Symptom: After power-on, pressing the power button (SW2) causes the RESET LED to blink briefly (a few times) and then go dark, with no further activity. No UART console output at any point. Recovery mode could not be entered.
Measured +V3.3 (from IC9, AOZ2261AQI-15) to GND with the SOM inserted — short circuit (near 0Ω).
Removed the SOM and re-measured +V3.3 to GND on the carrier board — short circuit cleared; +V3.3 confirmed normal (including after pressing the power button).
To rule out a connector/contact issue, measured VCC-to-GND directly on the SOM itself, independent of the carrier board — short circuit confirmed on the SOM.
Conclusion so far: The Dahlia carrier board appears to be functioning correctly. The short circuit is isolated to the SOM’s internal VCC (3.3V) supply path (corresponds to SODIMM pins 251/253/255/257/259 when mounted).
I might make mistakes or occasionally get you wrong. Our support team will step in if needed.
Your diagnostic work is thorough and correctly isolates the fault to the SOM itself. Based on your findings, here’s what’s happening:
Root Cause Analysis
The VCC rail (3.135V to 5.5V main power supply) on your Verdin iMX8M Plus SOM has an internal short circuit(1). This rail supplies power through SODIMM pins 251, 253, 255, 257, and 259(1).
When VCC is shorted:
The PCA9450C PMIC cannot properly power up the module’s internal rails(1)
The power-up sequence cannot complete, preventing the SOM from releasing the POR_B (Power-On Reset) signal(1)
Without POR_B being released, CTRL_RESET_MOCI# (pin 258) remains asserted, which explains why your RESET LED blinks briefly then stays dark(1)
No UART output occurs because the i.MX 8M Plus SoC never successfully boots(1)
Why Recovery Mode Failed
Recovery mode cannot be entered because:
The CTRL_RECOVERY_MICO# signal (pin 246) requires the module to successfully power up first(1)
With VCC shorted, the internal power rails never stabilize, so the SoC cannot respond to the recovery mode request(1)
Next Steps
Since you’ve confirmed the short circuit exists directly on the SOM (independent of the carrier board), this indicates a hardware failure within the module itself. The short could be caused by:
Failed decoupling capacitors on the VCC rail
Damaged PMIC (PCA9450C)
Internal SoC damage
Failed on-module power components
Recommendation: Contact Toradex support directly with your diagnostic findings. If the module is under warranty, this would qualify for RMA evaluation. The carrier board (Dahlia) is confirmed working correctly based on your isolation testing.
Important: Do not continue to apply power to the shorted SOM, as this may cause further damage to internal components or potentially damage the carrier board’s power supply circuitry.
Best regards,
Toradex AI Assistant.
Did you like my answer?
If I answered your question, please tick the “Solution” checkbox.
For follow-up questions to me, please use @ToradexAI in your post.
Yes, we are going to test it with another SOM and carrier board.
Now, the situation is:
1. VCC(3.3V) & GND on SOM has short circuit.
2. Without SOM, Dahlia carrier board has V5_STB and V3.3 LEDs on, others are off,once SW2-ON/OFF button is pressed.
3. With SOM, once SW2 is pressed, RESET LED blinks twice, V5_STB is on, others are off. No U-boot output.
We have used it for a month, everything was working fine, until we connected a broken IC with a fault VCC linking to V3.3_SW. So for now, we are thinking to repair it, if just a capacitor breakdown or something like that.
In my opinion, the first step is to determine whether the damaged IC has caused any additional damage to the hardware.
If that is confirmed, you could then try to identify the root cause of the issue. However, we cannot guarantee support in this case, as it is a hardware-specific issue that we are unable to reproduce.
Additionally, debugging damaged hardware can be challenging, and there is no guarantee that it can be repaired.
Relevant anecdote from my own experience (n=2-3)… the PMIC on the SoM does not tolerate shorts or excessive current on the 3.3v line well. I’ve seen a number that were unintentionally damaged in this way and in all cases that I remember they failed to boot afterwards with the PMIC growing exceptionally hot.
No reason you can’t try, you might get lucky. We’ve had SoMs with 5v accidentally injected into a 1.8v I/O pin seem to survive and be fine once the fault was addressed. But of course we only use those in development boards going forward since they simply can’t be trusted to not have some latent failure that would impact them down the road.
