How to test the power consumption of "System ON IDLE with GRTC (XOSC) and 256 KB RAM"

[666]

HI ALL,

   I used a system off demo and added the following content.

The power consumption tested was around 1uA, which should be normal according to the errata. Then, I wanted to test the idle state. Based on the above, I turned off the following content and tested that the power consumption was around 6uA.

This doesn't seem to match the specification book.

Is there any other way to test the power consumption of "System ON IDLE with GRTC (XOSC) and 256 KB RAM"？

note:v2.8.0,NRF54L15,QFAABB

[Sigurd Hellesvik]

Hi,

Can you try to measure the same for the blinky sample and let us know what current draw that has (when the LED is off in the blink cycle)?

Regards,
Sigurd Hellesvik

[666]

[Hieu]

The issues here are the spikes that you see all over the graph. They are not true spikes but rather false measurements due to limitations of the PPK.

I don't always see this issue, but when I do, changing the voltage in small steps help me find a threshold where such spikes occur. That doesn't help you see the real current draw at 3.3V, but the idea is, you can see that as supply voltage increase, current should decrease. Beyond this, at 3.3V, you would have to zoom in and selectively measure between the spikes.

I am asking internally what those spikes are and what can be done about them. I will get back to you once I get anything.

[666]

Hieu said:

you can see that as supply voltage increase, current should decrease.

From the graph, it appears to be limited to voltages below 3V?

Hieu said:

Beyond this, at 3.3V, you would have to zoom in and selectively measure between the spikes.

Is this the first two images? Is the power consumption 1.8uA at this time

[Hieu]

666 said:

Hieu said:

Beyond this, at 3.3V, you would have to zoom in and selectively measure between the spikes.

Is this the first two images? Is the power consumption 1.8uA at this time

No this also seems wrong.

Could you please measure VDDM and VDD_nRF with a voltmeter and let me know what the voltages are?

[666]

The two in the red box are both 3.43V.

The two in the red box are both 2.93V.

[Hieu]

I am trying to reproduce the issue on my end but haven't been able to yet. I only have earlier prototyping revisions of the DK, so I wonder if that is the difference. I will try to acquire a DK from the same revision as you.

Meanwhile, could you please share what it looks like when you measure in Ampere Meter mode? With all three wires connected, that is.

Do you also have a higher performance power analyzer, a different PPK, or a different nRF54L15 DK to switch up the setup? This is to help isolating where the problem is.

[666]

[Hieu]

Oh, I mean, what does the Power Profiler app look like when you use Ampere Meter mode. And when you use Ampere Meter mode, you need to connect all three pins, VIN, VOUT, and GND.

What about substituting parts for the measurement? Is it possible for you?

[Hieu]

I was tipped that this might be Errata 31 at play. There might be little peaks due to it.

The peaks are not milliamperes high like we see in the Power Profiler, but this compounds with a limitation of the PPK, where the voltage is measured in multiple level ranges, and if the actual level crosses between ranges, a false peak is generated.

Could you please try to install NCS at the main branch, and try again? A fix was made for it.

[666]

Can I use 'west update' directly? It's currently version 2.8.0

[Hieu]

Yes and no. west update will synchronize the SDK to what is defined in nrf/west.yml, so simply runs it won't change anything, unless you replace the entire west.yml with what is on main.