I'm trying to build my own sensor network, and I'm hoping to eventually have about half to a dozen sensor all over the house along with a weather station or two outside... Hence me thinking of using xbees for the wireless part of the network. If anyone has a better/cheaper RF that they can recommend, I'm willing to consider it.

Anyway the door bell & closing/opening of the door is kind of a special case, because it can happen at any time so I was trying to design a way for me to wake up the xbee when the event happen, otherwise just wake up every minute to five minute and just report that nothing has changed/happened.

So basically what I was thinking of doing was doing the following:

  1. Op-isolator (ac->dc) for attaching to the 16v doorbell chime.
  2. Hall Effects sensor + magnet for attaching to the door to detect opening/closing.
  3. Some form of logic that would trigger the xbee to wake up when the door bell is pressed, or there is a state change on the door (IE open to closed, or closed to open).
  4. Some sort of latch? to hold what the state is till the xbee is woken up enough to read and transmit the state.

Basically the goals of this is to get the xbee and the rest of the sensor to be as low powered and stay in a sleeping state as much as possible hence triggering the xbee to wake up. My current issue is how to design the circuit...

Also from reading it seems that you have to hold the line HIGH to keep the xbee in sleeping state, so I was wondering how low I can get that current consumption to be, and also it will still need a low amount of current in the op-isolator and the halls sensor to detect a state change there.

  • \$\begingroup\$ That sounds like a nice project, but you haven't asked a question here. Please be explicit about the problem you're trying to solve. Also understand that we discourage 'brainstorming' questions here, those are much better suited for chat. \$\endgroup\$ Commented Nov 23, 2011 at 20:39
  • \$\begingroup\$ I just made my own PCB based on the TI-2533 with onboard antenna and made my own communications layer that will be controleld via a central panel.I just ordered PCB's from china and going to assemble them. 1 pcb will cost me about 5usd max to make and can run on a lithuim coin cell for 2 years! Hehe... I know you are looking for something now- but if you look at my page soon i will be selling these modules- because they are freaking awesome! \$\endgroup\$
    – Piotr Kula
    Commented Nov 23, 2011 at 22:47
  • \$\begingroup\$ You can put many devices into deep sleep and use External Interrupts to bring it back to life- send a burst of packets saying "door open" - then go back to deep sleep- The recieving side should always be "ON" and ready. \$\endgroup\$
    – Piotr Kula
    Commented Nov 24, 2011 at 8:13

4 Answers 4


Why not have a reed-switch instead of a hall effect sensor? These take no power and act like a perfectly normal switch.

And how about having the reed-switch / doorbell control the power to the unit? You could arrange it so that when the button is pressed or the door is opened the power is then turned on. The turning on of the power could trigger something else that turns on the power (such as a relay in parallel with the button) so the power latches on.

The device will then boot up and do what it needs to do. After that it can then interrupt the power (turn off the relay?) and it all goes silent again.

You can't get a lower idle power than that now can you?

  • \$\begingroup\$ This could require a filter capacitor and voltage drop detector to discover when you are loosing power again, if expected to announce that fact over the RF link. \$\endgroup\$ Commented Nov 23, 2011 at 21:18

Wicked Device (admittedly my company) makes a very low cost sensor networking product offering called the Wicked Node and Wicked Receiver. One Receiver can receive from multiple Nodes. It's going to be a hell of a lot cheaper than an XBee based solution. All you need is an Arduino (or clone) platform to plug the receiver into.

It's good up to about 300 feet line of sight depending on terrain and works well within a house / through walls. Each Wicked Node can host up to 3 analog sensors and set up to transmit on a regular interval (10 second for test, 1 minute or 10 minutes), or on a falling edge of separate input (like a magnetic switch for example). It's basically designed for something very similar to what you are wanting to do, you should definitely check it out at least.

Also take a look at my blog entry I wrote about using these to alert me to when my mailbox had been opened. It's very similar to what you are wanting to do with respect to your doorbell.

A lot of people are pairing these with Nanodes (low cost Arduino clone with Ethernet) to collect sensor data and post it to the internet.

Hopefully this doesn't come across as too much of a "sales pitch" but I couldn't resist - your application sounds like a really good fit.


Another RF solution to consider is TI's 430BOOST-CC110L - CC110L RF Module BoosterPack (https://estore.ti.com/430BOOST-CC110L-CC110L-RF-Module-BoosterPack-P2734.aspx ) which works with their MSP430 LaunchPad kits.

The MSP Launchpads cost $4.30 each and the RF BoosterPack for $19 includes two(2) radios. Shipping is free.

TI claims that their MSP430s offer excellent low power solutions. As they say, your mileage may vary...

In any case this would be a relatively low cost ($15 per node) solution using readily available boards from a major manufacturer.

  • \$\begingroup\$ Yea- I just designed a PCB around this tech from TI my gaol was cheap and power efficient. It can run for up too 2 years of single coin cell and woken up remotely by sensing probe signals while in low power mode. TI CC-2540 claims to be even better an is SoC# \$\endgroup\$
    – Piotr Kula
    Commented Nov 24, 2011 at 8:11

As long as you do not need your central unit to be powered down or of very low power then between doorbell actions (button press, button release) and door actions once the door is fully open or fully closed, the system need draw no current at all - or so little as to not matter.

Zero current may be problematic if you want to maintain registers and memory state but even those can be copied to flash eerom prior to shutdown if true zero power is wanted. Usually a whiff of curret is acceptable if it conveys benefit.

For example, some processors have a keep alive state that maintains registers and on board memory and some status so that they can be woken to continue a pre-sleep state. This may require say 1 uA (microamp of backup current - more or less depending on processor. You need about 1 mAh of battery capacity per year per uA of continuous current. In most cases this amount of drain is trivial - especially so if you have even very low power "energy harvesting" available such as a sliver of PV panel supplying your battery using even indoor light.

Majenko suggested reed rather than hall switches. They have good (zero off current) and bad points (somewhat fragile, may have minimum wetting current) , but you can get strobed Hall cells with say a 100 mS or even 50mS inter- wakeup period

A nice examples is ROHM's BU520xx series of ultra low current Hall switches with 3.5 uA mean current draw and 50 mS inter-wakeup periods.

enter image description here

  • \$\begingroup\$ The decision may actually depend on if one state (open or closed) overwhelmingly dominates, or if they are comparably frequent. If comparable, the current draw of a pullup resistor shorted to ground via a closed reed switch will likely exceed that of this self-interval-polling sensor. If unequal, choose an NC or NO reed switch so the switch is closed only in the less common state. \$\endgroup\$ Commented Nov 23, 2011 at 21:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.