How to know when heater is on, any idea?

Hi, people,

I want to do a small project to know the state of my heater remotely using an arduino as a sensor. I've solved the data sending (using a raspberry) but I want to do the sensor battery operated, so power is an issue.

I'd like your advice to know which could be the best way to detect if the heater is on. I don't need precission (I won't use it for any safety purposes), just if it's on or off. The heater uses natural gas, and has a little flame when is idle and nothing, of course, if it's off.

I've thought about the best ways to know if that little flame is on or off:

  • Using a CO2 sensor. Probably the best option, but this kind of sensor needs several hundreds of mA to work.
  • Using a temperature sensor. I still have to test this (I have to put the sensor quite close to the flame so the little flame makes an effect to it), but, for the moment, it's my best approach.
  • Trying to detect if the gas meter is rolling using a hall effect sensor. I can't do this because the gas meter is outside the window and in a not very reachable position :frowning:

So, I'd like to hear your ideas about which way (thinking about battery life) you think would be the best way to detect this.

Thanks!

Has your heater got a light on it when it is on? If so a simple LDR over the light would do.

Or how about a PIR
http://www.bombayharbor.com/Product/59726/Senba_D203s_Pir_Sensor.html

Grumpy_Mike:
Has your heater got a light on it when it is on? If so a simple LDR over the light would do.

Or how about a PIR

I'm almost sure it doesn't have a light, although I'll check. That would have been a clever idea, because I could have taped the LDR directly onto the light. I can try, anyway. Maybe even the light of the flame can trigger the LDR somehow, although I guess I'd have many false positives.

And the PIR sounds good, too. I'll see if there is a clear line of sight to place it. Thanks for your ideas!

Hack a non-contact themometer?

Grumpy, doesn't that sensor require a change in the IR sensed? It looks like those in the standard alarm sensor.
A flame may move enough if this is the case. I am just asking.

Would a simple IR photodtransistor work?

Weedpharma

if the unit turns on and off, I would think there is power involved in some form ? electrical power that is.

if there is AC and there is a fan or even an electrical valve that opens, then there is current. a non-contact device, a current transformer, can generate voltage. fan running, voltage is present, fan not running, voltage is not present.

interestingly, for a home application, you can use the power available from the CT to be the power source for your project.

Perhaps with these parts you could detect all 3 states of your heater ... OFF (no pilot), IDLE (pilot on), and ON. Probably not the cheapest solution though.

How about a thermocouple in the flame?

weedpharma:
Grumpy, doesn’t that sensor require a change in the IR sensed? It looks like those in the standard alarm sensor.

It is the way that the alarm is constructed that requires movement. The signal is very small from the heat off a body and you need that fluctuation to make the electronics easy. With the larger signal from a flame you can use the absolute reason.

If contact with the flame is okay, maybe rig up a thermoelectric generator?

The heater is pretty old. It has just a gas entry and a button to light a spark (a piezoelectric? not sure about the english word). There is a handle to close the gas pipe (maybe I could check its position, but I don't want to mess with the handle because I have to open and close it everyday).

There is no electric switchs or power, so I can't rely on that. Also the flame is visible, but quite inside the heater, so no easy access.

After some study, I think I'll try the easiest solution: since I'm already using some temperature sensors (DS18B20) around the house, I'll put another one next to the heater. I have yet to see where, because when the heater is on, it gets hot, but when it's idle, I'm not sure if it will get hot enough.

If that doesn't work (specially if, when idle, the flame doesn't affect the sensor), I'll try the PIR.

Thanks everybody for your help. It's nice to see so many different ways to solve a problem!