5V power measurement PCA10040

[Hugo]

I need to measure full power consumption of nrf52, on PCA10040 from... a 5V source (because including some other 5V devices).
I measure it from 5V pin and not usb plug.
I cut SB29, to remove atmel processor leak. But I stil get around 4.5mA leakage (all led off), I don't know where from, and it seems huge for a regulator. (around 5mA +peaks while advertising)
Shoud I cut some other lines to only keep the regulator and nrf52 powered?

[Simonr]

Hi Hugo

Sorry if I haven't been clear enough in my previous replies. In order to measure the current draw on the nRF52832 DK when powered from the 5V, please do the following steps:

• Cut SB9 in order to enable the nRF current measurement pin header (P22). Keep in mind that P22 will need a jumper in order to reprogram the device when this SB is cut.
• (Optional): Cut SB10 in order to bypass the reverse voltage protection diode on the USB power. See section 6.7.2 in the DK user guide for how to measure current using an oscilloscope.
• (Optional): Cut SB29 in order to disable power for interface MCU.
• Plug your current measurement tool onto the P22 header.
• Power your device with a 5V source onto the 5V pin on P13.

This should be the required setup to measure the current consumption of the nRF52832 DK. This should not give you a current draw as high as 4.5mA, and if the graph is still as noisy as the one posted below, I think there's an issue in the setup either between the voltage source to the board or from the board to the current measurement tool.

Best regards,

Simon

[Hugo]

ok, I did it, from a siglent SDM3055, on my application, during advertising I get 3.45mA usage on P22, I get 5mA on the SMU. So +45% energy lost, I don't know where. 3.45 still seems high for advertising (no led on), but still better. What do you think?
nb: my application use less power than your blink example, and as I vut more and more jumpers, I'd like to have a useful application left on it.
If I run with "P22" cut (no nrf52), I get around 1.1mA usage. Maybe SB10, but it still seems huge.
SB9 and SB29 are cut, measurement during led off.

[Simonr]

Hi Hugo

What do you mean about "P22 cut (no nRF52)"? If your application just advertises and not does much else I agree that 1.1mA still is too high.

What are your advertising parameters, etc.?

Best regards,

Simon

[Hugo]

I meant SB9.
1.1mA is when SB9 is cut and P22 unplugged...so no current through the nrf.
when advertising, just one service every 8 seconds,

    init.advdata.name_type          = BLE_ADVDATA_FULL_NAME;
    init.advdata.include_appearance = false;
    init.advdata.flags              = BLE_GAP_ADV_FLAGS_LE_ONLY_LIMITED_DISC_MODE;

    init.srdata.uuids_complete.uuid_cnt = sizeof(m_adv_uuids) / sizeof(m_adv_uuids[0]);
    init.srdata.uuids_complete.p_uuids  = m_adv_uuids;

    init.config.ble_adv_fast_enabled  = true;
    init.config.ble_adv_fast_interval = 64;//(in units of 0.625 ms, so 40ms)
    init.config.ble_adv_fast_timeout  = 6000;//60 seconds
    init.evt_handler = on_adv_evt;

[Simonr]

Hugo said:

1.1mA is when SB9 is cut and P22 unplugged...so no current through the nrf.

What are you measuring the current consumption over where the consumption is 1.1mA if it's not the nRF? What is drawing this 1.1mA? And if P22 is unplugged, where are you measuring the current from?

Best regards,

Simon

[Hugo]

I measure (from the beginning) the current from the 5V power supply. That's what a SMU does :)
And the PCA10040 seeems to leak current, from Atmel processor (so sb29 cut), and from something else...