ISS2MDC High Power Consumption on Idle

[nikosandreadakis]

Hello,

I am evaluating the STMicroelectronics IIS2MDC magnetometer and measuring its supply current using a Joulescope. According to the datasheet, the device should draw approximately ~1.5 µA in idle mode, yet I consistently measure about 450 µA, which is far from the specified behavior.

Test setup & observations

• Supply: 3.0–3.3 V (also tested across the allowed 1.71–3.6 V range).

• Measurement instrument: Joulescope (µA resolution).

• Verified results across multiple PCBs and multiple IC samples.

• Tested the sensor stand-alone (no MCU connected, only power rails).

• Same behavior observed on the STMicroelectronics evaluation board.

Datasheet behavior (expected)

• After power-on, the device performs a ~20 ms boot procedure to load trimming parameters and then enters Idle mode automatically.

• In Idle mode, most internal blocks are switched off to minimize power consumption, while I²C/SPI remain active.

• Idle mode is selected with MD[1:0] = 10 or 11 in CFG_REG_A.

• A single-measurement cycle (MD=01) performs one conversion and then returns to Idle automatically.

Configuration checks performed

• Verified successful writes to CFG_REG_A (0x60).

• Tested mode settings:

  • MD = 10 and 11 → Idle mode

  • MD = 01 → Single measurement then auto-idle

• Tested both:

  • Low-power mode (LP = 1)

  • High-resolution mode (LP = 0)

• Verified default ODR settings (10 Hz) and tested other ODR values.

• Confirmed device exits measurement and returns to Idle.

Current consumption references from datasheet

• Continuous mode current at 10 Hz:

  • ~100 µA (high-resolution)

  • ~25 µA (low-power)

• After a single measurement, the device should return to Idle, where only leakage and interface logic remain active.

Measured behavior

• After boot and automatic idle entry → ~450 µA

• After single measurement and return to idle → ~450 µA

• No change when:

  • sensor runs stand-alone

  • MD bits changed

  • LP mode enabled

  • different boards or IC batches used

This current level corresponds roughly to continuous measurement at ~50 Hz rather than Idle mode.

---

Has anyone observed similar behavior with the IIS2MDC?
Could there be a condition that prevents the device from truly entering Idle mode (e.g., pin states, offset cancellation, DRDY configuration, or supply sequencing)?

Any insights would be greatly appreciated.

[Federica Bossi]

Hi @nikosandreadakis ,

A few things you can check to narrow down where the discrepancy comes from:

1. Make sure you are measuring only the IIS2MDC current
   On the ST eval board, the supply rail may also feed:

   • LDO/regulators
   • Level shifters / buffers
   • USB bridge, LEDs, pull‑ups, etc.
     If your Joulescope is in series with the board input, 450 uA could easily be the whole board consumption, not just the IIS2MDC.
     Try cutting the trace or lifting VDD to measure only the magnetometer VDD pin.

2. Do a full register dump after your initialization
   Don’t rely only on MD in CFG_REG_A. After your configuration sequence, read back all relevant registers (at least 0x60–0x62) and check:

   • MD[1:0] = 10 or 11 (Idle)
   • LP = 1 (low power)
   • Self‑test and any offset cancellation / compensation features disabled
   • No reboot/reset bits stuck high
     A single wrong bit in CFG_REG_B or CFG_REG_C can keep analog blocks active.

3. Minimal test setup on a single device
   With just the IIS2MDC on a simple PCB:

   • Power it from a lab supply, Joulescope/DMM directly in series with VDD
   • Wait >50 ms after power‑up
   • Write a minimal low‑power configuration:
     • MD = 10 (Idle)
     • LP = 1
     • Lowest ODR, all extra features off
   • Read back the registers to confirm, then measure the current after a few hundred ms.

4. Exclude accidental continuous/single conversions
   Make sure your firmware is not:

   • Periodically writing MD = 01 (single) faster than the ODR, effectively keeping the device always converting
   • Rebooting or resetting the device so often that it rarely sits in Idle.

Hope this helps :)

[nikosandreadakis]

Hello Federica,

Thank you for the detailed suggestions.

Step 1 — Evaluation board isolation
I have already tested this on the evaluation board. All connections to on-board peripherals were disconnected so that only the IIS2MDC supply path remained active during the measurement.

Step 2 — Register verification
After switching modes, I dumped CFG_REG_A and verified that the configuration is correct.
I also tested CFG_REG_B and CFG_REG_C with all bits cleared; however, the current consumption remains high.

Minimal setup on a single device
I repeated the measurements on a custom board where I have full control of the hardware and the IIS2MDC supply path. The measured current is the same.

Additionally, the same behavior is observed across all 60 test boards, which suggests this is not an isolated assembly issue.

I also found the following discussion on the ST Community regarding leakage current on the LIS2MDL package: link

In that case, the increased current consumption was attributed to leakage caused by flux residues during PCB production. The contamination created parasitic leakage paths between pins, leading to significantly higher standby current.

Given that we observe the same elevated consumption across all boards, could a similar flux-related leakage mechanism be possible in our case? Do you have specific cleaning process recommendations (e.g., mandatory washing for LGA packages) or inspection guidelines for this device?

At this point, the device appears to remain in a higher-than-expected consumption state even when configured for Idle + low-power mode.

Looking forward to your feedback.

Best regards,
Nikos

[Mallik18]

Hi Nikos,

We are trying to repeat tests by updating with different footprints (More solder mask open, extended pads to allow flux residue come out from under IC area/ IC pads area). So that there will be chance of cleaning.

Once tested we will share results and observations.

Best Regards,

Mallik

[npreel]

Hello everybody, 

I don't know if this subject is still up, but I am experiencing the same behavior. 
Interestingly, this only happens on newly assembled PCBs. I have older boards with the same design (and likely older IIS2MDC batches), and they do not have this issue.

I tested several solution and one can maybe help you solve this issue. 

I don't use the INT/DRDY pin and I noticed that pulling down this pin to GND, completely erase the consumption problem. Don't ask me why (this pin is supposed to be High Z output). I have opened a support ticket regarding this specific issue here.

Best Regards,
Nicolas

[nikosandreadakis]

Hello @npreel 

I have already pulled the INT/DRDY pin to GND, but the power consumption is still higher than the value specified in the datasheet.

Could you share what equipment and setup you are using for the current measurements?

Also, please make sure the device is not being unintentionally powered during low-power mode through other pins, such as communication lines with external pull-ups. To verify this, try completely isolating the IC from the communication pins and repeat the measurement. 

Below i have upload a X-ray sample of effected chip.

Best Regards,
Nikos.

[npreel]

Hello @nikosandreadakis 

I am using a Power Profiler Kit 2 and a lab generator to do my tests. 
I measure the consumption of my entire board that must be around 30 µA.
With the schematics below, I measure ~450 µA. 
The current consumption is unstable and is sensitive to magnetic field variations, which is why I initially suspected the floating INT/DRDY pin.
As I sais, with a pull down (20K) this consumption disappears.

Unfortunately, my IIS2MDC is integrated on a PCB and I will be unable to disconnect communication pins without being too destructive. But I will end up doing it, and share my updates here.

Indeed, my consumption may come from back powering thought I2C and its 2 pull up. However, this consumption didn't happen on my previous PCB generation. The magneto routing didn't change at all. Its package is the following : LGA-12_L2.0-W2.0-P0.50-TL

Here are my schematics : 

Best Regards,
Nicolas

[nikosandreadakis]

Hello @npreel 

I will repeat the power consumption measurements with the device operating standalone, both with and without the INT/DRDY pin connected through the 20 kΩ resistor, and I will report back with the results captured using the Joulescope.

According to the datasheet, the IC automatically enters low-power mode after boot, without requiring any communication.

Additionally, a small recommendation: if you are using the Z-axis measurements of the sensor, it is good practice to remove copper from all PCB layers directly below the IC. This can help improve the sensor performance and accuracy on the Z axis. Copper is not magnetic but interacts uniquely with magnetic fields.

Best regards,

Nikos