Go Down

Topic: Lowest consumption issue (Read 8063 times) previous topic - next topic



When using my MKRFOX1200 supplied through 3.3V / Vin , I Can't get a lower consumption than 1.2 mA during LowPower.deepsleep(120000); instruction time (using <ArduinoLowPower.h>).

My test bench uses the simpliest hardware and software (nothing enabled, no LED ... ).

I've double checked this measure using my multimeter (very low quality) in serie (current measure) and in parallel (voltage measure) accross a 100 Ohm resistor in serie with the power supply : 120 mV drop ...

Unfortunately it is not enough efficient to be powered during monthes through a pair of AA alcaline batteries, even if my module sleeps always ... I was expecting less than 100 uA reading the Atmel SAMD data sheet ...

I could see two possible issues :

- The mkrfox1200 hardware design doesn't allow to get better results (ex. due to the clock input ... I don't know ...)
- The Arduino LowPower library doesn't do enough and could be optimized ...

As I've not got enough time to check by myself those two issues, I just hope that someone or a support can tell me what objective can be set  and how to proceed in order to get it. I wouldn't like to spend efforts targetting a non feasible objective, and some other messages on the forum are not enough conclusive on this topic ...

Sleep mode conumption parameter should be written in the MKRFOX1200 documentation, when it's said that the board can be supplied through AA batteries ...

Sorry for the English and the message form, I'm constrained by a Parkinson disease ... however compatible Arduino practice  :)


I did a mistake in my description, sorry : the board is supplied through the battery connector and not through Vin !


I've got a piece of information from technical support / Carla :

"After run some tests on this board, we have reached 300uA with a MKR Fox 1200, so we recommend you to update the low power library to get the improvements."

I had not yet time enough to experiment. I've seen on github a 1.2.0 version (instead of 1.0.0 in my Arduino IDE) but I can't see any major change which could improve my configuration. Anyway, I will do the comparison, for sure.

And now, I know the target/feasibility thanks to Carla.

Again, I would appreciate to get results from others. Thanks. Philippe.


Reading the code of the low power library, I can see that the option BOARD_HAS_COMPANION_CHIP is not setup. Unfortunately this doesn't give us the capability to use the instruction LowPower.companionSleep(); and to ensure that the sigfox module by itself is set in deep sleep mode as well as the SAMD CPU.

In my test bench I did'nt initialised the sigfox module. I was expecting  the sigfox module to be its lowest consumption mode after reset. Furthermore I didn't call a SigFox.end() instruction. Maybe the mistake is there, I don't know ...

I definitively would appreciate to get support on this topic.

Please tell me anyway how you've got  happy with the lifetime of your "AA" battery back !



Sep 13, 2017, 02:22 am Last Edit: Nov 27, 2019, 10:17 am by racpra16
I've updated the Arduino Low Power library to 1.2.0, I've called sigfox.end();   : I get 500 uA (+/- 100 uA) during the sleep mode. Progress, yes, but not as much as expected reading the components data sheet + MKRFOX1200 schematics.

End of my posts on that subject.


Sorry, but I'm coming again a last time in order to provide a fruitful issue .

I've tried "to remove  the pin initialization in wiring.c", as suggested in other forums :


And now my consumtion in standby is about 50uA, difficult to be more precise with my multimeter ...

I've not seen negative side effects, anyway.

I can close the topic as SOLVED, and maybe my trials will help others ...


@philgeek, Wiring.C file in Arduino installation folder on my computer is completely different than this one and there is no pin initialization in it. Do I have to substitute it with the one on Github?


Mar 18, 2018, 10:02 pm Last Edit: Mar 18, 2018, 10:07 pm by philgeek
Here is the updated file with comments & installation path added in the header. I hope this will help to get the low consumption figure on mkrfox1200.

Please take also into account the great benefits of a setup boost 3.3V circuit, which makes batttery life about 10 times longer. My reference is the NCP1402, and its implementation by Sparkfun.

  Copyright (c) 2015 Arduino LLC.  All right reserved.

  This library is free software; you can redistribute it and/or
  modify it under the terms of the GNU Lesser General Public
  License as published by the Free Software Foundation; either
  version 2.1 of the License, or (at your option) any later version.

  This library is distributed in the hope that it will be useful,
  but WITHOUT ANY WARRANTY; without even the implied warranty of
  See the GNU Lesser General Public License for more details.

  You should have received a copy of the GNU Lesser General Public
  License along with this library; if not, write to the Free Software
  Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
// Philgeek on www.philoc.fr
// Edited on 2018/03/17 thanks to question / my post on Arduino forum mkrfox1200/lowest consumption issue
// I just hope it can help but I cannot commit ;-)

// This file has been updated as I looked for the lowest consumption on mkrfox1200
// The modification below is tagged with "phg" embedded in the comment
// Only one line had to be placed as comment ...
// Notes :
// The path on my configuration is .arduino15/packages/arduino/hardware/samd/1..6.16/cores/arduino/wiring.c
// Notes / Prerequisites
// All unused components, and especially sigfox and usb must have been reset by your application before entering deep sleep
// And Arduino Low Power library must be in its 1.2.0 version, at least in my configuration
// Lastly check any pin configuration, be careful with pullups and leds ... but you did that already
// And at he end of the day, you should get around 50 mA, as I've got by myself ... because the data sheet figures & the mkrfox1200 schematics are done that way.
#include "Arduino.h"

#ifdef __cplusplus
extern "C" {

 * System Core Clock is at 1MHz (8MHz/8) at Reset.
 * It is switched to 48MHz in the Reset Handler (startup.c)
uint32_t SystemCoreClock=1000000ul ;

void calibrateADC()
  volatile uint32_t valeur = 0;

  for(int i = 0; i < 5; ++i)
    ADC->SWTRIG.bit.START = 1;
    while( ADC->INTFLAG.bit.RESRDY == 0 || ADC->STATUS.bit.SYNCBUSY == 1 )
      // Waiting for a complete conversion and complete synchronization

    valeur += ADC->RESULT.bit.RESULT;

  valeur = valeur/5;

 * Arduino Zero board initialization
 * Good to know:
 *   - At reset, ResetHandler did the system clock configuration. Core is running at 48MHz.
 *   - Watchdog is disabled by default, unless someone plays with NVM User page
 *   - During reset, all PORT lines are configured as inputs with input buffers, output buffers and pull disabled.
void init( void )
  // Set Systick to 1ms interval, common to all Cortex-M variants
  if ( SysTick_Config( SystemCoreClock / 1000 ) )
    // Capture error
    while ( 1 ) ;
  NVIC_SetPriority (SysTick_IRQn,  (1 << __NVIC_PRIO_BITS) - 2);  /* set Priority for Systick Interrupt (2nd lowest) */

  // Clock PORT for Digital I/O
//  // Clock EIC for I/O interrupts

  // Clock SERCOM for Serial

  // Clock TC/TCC for Pulse and Analog

  // Clock ADC/DAC for Analog

  // Setup all pins (digital and analog) in INPUT mode (default is nothing)
  for (uint32_t ul = 0 ; ul < NUM_DIGITAL_PINS ; ul++ )
 // modification - PHG - 2017-09-13 : line below suppressed in order to get lower consumption in standby /deep sleep mode '<50uA)
//    pinMode( ul, INPUT ) ;

  // Initialize Analog Controller
  // Setting clock

  GCLK->CLKCTRL.reg = GCLK_CLKCTRL_ID( GCM_ADC ) | // Generic Clock ADC
                      GCLK_CLKCTRL_GEN_GCLK0     | // Generic Clock Generator 0 is source
                      GCLK_CLKCTRL_CLKEN ;

  while( ADC->STATUS.bit.SYNCBUSY == 1 );          // Wait for synchronization of registers between the clock domains

  ADC->CTRLB.reg = ADC_CTRLB_PRESCALER_DIV512 |    // Divide Clock by 512.
                   ADC_CTRLB_RESSEL_10BIT;         // 10 bits resolution as default

  ADC->SAMPCTRL.reg = 0x3f;                        // Set max Sampling Time Length

  while( ADC->STATUS.bit.SYNCBUSY == 1 );          // Wait for synchronization of registers between the clock domains

  ADC->INPUTCTRL.reg = ADC_INPUTCTRL_MUXNEG_GND;   // No Negative input (Internal Ground)

  // Averaging (see datasheet table in AVGCTRL register description)
  ADC->AVGCTRL.reg = ADC_AVGCTRL_SAMPLENUM_1 |    // 1 sample only (no oversampling nor averaging)
                     ADC_AVGCTRL_ADJRES(0x0ul);   // Adjusting result by 0

  analogReference( AR_DEFAULT ) ; // Analog Reference is AREF pin (3.3v)

  // Initialize DAC
  // Setting clock
  GCLK->CLKCTRL.reg = GCLK_CLKCTRL_ID( GCM_DAC ) | // Generic Clock ADC
                      GCLK_CLKCTRL_GEN_GCLK0     | // Generic Clock Generator 0 is source
                      GCLK_CLKCTRL_CLKEN ;

  while ( DAC->STATUS.bit.SYNCBUSY == 1 ); // Wait for synchronization of registers between the clock domains
  DAC->CTRLB.reg = DAC_CTRLB_REFSEL_AVCC | // Using the 3.3V reference
                   DAC_CTRLB_EOEN ;        // External Output Enable (Vout)

#ifdef __cplusplus


Hi there philgeek

I have also tried the Mkrfox1200 in Lowpower deepspleep. (The board has its own way of behaving).

I need a board for long life. As a test I have put the board in deep sleep for intervals with a tmp36 temp measure sensor. I have measured the voltage  between vcc / ground and the current used by the sensor. When the board changed to sleeping mode the values did not change.

I think thats not good for a long battery life - did you find a way to make the board "total sleep" .


Hi philgeek and Gormd,

I'm in a similar project and need a low power board.

Just beginning tests, so I will take advantage of your work and keep you also updated about my progresses.

I cant't find wiring.c in my installation. Where should I find it in a MAC OS X installation?

And, what do you mean philgeek with setup boost 3.3V circuit? It's an additional circuit I should put in front of the board? Could you give more information?

By the moment I'm feeding the board using 2xAA cells.

Best regards.



I finally found wiring.c and it looks like the one indicated by philgeek.

Regarding the functions to get the board into sleep mode, I have seen that using LowPower.sleep() function the board awakes without problem, but using LowPower.deepsleep() freezes the board and doesn'n wake up. Have you experienced the same problem?

I've seen one post talking about the same problem, but there is no answers.

I have still not made any mesure about the power consumption.

I'll try tomorrow, maybe.

Keep you updated.



Neither the sleep() nor the deepSleep() functions work. And I'm pretty sure they were working in the first tests. I have never changed any core library. Any suggestion?

It's really weird...

Maybe a kind of hard reset? ...


Hi jmsalomr

A lot of time can be "spild" on this - sometimes its working - sometimes not.

When you use sleep(yourtime to wakeup) you have to add the:
LowPower.attachInterruptWakeup(RTC_ALARM_WAKEUP, alarmEvent0, CHANGE); -in your void setup().. that worked for me.

Have not tried to use the deepsleep - I will try it now....

What will you use your Mkrfox1200 for?

(be aware, that if you have your deve going in sleep mode you have a hard time to connect it to your usb!!)


Hi there

Deepsleep is also ok with attachInterruptWakeup as mentioned.


Hi Gormd,

It was a little bit weird, but it works for me now. I have just to disconnect the board and connect it again. If I left it connected through the USB, even if I push the reset button, it doesn't work - no wake up. But if I disconnect the board and I connect it again, it works. Both with sleep() and deepSleep(). It seems I need just a switch off.

I've been measuring with both sleep() and deepSleep(), without SigFox initialization, and I measure 350 uA in sleep() and deepSleep() modes. No differences between them.

I'm using MKRFOX1200 for an agriculture application. I've been testing Pycom SiPy + PySense with quite good results, but I would like also to test this board.

I'm going to test removing the pin initialization from wiring.c.

Go Up