[How-To] Measure Gas Water Heater Energy Usage

I noticed on the pictures that it seems that there is a insulation wrap on the water heater. I just wanted to let you know that it is a code violation to use a insulation wrap like that on a gas water heater. Something about it possibly catching on fire ;)

Just wanted to mention it as another poster was also talking about adding one to his water heater. Use it for electric heaters, not gas.

Other than that nick picky detail, great write up. I am just getting started in the whole HA scene and I have a gas water heater, so this definitely helps me.

Keep up the good work.
I tried using the modification as shown, but was having no luck due to the Hawkeye I was using not being able to reach my W800 with RF from the location of my hot water heater. I've had better luck with DS10A sensors, and sure enough, it had sufficient range. So, that might be an option if others are having that problem.

Then, I couldn't get the sensor to recognize the flame, even while varying the resistance from 47K - 267K. So, I stopped by Radio Shack and bought a pack of their photocells (a variety of 5) for less than $3. After some experimentation with the largest photocell in the pack (my theory being it had more surface area) I was able to get a reliable response using a 4.7K ohm resistor.

Just thought others might want to know this in case you were having problems with the original, excellent idea here.
Have just discovered this forum - extremely informative & stimulates some interesting HA interfacing concepts & ideas.

I had separately been contemplating how to monitor gas hot water heater "status", I have planned - but not as yet implemented this idea.
I wonder if any others have attempted this or similar idea??

Method 1 Thermocouple - which generates a very small voltage based upon "flame" temperature (an appropriate thermocouple can even be inserted directly into the flame)
This technique is typically used for monitoring the pilot light in gas appliances & when hooked to an appropriate circuit (more below) will generate a voltage that is proportional to flame temp.

Accordingly this has potential uses for monitoring "presence of flame" - some that come to mind :
- monitor pilot light on gas appliances (eg hot water heater, heating furnace)
- monitor status of hot water "heating" (ie when the main gas jets & thus flame is operating in heat mode)
- monitor gas fireplace status ("on/off")

Thermocouple types : there are a number of different thermocouples - the one required (most common) for above applications is "Type K"
(this type of thermocouple has Sensitivity approximately 41 µV/°C.) with temp range up to 1200 °C
A suitable Type K from Digikey could be : BK Precision TP-29-ND (approx $11) - note this is limited to +392F (hence not suitable for insertion into flame)
Thermocouple would need to be positioned in proximity to flame to ensure max temp is well within the max rating of thermocouple.

Thermocouple Amplifier Circuit : to amplify the very small voltage produced by a thermocouple
Analog Devices AD597AR digikey $6

Thus the thermocouple temp will produce a voltage output proportional to temperature, this can be used to
a) close a relay & thus become an input for DS10, 1 wire switch, input to digital IO board input
:) input to a analog to digital IO board, HVAC 1 wire board

Method 2 Encapsulated Temp Switch
An "encapsulated thermal switch" is used to measure surface temperature - are available in +5C steps & are (claimed) accurate to +/-5C
Digikey has these units (approx $9 each) model 317-10XX-ND in both normally open & normally closed contacts. (select XX according to temp)
They are available in nominal temp steps of +5C from 70C to 150C
Note that max permanent temp of housing is +220C - hence the unit needs to be mounted to a surface with a good differential temp between ambient & the internal heating areas, BUT is always within rated max temp.
Note that the "conducting surface" that brings heat from the flame area needs to not be "too large" - as otherwise it will have a high thermal hysterisis that will slow the response (ie the switch off when flame is extinguished will have a time lag)

The thermal switch method is much simpler, requires no power (its a simple passive switch) or knowledge of electronics - but does require some "mechanical" dexterity & also experimentation to get temperatures in the correct range to trigger the switch. It also has some hysterisis & thus will have a time lag both for "flame on" & "flame off"

The thermal switch can be used directly as an input to DS10, IO board, 1 wire switch, powerflash module, etc.
Hi Greg;

Implementing a thermocouple might be a touch harder as you have to have a reference temperature for it, plus it doesn't output a "simple" linear voltage proportional to the temperature. Rather is uses a "look up table" or polynomial calculation. This is why you would want to use a datalogger that is made for thermocouple data collection as they have this algorithm built in (based on the "type" you are using).

But, if all you are doing is measuring a "presence" this might work. I'll have to look into this some more.

BTW, welcome to CocoonTech. I personally like your creative thinking!


BSR - thanks for the welcome.

The AD597 Thermocouple Conditioner and Setpoint Controller module that I identified takes care of temp reference point & has internal circuitry to produce a "linear" output of 10mV/C - the advantage of this module is that it takes care of all the control & other things - & needs minimal external circuitry.
Hence with use of this module - which has wide supply voltage - it should produce proportional voltage (unless Im reading it wrong?)
I wasnt planning to worry about the exact flame temp - given that its a binary control (gas/flame on or off) - I was planning only to use this as a direct indication of presence of flame - thus negating a need to lookup voltages etc
This would produce a wide swing in voltage between "flame & no flame" - that should enable control of relay or similar.

I will try both techniques I described - however it will be some time before I can get to the heater - but I will in due course post my results.