Garage door status sensors ideas

[rozpruwacz]

can't understand why reed switch makes You unable to make wireless sensor ? these are orthogonal things. Just make an atmega board with nrf24 and reed switch connected to it and voila :) event better, make two sensors, one for checking if the garage door is closed, and second for checking if garage door is fully open. But if you need to report the actual angle of the door than maybe something like ADXL345 ?

[dbemowsk]

@mcqueen Is your garage door opener a chain drive or a lead screw type?

[McQueen]

@dbemowsk it's actually a rubber chain like. I read your post. I have the same issue currently I have a reed switch at the top and the door sometimes slides past the switch.

[McQueen]

@yveaux is it like an accelerometer? Like the ADXL345?

[dbemowsk]

@mcqueen The reason I asked about how your door is driven is that I have heard of some people that have put the magnets on the chain and put the reed switches somewhere near the magnet on the chain. If you had one of the lead screw type that wouldn't have worked.

Have you looked inside your controller? Maybe you have a setup like mine where you could tap into the internal limit switches somehow.

[ileneken3]

I use an accelerometer ( search for MMA8452Q in AliExpress), along with code that checks the orientation:

byte pl = accel.readPL();
   switch (pl)
   {
     case PORTRAIT_U:
       Serial.println("Portrait Up");
       break;
     case PORTRAIT_D:
       Serial.println("Portrait Down");
       break;
     case LANDSCAPE_R:
       Serial.println("Landscape Right");
       break;
     case LANDSCAPE_L:
       Serial.println("Landscape Left");
       break;
     case LOCKOUT:
       Serial.println("Flat");
       break;
   }

   if ( pl == LOCKOUT ) { // flat
     garage_state = GARAGE_CLOSED;
   } else {
     garage_state = GARAGE_OPEN;
   }

[Nca78]

@ileneken3 said in Garage door status sensors ideas:

I use an accelerometer ( search for MMA8452Q in AliExpress), along with code that checks the orientation:

I second this solution it's the best to have 0-90-180-270 orientation with no pain and low power consumption. And with some (complex) calculations you can even calculate a more precise angle, depending on the resolution of the accelerometer.
If you already have an accelereometer breakout at home check if it has this function, it's included in a bunch of chips like LIS3DH.

And be careful on AliExpress some sellers sell MMA7361 with current consumption of 400uA with "MMA8452Q" reference, it's not suitable for battery operation (real MMA8452Q is 6uA in low power mode)

[Yveaux]

@mcqueen said in Garage door status sensors ideas:

@yveaux is it like an accelerometer? Like the ADXL345?

No, more like an HMC5883L compass senosor. Gives you the xyz-angles w.r.t. the earth magnetic field.
MMA8452Q also seems like a good choice though.

[Nca78]

@yveaux said in Garage door status sensors ideas:

@mcqueen said in Garage door status sensors ideas:

@yveaux is it like an accelerometer? Like the ADXL345?

No, more like an HMC5883L compass senosor. Gives you the xyz-angles w.r.t. the earth magnetic field.
MMA8452Q also seems like a good choice though.

Compass are better at this (and can measure while moving) but they are much more power hungry. I guess the best would be to mix the 2 sensors: keep only accelerometer active while there's no movement, measure with compass while door is moving. That way you get the real angle of the door and not an approximation that can tell you it's opened or it's closed if it's stucked in the middle or not fully closed because it's blocked by an obstacle.

[McQueen]

Thanks for the suggestions, everyone! This is very helpful.

[dbemowsk]

When you start talking about using compass sensors and the like, you are talking about sensors that would need to be attached to the door itself. Being that the door is moving, it is not easy to run power to it which means that it would need to be a battery operated node. The other part of it is that it would be hard to combine that with the garage door control for opening and closing the door. I can't imagine the need to know the exact position of the door other than the fully up and/or the fully down positions. If you needed something in between that you can include some logic to estimate position based on timing. The door travels at a certain speed. Based on that you can calculate an average time that it should take for the door to open or close. Add a couple logic checks for what position the door was in when the opener was started, you can calculate whether the door is opening or closing. If the door is closing and you trip the safety sensors causing the door to go back up, you add another check to see if the door returns back to the upper position instead of stopping at the lower position. Based on all of that data you should be able to get a rough idea of the position of the door at any given time.

[gohan]

I still think the reed switch on the side of the door is the best solution, as you can have a battery powered node that can wake up only when sensor is triggered

[zboblamont]

@gohan Tend to second this on the KISS principle, but I recall a comment that the door would sometimes overshoot the optimal contact point (possibly door flexing?) on the 'closed' position...

Perhaps it might be worth re-examining WHERE the switching contact is located to minimise the possibility of 'passing', or change from reed to microswitch with a finger actuator to give a range of detection of say 10mm.
A c/o contact micro would also give confirmation the door was mid transit, the node could have a rundown timer to alarm out if the closed or open signal not received as expected?

[gohan]

Of course you need to find the right spot for the magnet to trigger the reed switch

[iahim67]

Hi, maybe you can use the reed switch in a different way, I mean do not use it like contact closed or contact opened as this would cause you overshoot issues.
When the magnet attached to the door slides in front of the reed switch, then you'll have the reed switch closed for a short time at least, most likely, you can test that. You can use the reed just like you would use a motion detector.
I mean connect the reed to pin 3 - interrupt - and put this at the end of your code, so when an interrupt occurs you will know the door has moved (and you can keep track of movement of course so you would know if the door is open or close) ... just an idea ... :

sleep(digitalPinToInterrupt(DIGITAL_INPUT_SENSOR), CHANGE, SLEEP_TIME);

[rozpruwacz]

another option is to place couple of reed switches in parallel in several places. Only one is needed to close the circuit and trigger the interrupt.

[dbemowsk]

@iahim67 That is a creative option.

[Tim Nagy183]

Couldn't you just use a proximity sensor mounted to the ceiling. That way you can see if the door is up, or if something is parked there, or if the door is down with nothing parked there - all based on the measured distance.

[gohan]

Remember that if you have a battery powered node, try to find normally closed reed switches as they will not drain battery as the magnet will keep the reed switch open.

[iahim67]

@dbemowsk Thank you :smile:
To be a bit more creative I think, @McQueen you can use the interrupt as RISING at the end of your code like this:

sleep(digitalPinToInterrupt(DIGITAL_INPUT_SENSOR), RISING, SLEEP_TIME);

The goal is to set a flag in your main loop every time you get an interrupt and then put this flag in the EEPROM, use the Relay Actuator sketch example again to see how to do that:

        // Store state in eeprom
        saveState(message.sensor, message.getBool());

And read the flag from the EEPROM like this:

        // Set relay to last known state (using eeprom storage)
        digitalWrite(pin, loadState(sensor)?RELAY_ON:RELAY_OFF);

This way you'll know the status of your door. This only works however is during the magnet slide you get just one interrupt trigger, that you can check for yourself, use Arduino itself to print out how many times you get interrupted while closing the door or opening the door. Hope it helps ...