Battery Powered Motion Sensor



  • Has anyone successfully assembled a battery motion sensor that lasts for more than a few days, before the batteries are drained?

    If so could you please provide a list of components you used.

    Components I have tried in may combinations:

    Arduino Pro Mini 3.3V http://www.ebay.com/itm/200914924969
    LED and regulator removed

    NRF24L01+ http://www.ebay.com/itm/170819069271
    with CAP

    DC/DC Input 0.8-3V, Output 3.3V. Step Up Power Converter http://www.ebay.com/itm/231083181020
    LED removed

    HC-SR501 PIR Motion Sensor Module http://www.ebay.com/itm/310574919531
    5v and 3.3v power mod

    Mini PIR Motion Sensor Module http://www.ebay.com/itm/221073558145
    5v

    DC/DC Input 1-5V, Output 5V. Step Up Power Converter http://www.ebay.com/itm/400528583414
    LED removed

    Supper mini 3A DC-DC Converter Step Down buck adjustable http://www.ebay.com/itm/201073184133?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1497.l2649

    Power Sources:
    23AE battery powered
    CR2032 battery powered
    2-AA batteries powered

    motion.jpg


  • Hero Member

    Haven't tried building this sensor myself, but it does seem odd that it wouldn't last more than a few days. On a CR2032 I could believe it since it doesn't have that many mAh (perhaps a week tops), but 2xAA, even with the step-up wasting 1mA and the sketch checking in every 2 minutes by default (what is the point of that anyway?), the sensor should still run for 40+ days shouldn't it? http://oregonembedded.com/batterycalc.htm

    Perhaps it never goes to sleep or at least doesn't sleep for long.

    Is the example code suitable for battery powering anyway? I mean, interrupt CHANGE and no debouncer (hardware or software)?



  • I'm building my own Motion/Temperature/Humidity/Light Level sensor and with my calcul, it will run more than 6 mounth I think... The only problem is the PIR run from 4.5V To 20V but with a 3.3V regulator... I would like ton remove the regulator and power it with 3.3V but when I Do it, the PIR Output is verry unstable and do not even work by motion... just blink. Anybody?


  • Admin



  • @bjornhallberg

    I am using the following sleep time

    unsigned long SLEEP_TIME = 1800000; // Sleep time between wakeup and voltage reporting (in milliseconds)

    // Sleep until interrupt comes in on motion sensor. Send update every thirty minutes
    gw.sleep(INTERRUPT,CHANGE,SLEEP_TIME);



  • @hek Thank you very much. I didn't fond it and I will try today... I just removed the 3.3V regulator but it didn't work... I Will try by solder directly on the H pin... Thanks!


  • Hero Member

    @lininger And you're not getting a ton of false positives from the sensor either I take it? Can you check if the sensor is actually sleeping reliably between interrupts/timers (even if it isn't detecting motion / HIGH)?

    Looking at the forum I found two posts which may suggest that getting the motion sensor to work isn't entirely straightforward:
    New interrupt: http://forum.mysensors.org/topic/225/hc-sr501-motion-sensor/18
    Debouncer: http://forum.mysensors.org/topic/115/implementing-multiple-sensors/47
    I'm not experienced enough with interrupts, debouncers and this particular sensor to make any suggestions but it seems likely that the solution is related.


  • Mod

    @bjornhallberg said:

    debouncers

    Debouncing is only required for mechanical switches. Digital signals coming from a microntroller or sensor module of some kind normally don't require debouncing, as is the case with motion sensors.



  • @bjornhallberg said:

    And you're not getting a ton of false positives from the sensor either I take it? Can you check if the sensor is actually sleeping reliably between interrupts/timers (even if it isn't detecting motion / HIGH)?

    3.3v modification of the HC-SR501. No floating, just make sure you run the HC-GND to the Arduino-GND close to the RAW Pin (I read about this somewhere on this board I believe). If I placed the HC-GND at any other point (including the Arduino-GND next to pin 2) the HC-OUT would float all over the place. Also allow a few seconds for the sensor to settle after applying power.



  • @bjornhallberg said:

    Can you check if the sensor is actually sleeping reliably between interrupts/timers

    Yep it is sleeping correctly. I output the trigger event to the console when testing locally and it only triggers on motion detect (1) and motion timeout (0) and when I exit sleep with a (0) I send the battery level back to the gateway when I exit sleep mode.

    I just replaced the DC/DC Input 0.8-3V, Output 3.3V. Step Up Power Converter - it was outputting 5v. The last batch of 5 DC/DC Input 0.8-3V, Output 3.3V I ordered 3 had 5v output and 2 had 3.3v output. I am surprised it did not fry the pro mini as I bypass the regulator. Maybe this will fix my battery drain issue.

    Be sure to check the components you get from china to make what you get is what you ordered.


  • Hero Member

    @lininger Hope it really was that simple.

    I just got around to making my first battery powered motion sensor (with the voltage divider to read battery level) and it does indeed seem to be working as intended for now. Knock on wood. Detection seems to work. Battery level hasn't changed since last night and it is doing its regular reporting per default example just as intended. No change to the interrupt or anything. I have no step-up in the circuit, just 2xAA that powers the NRF24 directly and then the HC-SR501 is powered from the Arduino Pro Mini, not grounded next to the RAW pin but it seems to work anyway. No idea at what voltage the HC-SR501 will start to experience problems but I guess we'll find out as the battery level drops.



  • @bjornhallberg said:

    I have no step-up in the circuit, just 2xAA that powers the NRF24 directly and then the HC-SR501 is powered from the Arduino Pro Mini, not grounded next to the RAW pin but it seems to work anyway. No idea at what voltage the HC-SR501 will start to experience problems but I guess we'll find out as the battery level drops.

    Are the 2xAA that power the NRF24 also powering the Pro Mini or are you using another power source?
    Are you running the HC-SR501 in its 4.5v mode or the 3.3v modification?

    I believe it was the step up converter in my case as it has been 3 days, working fine, and the 2-AA's still report 3.3v when not connected to the sensor.

    Thanks


  • Hero Member

    @lininger Yep, 2xAA power everything. But only the HC-SR501 get its power from the APM. 3.3V soldered to the H-pad. I used an older design that I bought by mistake so it has no convenient H pin.

    IMG_20141004_191011.jpg
    IMG_20141004_191120.jpg

    Don't know how smart this looks, but I have very little experience with electronics and soldering. Would kill for some printed PCBs or the official MySensors battery powered sensor. I have the components to build three proper boost circuits but these TI chips are just so hopelessly tiny. And I still have no idea if the inductors and capacitors I bought are up to the task or AliExpress junk 😕



  • This looks great. Are you still happy with it? Any chance you could post the sketch and wiring diagram/schematic?

    thanks



  • I was working on something like this about 2 years ago and it still works for me for about 6months or so on 2xAA.
    Please, check the following github project page
    https://github.com/endysh/Wireless-PIR-sensor
    With some improvements (I did not have a chance to finish yet) the power consumption in detection mode can be reduced down to 60 - 70 uA, which can give you 1+ year, but it depends on traffic and how often it will send notifications


Log in to reply
 

Suggested Topics

  • 87
  • 6
  • 7
  • 1
  • 3
  • 9

0
Online

11.3k
Users

11.1k
Topics

112.5k
Posts