Verdin AM623 - how to enable rtc1 battery switchover via device tree overlay?

Technical Support
,
1

Hi all,

I am trying to enable the rtc1 (in the Verdin dtsi is defined as rtc_i2c) to be kept powered by battery connected to pin 249 VCC_BACKUP.

I tried to add in an overlay this definition:

&rtc_i2c {
status = “okay”;
battery-switchover = <2>; // Non-rechargeable battery backup, I2C & FOUT always on
};

but from my tests seems the rtc1 is not kept enabled and does not retains the last date time settings. How can I set it properly and test it?

2

Hi, I’m the Toradex AI Assistant.

I might make mistakes. Our support team will step in if needed. If I answered your question, please tick the “Solution” checkbox. For follow-up questions to me, please use @ToradexAI in your post.

Did you like my answer? :+1: :-1:

RTC Configuration on Verdin iMX8M Plus

Based on the Verdin iMX8M Plus V1.1 datasheet, the module features an RTC IC with specific backup power requirements(1).

Hardware Configuration

The Verdin iMX8M Plus module features an RTC IC equipped with an accurate 32.768 kHz quartz crystal(1). For backup power operation:

  • VCC_BACKUP (pin 249): A coin cell must be applied to this supply pin if the RTC needs to be retained without the module’s main voltage(1)
  • Charging capability: While the module’s main voltage is available, the onboard RTC can recharge a supercapacitor or rechargeable battery via the VCC_BACKUP pin with very low charging current (< 500µA at 3V)(1)
  • Current-limiting resistor: When using the recharging capabilities, a current-limiting resistor with a value >47kΩ must be added between the battery and VCC_BACKUP(1)

Important Hardware Limitation

The documentation specifically states: “If a resistor with a lower resistance is used, the module will fail to boot”(1). This is a critical hardware constraint that must be observed.

Testing RTC Backup Functionality

To test if your RTC backup is working properly:

  1. Set the time while the module is powered
  2. Power down the module completely
  3. Wait for a period of time
  4. Power up the module and check if the time was retained

Hardware Verification

Ensure that:

  • A proper coin cell battery (like CR2032) is connected to pin 249 (VCC_BACKUP)
  • If using a rechargeable solution, the current-limiting resistor >47kΩ is in place
  • The battery voltage is sufficient for RTC operation

The device tree overlay approach you’re using should be correct for enabling the RTC functionality, but the hardware setup is crucial for backup power retention.

Best regards, Toradex AI Assistant.

3

The RTC on Verdin AM62 modules should work out of the box and keep time when the main VCC is not present, provided that VCC_BACKUP is at the proper level (3.3 V) and remains uninterrupted. No additional overlay is required.

Could you please specify exactly which image you are using and share your full boot log (as text)?

For instructions hoe to test and set RTC please refer to this article -

4

Dear Alex,

this is the result of the uname -a command:

Linux verdin-am62 6.6.94-0 #1 SMP PREEMPT Mon Jul 7 17:28:02 UTC 2025 aarch64 GNU/Linux

Attached as text file the boot log of my system.

Here briefly the result of dmesg | grep -i “rtc” :

[    0.285807] rtc-ds1307 0-0032: registered as rtc0
[    0.286594] rtc-ds1307 0-0032: setting system clock to 2025-08-28T07:33:20 UTC (1756366400)
[   12.715427] rtc-ti-k3 2b1f0000.rtc: registered as rtc1

I tried also this simple test sequence to verify time preservation over a simple reboot (no power loss) and seems anyway to fail:


RTC_DEV=/dev/rtc0

date -u -s "2025-08-28 07:25:00"

hwclock -w -f "$RTC_DEV"

hwclock -r -f "$RTC_DEV"

RTC_DEV=/dev/rtc1

date -u -s "2025-08-28 07:25:00"

hwclock -w -f "$RTC_DEV"

hwclock -r -f "$RTC_DEV"

sync; reboot

After reboot completed I read back both rtc:

root@verdin-am62:~# hwclock -r -f /dev/rtc0
2025-08-28 07:28:28.818363+00:00
root@verdin-am62:~# hwclock -r -f /dev/rtc1
1970-01-01 00:00:48.997641+00:00

As you can see, rtc1 is always reset to 1970, no persistence at all even if it’s just a soft reboot
log_boot.txt (38.8 KB)

5

/dev/rtc0 appears to be working as expected and corresponds to the on-module low-power Epson RX8130CE real-time clock. rtc1 refers to the AM62 internal RTC, which is not backed up by VCC_Backup, as it consumes significantly more power than the Epson RTC.

1 Like
6

Thanks for the clarification, now I have a reason of why they behave differently. And yes you are right, I confused the 2 devices.

Thanks again for cross check and confirmation!