Confusion with online power profiler.

RE: Confusion with online power profiler.

GLNeedham — Fri, 03 Jul 2020 10:32:23 GMT

Simon,

Thanks for your help in understanding this. It is quite a challenge to eliminate the last few microamps but we now have the knowledge to sort this out on this project and future ones

Thank you

Glenn

RE: Confusion with online power profiler.

Simonr — Fri, 03 Jul 2020 10:27:18 GMT

Hi Glenn

This makes sense. I think you got it correctly now! Yes, the HF XTAL will only start when it is called for by the SoftDevice to start the radio (or other operations requiring the XTAL oscillator).

Best regards,

Simon

RE: Confusion with online power profiler.

GLNeedham — Fri, 03 Jul 2020 08:20:58 GMT

Simon,

I can confirm that I am NOT measuring between the wake-up/HFCLK start until the radio starts up but am in fact measuring between advertising so the 100uA does make sense in that context.

As we have SPI initialised but no CSN activity I suspect that block is consuming more or less nothing, the RTC also draws very little. The PWM block is initialised but not generating an output therefore should be consuming more or less nothing. The I2C is initialised and will likely consume typically 50uA

Also, because we have peripherals initialised then the internal HF OSC is likely running which consumes typ. 60uA

I assume the HF XTAL OSC only starts when the radio is needed (in our application)?

So, in theory we have ~50uA for the I2C plus ~60uA for the internal OSC giving a total of ~110uA which fits more or less perfectly with measurements

Your other comment about me misunderstanding the OPP was correct, I had assumed 'standby' meant between advertising, I had not fully taken notice of the ~1000us duration!!

Assuming my comments above are correct then I think we fully understand the situation now so many thanks for your help

Regards

Glenn

RE: Confusion with online power profiler.

Simonr — Fri, 03 Jul 2020 07:55:19 GMT

Hi Glenn

I just ran the ble_app_template project on my side (on an nRF52840DK, didn't have an nRF52832 at hand), and measured the current consumption for some scenarios using the PPK and the Power Profiler application in nRFConnect for Desktop.

Average current while advertising (~140µA):

Average current in IDLE mode between advertising events (~24µA):

I think you're misunderstanding the Online Power Profiler, as the standby you're pointing to is what is between the wake-up/HFCLK start until the radio starts up, which is why it is stated as 0.2mA in the OPP. Can you confirm whether this is the timeslot you're measuring the current consumption in as well (see picture below), or are you measuring between the advertising events themselves?

If you're seeing 100µA in the idle between ramp-up and radio, there must be something wrong with the way you're measuring the current consumption, as the HFCLK is definitely running in this timeslot. If you're seeing 100µA between advertising events, that likely means that one of the peripherals you're using is not uninitialized properly.

Best regards,

Simon

RE: Confusion with online power profiler.

GLNeedham — Thu, 02 Jul 2020 14:09:19 GMT

Simon,

Our FW is based on your template:

C:\Nordic SDK\nRF5_SDK_15.3.0_59ac345\examples\ble_peripheral\ble_app_template

and we have not modified the power management code at all.

Using the Keysight power analyser I can examine the current consumption at any moment in time and the 100uA current is the instantaneous current between advertising and between our 100ms wake up so it is at the time when the CPU should be asleep.

What I can say is that we have 'inited' SPI, RTC, I2C and PWM but while asleep they are not active therefore I suspect that the current being consumed is the IDLE current for each of these peripherals. We do not call the uninit functions because we need to maintain the pin states.

I don't think the HF OSC can be running as the current is not high enough for that (datasheet says 250uA typ.)

Perhaps just the SPI and I2C block current is accounting for what we see?

Regards

Glenn

RE: Confusion with online power profiler.

Simonr — Thu, 02 Jul 2020 13:13:31 GMT

Hi Glenn

This standby current you're seeing in the online power profiler is the standby current before the radio turns on, where the HFCLK is running (consuming ~200µA). The sleep current should not be as much as 100µA either like what you're seeing on your device. I assume you have a peripheral still running that is drawing this extra current, as an nRF52832 in sleep mode running the RTC in sleep mode shouldn't consume more than ~2.6µA. Please see this table for more information on the sleep mode currents (yes this is from the nRF52833 PS, as we haven't updated the nRF52832 PS to include these current consumption scenarios, but they are very similar in terms of current consumption).

Can you show me a snippet of how your application goes to sleep and initializes the RTC? If you're seeing the total current consumption being ~100µA that might be because of what the application does every 100ms when it "wakes up to do some work". Are you able to measure the current consumption in specific parts of the application running, I.E. just when operating and just when it's asleep?

Please note that we're entering the summer vacation period here in Norway, so delays must be expected as we are low on staff for the month of July. Sorry about the inconvenience!

Best regards,

Simon