Softdevice high idle state power consumption

RE: Softdevice high idle state power consumption

Aamir — Mon, 10 Jun 2019 06:55:49 GMT

Thanks Stian for your great great support, your efforts are commendable.

We really appreciate it.

Thanks !

RE: Softdevice high idle state power consumption

Stian Røed Hafskjold — Fri, 07 Jun 2019 08:05:41 GMT

OK great, this seems a lot better. Let me know if you have more questions.

RE: Softdevice high idle state power consumption

Aamir — Fri, 07 Jun 2019 07:49:54 GMT

Hi, Stian

First of all, thank you for your wonderful support.

we carried out the test according to your instructions and also adjusted the scope settings according to your suggestions.

and we got pretty close to the expectations.

This is the reading we got for the 10 ohm - 1 M ohm voltage divider

for Voltage divider, we got around 45 uV corresponding to 4.5 uA across 10 ohm resistor

and We also did the testing on the Preview DK using the 'power profiler app'

and the readings we got is around 23 to 25 uV across a 10 ohms resistor corresponding to 2.3 to 2.5 uA of current consumption.

RE: Softdevice high idle state power consumption

Stian Røed Hafskjold — Tue, 04 Jun 2019 16:09:27 GMT

I tested here now to see if it's possible to measure current with a scope as you are trying. Here are some things you should be aware of.

You must set the mV/div to as low as possible. I did not get good results until I turned it down to 2mV/div
Use average not RMS
at 2mV/div you will get lots of noise, so turn on the bandwidth limit function. On my Agilent it's called 9-bit BW limit (in the acquisition setup)
A 10x attenuation probe did not work because I was limited to 10mV/div. Use a 1x probe. Remember to set the correct probe type in the scope.
Do not use the averaging function.

RE: Softdevice high idle state power consumption

Stian Røed Hafskjold — Tue, 04 Jun 2019 15:46:53 GMT

Well, you need to measure over a 10 Ohm resistor. A 10k resistor will give you a voltage drop in the mV range. The questions is what happens if the voltage on the probe is in the uV range.

So two resistors in series with 3V across. One that limits the current to 3uA (1MOhm), and one 10Ohm measurement resistor. You should be able to measure the 30uV voltage drop on the 10Ohm resistor.

RE: Softdevice high idle state power consumption

Aamir — Tue, 04 Jun 2019 12:36:28 GMT

Hi Stian,

Thanks for your answer,

As you suggested we verified that our instrument is capable of measuring into milli-amperes

we measured 30 mV across 10K ohms resistor, which turns out to be near 3.1 micro amps

On the subject of Eng A. we also have our company's proprietary board with Eng C. chips

so we measured the power profiling app on that as well, unfortunately, I can't show you the board but here is the measurement as taken on the oscilloscope

we got around 45 uA, which is very close to Eng A. chip on the development board.

RE: Softdevice high idle state power consumption

Stian Røed Hafskjold — Mon, 03 Jun 2019 11:27:12 GMT

Hi, here in Europe (except UK) we are using comma as the decimal separator, so it's 1.5uA. I'm using a Power Analyzer to measure current which can supply voltage and measure current at the same time. So the green wire is a 3V output from the power analyzer on channel 2. And the yellow wire is channel 1 in ampere-meter mode.

In general it's hard to measure current in the uA range, you need to have equipment which can handle it. So first you need to figure out if the scope is actually showing you the right consumption. Do a test measurement over an unconnected resistor to see what the base current of the equipment is.

If it turns out that the current flowing into the nRF5 chip is actually 55uA when running the unmodified ble_app_pwr_profiling example, I don't really know what's wrong. Maybe the chip is ESD damaged or similar. I would suggest getting the newest version of the DK with a production ready version of the chip. The eng. A version of the chip is not recommended for new development anymore. The current consumption is different from the production silicon.

RE: Softdevice high idle state power consumption

Aamir — Fri, 31 May 2019 05:51:09 GMT

Hi Stian,

Sorry for the delayed reply, I was out on business trip

I have some of questions,

In your setup, you are using Green wire as 3V external supply, how are you measuring it on oscilloscope ?

In the oscilloscope snapshot you posted, I see can see the marker 1 at 1,505uA and 2,487 uA , both the readings have a comma(,) in them, doesn't that correspond to 1.5 mA and 2.48 mA ? and same with maker 2 and average current.

and we replicated your setup and this is the results we got

our oscilloscope can't measure current directly, so we take the reading across a 10 ohms resistor, which in this case measures around 55 uA. Can you please take a look at this and tell me what's wrong?

Thanks !

RE: Softdevice high idle state power consumption

Stian Røed Hafskjold — Fri, 24 May 2019 13:55:47 GMT

Engineering C is the same as revision 1, the newest chip is revision 2 (eng D). You should be able to verify the consumption on the eng C samples.

RE: Softdevice high idle state power consumption

Aamir — Fri, 24 May 2019 13:06:20 GMT

Yes, I have Engineering A chip the markings on the chip says

NRF52840 QIAAAA 1716AK

aside form this DK I also have our company's proprietary boards, with markings saying

NRF52840 QIAAC0

I think this is engineering C chip, can this chip be used to verify the power consumption ?

RE: Softdevice high idle state power consumption

Stian Røed Hafskjold — Fri, 24 May 2019 12:44:44 GMT

Can you check that the chip is eng A (QIAAAA markings on the chip)? On engineering B (QIAABB) the regulator was never able to fully go into low power mode (ULP) if I recall correctly, so if you have this version it could explain the high idle current you are seeing.

You should get the newest version of the DK, since a lot has changed since the first engineering samples of the chip.

RE: Softdevice high idle state power consumption

Stian Røed Hafskjold — Fri, 24 May 2019 12:39:12 GMT

Hi, not sure what is wrong. The ble_app_pwr_profiling is starting in system OFF mode. I see you have version 0.9.2 of the DK which has the first engineering version of the chip (Eng A), I think? It has lots of bugs and system OFF current should be around 1.5uA.

Here's my DK setup:

Supply power to the P21 "external supply" header (3V).
Set SW10 (VEXT->nRF) to the ON position
Keep USB cable plugged in
Set SW6 to nRF ONLY
Set SW9 in VDD position
Measure current over P22

Green wires are supplying the power and the yellow wire is measuring current

And here's the current measurement running the ble_app_pwr_profiling hex. Green plot is current going into the green wire, and same with yellow.

RE: Softdevice high idle state power consumption

Aamir — Fri, 24 May 2019 09:33:11 GMT

Hi Stian,

is there anything else we can try to measure the power consumption ?

RE: Softdevice high idle state power consumption

Aamir — Thu, 23 May 2019 05:35:27 GMT

Yes, the ground on scope is isolated from DK

and we measured it with a multi-meter, we got around 270 uA

RE: Softdevice high idle state power consumption

Stian Røed Hafskjold — Wed, 22 May 2019 15:27:35 GMT

Are you sure that the ground potential on the scope is completely isolated from the DK? Did you try with a multimeter?

RE: Softdevice high idle state power consumption

Aamir — Wed, 22 May 2019 12:28:01 GMT

the RMS value is fluctuating quite frequently, the moment I captured the above picture it was showing 290uV but it mostly remains aroung 500uV that is why I said 50 uA

We measured it with multimeter and we got around 270 uA

and on the oscilloscope, we switched to Average and the value is fluctuating very much it is even going into negative values

RE: Softdevice high idle state power consumption

Stian Røed Hafskjold — Wed, 22 May 2019 11:09:49 GMT

Yes this is less than version 1.0.0

Can you try to measure with a multimeter instead of the scope?

So is the 50 uA based on the 290uV RMS value? You have to use average value, not RMS.

Are you sure the accuracy of the scope is in the uA range?

RE: Softdevice high idle state power consumption

Aamir — Wed, 22 May 2019 06:22:28 GMT

Hi,

I flashed the ble_app_pwr_profiling hex file and we were getting several hundred microamps.

Then I followed the steps you laid down, except cutting SB58 because my DK version is less than 1.0.0, I think, the sticker on DK says PCA10056 V0.9.2 2017.26 (please confirm that this is less than version 1.0.0)

after following your given steps, now we are reading current consumption of around 50 uA

I have attached the images of oscilloscope and setup here

RE: Softdevice high idle state power consumption

Stian Røed Hafskjold — Tue, 21 May 2019 14:24:52 GMT

Can you try to flash the hex file located in the ble_app_pwr_profiling example folder? It starts in system OFF, so it will consume less than 1 uA. If you still see several microamps, then there's something wrong with the measurement setup.

Then follow this setup:

Supply power to the P21 "external supply" header (3V).
Set SW10 (VEXT->nRF) to the ON position
Cut SB58 (on DK version 1.0.0 or newer)
Keep USB cable plugged in
Set SW6 to nRF ONLY
Set SW9 in VDD position
Measure current over P22

RE: Softdevice high idle state power consumption

Aamir — Tue, 21 May 2019 13:13:28 GMT

We can see the TX, RX waveform of the BLE.

I don't know how to open the file either, the oscilloscope saves it by default in that file

We can see over 200 uA in idle

we use a 10 ohms resistor.

RE: Softdevice high idle state power consumption

Stian Røed Hafskjold — Tue, 21 May 2019 12:36:55 GMT

If you are connecting the scope like this you must be sure that ground on the scope is not connected to the same ground as the development kit. Otherwise you will short the USB power / power supply to ground.

I can't open the scp file. Which program do you use to open it?

So what is the idle current you measure?

Which value do you use on the measurement resistor?

Try to connect an amperemeter/multimeter on the measurement header (remove the resistor) instead of the scope. It's easier to setup and you can use it as a reference for your scope-setup.

RE: Softdevice high idle state power consumption

Aamir — Tue, 21 May 2019 10:58:21 GMT

This is how we have connected the oscilloscope.