: 0x18, aux_info: 0x13[2021-07-16 09:05:56.275]
[0000.146] I> MB1(1.5.1.3-t194-41334769-d2a21c57) BIT boot status dump :[2021-07-16 09:05:56.283]
0000000000011111111110111000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000[2021-07-16 09:05:56.305]
??[2021-07-16 09:05:56.332]
[0000.024] W> RATCHET: MB1 binary ratchet value 4 is too large than ratchet level 2 from HW fuses.[2021-07-16 09:05:56.350]
[0000.033] I> MB1 (prd-version: 1.5.1.3-t194-41334769-d2a21c57)[2021-07-16 09:05:56.350]
[0000.038] I> Boot-mode: Coldboot[2021-07-16 09:05:56.353]
[0000.041] I> Chip revision : A02P[2021-07-16 09:05:56.353]
[0000.044] I> Bootrom patch version : 15 (correctly patched)[2021-07-16 09:05:56.357]

Is your board ever flashed with sdkmanager?

Yes
when it happend,the power plug out and plug in ,it can work.
this case appear occasionally。we use ssd through PCIE*4 startup,not use emmc internal.
thanks.

I don’t think this issue would be resolved by just power in and out…
If this indeed recovered by just cold boot it again, please share me the log in that case again.

What does that mean pcie*4 startup? Boot from NVMe? If so, it is also not related to your issue. The bootloader is still on the QSPI of the module. Not your NVMe. What get broken now is the bootloader.

now it is not appear again.

yes.boot from NVMe ssd outside.you mean this log is in booloader part programme? the kernel is not loaded ?

thanks.

Yes, kernel is still far from booting. Just get error in early bootloader.

If this issue would recover from just power cycling, please try to move this module to devkit and reboot it as much as possible to see if it can be reproduced on devkit too.

Actually, I suspect this only happens to your custom board.

OK。
the total time is about one second.
PowerON,then error happen.so quickly!
thanks.