Current consumption difference nRF51422 and nRF51822

RE: Current consumption difference nRF51422 and nRF51822

Bart de Kramer — Wed, 06 Jan 2016 09:34:29 GMT

Hung,

I found the solution based on my last comment. As you can see in my design I have a resistor over the 16MHz crystal (R11) which should be not placed (it can be used for EMC issues). But somehow the PCB manufacturer did place a 0R resistor so the HF clock could never start! Thank you for your support!

RE: Current consumption difference nRF51422 and nRF51822

Bart de Kramer — Wed, 06 Jan 2016 09:09:34 GMT

Hung,

I've tried some other things. I've created an simple program what just initialize the clocks (LF and HF). When I only start the LF clock, I get a current draw of 10uA. Otherwise, if I also try to start the HF clock and then wait untill it is started (: while (! NRF_CLOCK->EVENTS_HFCLKSTARTED) {}) I never pass this while loop and the current consumption stays at 2mA! It looks like the HF clock won't start

RE: Current consumption difference nRF51422 and nRF51822

Bart de Kramer — Tue, 05 Jan 2016 13:10:49 GMT

I've tried SoftDevice S110 V7.1.0 but I still get that high current. As soon as I call SOFTDEVICE_HANDLER_INIT(NRF_CLOCK_LFCLKSRC_RC_250_PPM_TEMP_4000MS_CALIBRATION, NULL); that high current occurs

RE: Current consumption difference nRF51422 and nRF51822

Bart de Kramer — Mon, 04 Jan 2016 15:04:11 GMT

Hi Hung, The debug mode is not the issue. I send the command <nrfjprog --pinreset> after each flash cycle. But, I'm using Softdevice S110 V8.0.0! I'm not sure which SDK I'm using. Do you know where I can check the SDK version (I'm using Keil 5).

RE: Current consumption difference nRF51422 and nRF51822

Hung Bui — Mon, 04 Jan 2016 14:34:06 GMT

Hi Bart,

Thank you, and Happy New Year to you too.

The 2 test cases without softdevice showed the expected result => no issue with the chip or other peripheral on board. The test with softdevice showing 2mA during idle is pretty strange, have you made sure you power reset the chip after flashing ?

The chip will enter debug mode after flashing and can consume more power. ( it should be around 30uA average current consumption with 2-3 uA when in sleep mode - idle)

You can test if the radio cause the exceed power by removing the advertising_start() function, so that there is no RF activity but with the softdevice active.

Note that you should test with SDK v6.x (and S110 v7.0), SDK v7 and above were not tested with Revision 2 chip (only revision 3).

RE: Current consumption difference nRF51422 and nRF51822

Bart de Kramer — Mon, 04 Jan 2016 11:08:05 GMT

Hi Hung, First of all, happy newyear! Tomorrow I started again with the problem above. As you suggested I changed the values for L1 and C5 as showed in the nRF51822 PS (L1 = 10nF, C5 = 2.2pF) and removed the SWDCLK pulldown. But I still have that high current consumption.

The tests you suggested gives the following result: Erase SoftDevice:

4.26 mA current consumption
9.4 uA current consumption

Flash softdevice:

2.07 mA during idle and a peak of 2.65 mA every 500msec (I assume this is because of the advertising). But if I scan for bluetooth devices with my phone I also cannot find the unit.

RE: Current consumption difference nRF51422 and nRF51822

Hung Bui — Tue, 22 Dec 2015 13:37:53 GMT

Hi Bart,

The chip on the PCA100028 is the nRF51422 QFAC (revision3+) and has different matching network components' values. Your chip ( I assume N51822QFAAG3 not 63) is revision 2 chip, you can find the reference design for it in the nRF51822 PS. You don't need the pulldown on SWDCLK for it.

Please also try the tests I mentioned above.

RE: Current consumption difference nRF51422 and nRF51822

Bart de Kramer — Tue, 22 Dec 2015 12:31:46 GMT

I used the values from the schematics from the PCA100028 (Revision 1.0.0, date 24-10-2014). The SWDCLK pulldown should be removed if I have the N51822QFAA631440BK?

RE: Current consumption difference nRF51422 and nRF51822

Hung Bui — Tue, 22 Dec 2015 12:14:08 GMT

Wait a sec, I just looking again at the value of the RF components, could you explain why you use L1 =27nH, and C5 = 3.9 pF ? R12 for SWDCLK is also not needed (if you are not using the very old revision 1 chip from 2 years ago)

RE: Current consumption difference nRF51422 and nRF51822

Hung Bui — Tue, 22 Dec 2015 12:09:21 GMT

Hi Bart,

I have received your files. They looked fine for me. (except some improvement can be done with the layout on RF for optimization but we can discuss that after solving this issue)

I would suggest you to test with following firmwares in the nRF51 SDK and report the current consumption, in each test. Please make sure you removed or modified the firmware to match with your board's GPIO pin (the LED pins)

Erase softdevice

Test with \examples\peripheral\blinky (modify so that it blink only one LED, if you have no LED you can check if the voltage on one pin is changing)
Test with \examples\peripheral\pin_change_int (Add __WFE(); to put CPU to sleep)

Flash softdevice:

Test with \examples\ble_peripheral\ble_app_hrs (remove and LED output). Check if the device advertise, if it doesn't you can try to go to debug mode and walkthrough until the advertising start API call.

RE: Current consumption difference nRF51422 and nRF51822

Bart de Kramer — Tue, 22 Dec 2015 07:01:44 GMT

Hung, I've send the schematic and layout trough the support portal. unfortunately this is my last working day before Christmas. So maybe I can do some tests today, otherwise I'll be back at monday 4th January

RE: Current consumption difference nRF51422 and nRF51822

Hung Bui — Mon, 21 Dec 2015 16:36:38 GMT

Yes, but we need to know why on our board it doesn't draw more current but on yours it does.

I suspecting the RF part causing the issue. You can try to test with an example that advertise something. Also you can try to test with the radio_test example (the simple one, not with softdevice). You can open a case on our Support Portal (My Page) and send your schematic and layout there, then we can do better investigation.

RE: Current consumption difference nRF51422 and nRF51822

Bart de Kramer — Mon, 21 Dec 2015 14:16:54 GMT

Hung,

I made a very simple main loop:

int main(void) { ble_stack_init(); for (;;) { sd_power_system_off(); } }

Current draw is now 8.3uA, so for me it look likes the nRF51822 draws that excessive current.

RE: Current consumption difference nRF51422 and nRF51822

Hung Bui — Mon, 21 Dec 2015 13:54:45 GMT

@Bart: I would suggest you to test the board with some simple code. Try to use some simple peripheral, or simply put the CPU to sleep. Just to check it's not from other part of the board. Then you can try the example with softdevice in the SDK.

RE: Current consumption difference nRF51422 and nRF51822

Bart de Kramer — Mon, 21 Dec 2015 07:12:04 GMT

Hi Hung, I have some extra components, but I add these to the development board too. I removed the 32kHz crystal from my own PCB but also from the development board. I've only tested with an softdevice application (starting with ble_app_hrs_s110_with_dfu_pca10028 but add my own functionality to it). The softdevice init in the function BLE_Stack_Init is now like SOFTDEVICE_HANDLER_INIT(NRF_CLOCK_LFCLKSRC_RC_250_PPM_TEMP_4000MS_CALIBRATION, NULL);

At my own PCB I cannot establish a Bluetooth connection because I need to tune my antenna but the software is running on it. I have a serial port on it at which I can print some chars.

RE: Current consumption difference nRF51422 and nRF51822

Hung Bui — Fri, 18 Dec 2015 16:19:28 GMT

Hi Bart,

I would need to see your schematics, to investigate more. Have you made sure you don't have any extra components that may consume more power ? Do you have the 32 KHz crystal on the board ? Beside consuming more power, would it still works at it should ? What kind of application did you test ? with or without softdevice ?