Help with understanding onboard LDO power consumption

Products & Technology
1

I’m new to the XIAO nrf52840 but not low power design. I am using a XIAO nrf52850 plain (non-sense) board as a BLE beacon. I’m powering it using a 5V supply to the board (not via USB). I am measuring current from the supply using a uCurrent Gold.

With the board powered from the 5V supply, its uses a minimum of 3ma of current. This is odd since I would have assumed the LDO on the board would be super efficient. If I put the SOC into power-down mode deep sleep, it still draws 1ma from the 5V supply.

If I instead bypass the on-board LDO and supply 3.3V to the board, it draws only 4 uA when idle and about 25 uA from the 3.3V supply when it sends out the beacon. This is what I would expect from the board.

The schematic shows U6 as the 3.3V LDO without any part information. I find it hard to believe that this part would consume so much power.

For now, I’m powering it with a 3.3V source.

2

According to the circuit diagram, the BQ25100 charge controller consumes current when 5V is supplied, even if no battery is installed.

3

I took a look at the datasheet for the BQ25100 , and indeed it does draw some power but it from what I can see, it shouldn’t be more than 1 ma.

qc
qc718×141 17.7 KB

4

Hi there,

And Welcome here…

Can you post a pic , how you have it connected. WHich BSP are you compiling with ? and IDE? We have used the PPK2 for most if not all the measurements. You right in there on it with 3.3v :+1:
are you using a coin cell battery ? Which one?

HTH
GL :santa_claus: PJ :v:

5

Here is a pic of the setup.

The board has only two connections.

5V is connected and GND with the uCurrent Gold set connected and set to 1mv/ma. The board is running code that does nothing. It simply sleeps. The board is drawing 3ma when powered on the 5V pin.

In the second test, the board is connected to 3.3V pin and GND. The uCurrent Gold is set to read 1mv/uA. The board is drawing between 14/15 uA when powered on the 3.3V pin.

Code is compiled using gcc, arduino IDE level 0 (release) .

code
code564×231 9.85 KB

setup
setup1536×2048 438 KB

Setup2
Setup21536×2048 279 KB

3ma5V
3ma5V1536×2048 255 KB

setup3V
setup3V1536×2048 429 KB

14uA3V
14uA3V1536×2048 263 KB

1 Like
6

Hi there,

Excellent Documentation. :+1:

How about that compiler output, Verbose , Just the first 3 lines (BSP used)and the last 12-15 before it uploads. :v: (LIB-check) :grin:

HTH
GL :slight_smile: PJ :v:

7

First few lines

first_3lines
first_3lines1238×137 7.17 KB

last few lines

last15
last151242×229 17.1 KB

8

Please also see the link below. I actually looked at the current waveform, and a rectangular wave current of several mA is being repeatedly consumed.

9

Yes, I am aware of the charge controller which I was under the impression was a TI part.

TI bq2510x 250-mA Single Cell Li-Ion Battery Chargers, 1-mA Termination, 75-nA Battery Leakage

However, from what I can see, that part has a max quiescent current usage of 1ma.

qc
qc718×141 17.7 KB

1 ma is still not great, but in my case, I’m seeing 3ma on the 5V supply and that translates into more like a 5ma load on the battery which is driving a small boost converter.

I was hoping to be able to power with just 5V as I have other components in the design that require the 5V rail.

10

Hi there,

So Have you tried to Roll back the BSP to the 1.1.1 variety :grin: and test again?
more to rule out the softdevice and also I believe it’s not the TI part anymore but I may be wrong? this was an earlier conversation years ago, but SMH :v:

HTH
GL :santa_claus: PJ :v:

11

Took a closer look at the TI datasheet and indeed, the battery detection algorithm does use 2ma when the chip tries to detect the presence of a battery.

This coupled with the other 1ma quiescent current usage adds up to the 3ma I’m seeing.

2ma_no_battery
2ma_no_battery1150×557 148 KB

1 Like