Overcurrent Error Message on Heavy Load, next try

Robotics & Edge Computing Jetson & Embedded Systems Jetson Orin NX
,
1

Followup from Overcurrent Error Message on Heavy Load.

Setting:

  • Orin Nano Devkit
  • Orin NX 16G Module mounted on Devkit
  • Jetpack 5.1.1 fresh install
  • some PyTorch application

Devkit has been modified: R213 (10k) was polulated in order to enable 19V power supply to the Orin NX module. By default R213 is not populated, and the Devkit always supplies 5V to the module, ignoring the MODULE_ID line.

Problem:
System runs fine in 25W mode.
System displays error message “System throttled due to Over-current.” in MAXN mode when PyTouch application is running.

Power GUI output:

orin_power_gui
orin_power_gui1838×1043 69.2 KB

Why do I get this error message?

  • The current readings are well below the warn current and the crit current settings.
  • Voltage reading is below 20V
  • Current Temp readings are well below SW Throttling and SW Shutdown limits.
  • The sum of all power readings is 27.001W - is this the limitation?

Edit: This is a snapshot of an unmodified devkit:

orin1_5V_power_gui
orin1_5V_power_gui1843×1050 67.1 KB

Here the values are also well below the critical values. Sum of power readings is 27.038W.

5

I also ran into the same issue. For now I have underclocked the CPU a bit since PyTorch is GPU heavy. It’s not a proper fix but gets the job done without the annoying warning.

I followed this link:

https://forums.developer.nvidia.com/t/system-throttled-due-to-over-current-on-orin-nx/247300

7

Hi fchk,

What’s the current power mode in use?

The system-throttled may be triggered through following 3 reasons.

  1. OC1 (under voltage): 4.5V
  2. OC2 (VDD_IN average power): 25W
  3. OC3 (VDD_IN intant power): 30W

Could you help to run the following command and share the result to check which OC event is triggered in your case?

$ sudo su
# grep "" /sys/class/hwmon/hwmon*/oc*
8

I’ve got similar issue. I tried the command above and here is the result:

/sys/class/hwmon/hwmon1/oc1_event_cnt:0
/sys/class/hwmon/hwmon1/oc1_throt_en:1
/sys/class/hwmon/hwmon1/oc2_event_cnt:0
/sys/class/hwmon/hwmon1/oc2_throt_en:1
/sys/class/hwmon/hwmon1/oc3_event_cnt:313
/sys/class/hwmon/hwmon1/oc3_throt_en:1

Does it mean OC3 is triggered? I ran tegrastats and it said VDD_IN was around 15W, far below 30W…

9

Yes, OC is triggered 313 times in your case.
It means “instant” power so that it might be not recorded in tegrastats due to slow sampling rate.
What’s your current power mode in use?

10

It’s MAXN.

11

MAX-N mode is not a suggested power mode to be used.
You could try to create a custom power mode according the power usage of your application.
(e.g. if your application doesn’t have much CPU load, you could limit it to lower frequency).
After that, your OC3 issue might get fixed.

13

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.