I want to know when my gas hot water tank turns on and off.
The 3" exhaust pipe bends about 75 degrees immediately above the tank. I attached a thermistor at the top of the bend and used an Arduino analog port to read the sensor. The data shows a significant rise and fall when the tank cycles, so the cycles are quite detectable. But I want more precision. I want to detect the ON and OFF times within about 5 seconds or better, and I cannot get this by sensing the exhaust heat. There is too much latency and variation of rise and fall times from cycle to cycle.
I then tried a photo sensor (light dependent resistor). I taped the sensor to the glass window into the combustion chamber at the bottom of the tank. The pilot light is the only significant light when the heat cycle is off. I replaced the insulation and cover over the window so that there is no light leakage from the outside. The data shows an immediate drop in the resistance when the gas comes on, so the transition from off to ON works great. But the LDR is slow to respond when the cycle ends. It can take up to a minute for the LDR to return to the typical OFF resistance. There are also rather large variations in resistance during the OFF time.
I have looked online for other suggestions but have not found any. After 2 unsatisfactory attempts I decided maybe I should ask around. Does anyone have suggestions for a better sensor (e.g. infrared), or some other method of gas hot water tank ON / OFF detection?
An LDR should change resistance with light intensity, it seems strange that it would react quickly to the burner turning on but take over a minute to react to the burner going off.
Could it be that the LDR is reacting to infrared so that cooling has to take place before it detects an 'off'.
Perhaps you should check what wavelengths the LDR is sensitive to, and a photodiode may be another option.
Maybe having an 'on' resistance threshold and an a different 'off' resistance threshold would solve the problem.
It would probably be hard to isolate the LDR from heat.
However if something opaque to visible light was put in front of the LDR and the resistance measured as the boiler cycles that would prove if the LDR was reacting to heat either as infrared or simply because the LDR itself was getting hot.
Knowing any resistance changes without visible light might help 'on' and 'off' resistances to be chosen.
And Opto-copupler can work fine. What signal, voltage, do You intend to connect to the sending diod in thew coupler? If AC, use a diode managing the top-voltage connected in serial, in the correct direction!
As Railroader says, ldr's work fine at detecting the flame.
Tobacco burner dryers from 30 years ago used them successfully.
Perhaps it's the setup used ( whatever that is).
There is a basic ldr module available with all the necessary circuitry available for reliable operation and adjustable sense level with a comparator on Ebay etc. for a dollar or so.
Thanks for all of the replies. From the comments, my guess is that the LDR is being influenced by the heat of the burn chamber. Although the ldr is outside of the chamber, and separated from it by a glass window that is slightly below the gas burn level, I suspect that there may be a small heat rise during a heat cycle. Thanks for this suggestion. I will do some experimentation to see if this is true, and if so, if I can attempt to further isolate the ldr from the heat, as mentioned above.
The following is a more detailed description of the hot water tank.
This is a conventional 40 gallon gas hot water heater. It consists of an insulated cylindrical "donut" tank. I call it "donut" because the tank is made with a 4" exhaust flu that extends all the way through the middle of the tank - from the burner, which is under the tank, to the top, where it is connected to an exhaust vent line during installation. The tank is about 2 feet in diameter and 5 feet high.
The tank is made with a control box mounted outside of the insulated cylindrical enclosure and near the bottom of the tank. The control box has a water temperature probe that goes through the enclosure and into the tank. The control box includes a temperature controlled gas valve. A gas line (part of the tank) goes from the control box to the burner. The burner is mounted directly under the tank. When the tank is installed a gas supply line is connected to the control box.
There is a service cover at the bottom of the tank. Removing the service cover and some insulation exposes a small window into the burn chamber. To start the tank after installation one presses a button on the control box to allow gas to flow to the pilot. While holding the button down, one presses a piezo igniter button every second or so until the pilot ignites. After the pilot is ignited, the tank runs on its own.
There are no external electrical connections to the tank, and the control box is completely sealed. The tank operates completely on the small electricity that is generated by a thermocouple (mounted next to the pilot), and the mechanical action of the water temperature probe.
hammy:
Put a temperature sensor around the hot water outlet , that will tell you if the boiler is running .
Such systems are slow by the very nature , but this can be suprisingly quick
I would agree with that. I once taped digital thermometers to the inlet and outlet of a gas boiler and from memory the outlet rose quite soon after the burner turned on, I am not sure if it was 3s or less though. Also my present boiler has a digital temperature display and it starts to rise soon after the burner turns on, in this case I am pretty sure it is more than 3s though. I can hear the boiler igniting, then I see a little flame signal to show that the gas has ignited then a little later the temperature starts to rise.
... From the comments, my guess is that the LDR is being influenced by the heat of the burn chamber. Although the ldr is outside of the chamber, and separated from it by a glass window that is slightly below the gas burn level, I suspect that there may be a small heat rise during a heat cycle.....
You should not give up on the LDR, as you presently have it installed, until you have done your experimentation.
Firstly get the data sheet for the LDR and see if it mentions what wavelengths it is sensitive to and also how the resistance might change with ambient temperature.
Then you should definitely look at how the resistance changes over a full cycle from cold, through turning on, reaching maximum temperature, then turning off, and reaching cold again.
When the heater turns on from cold you would expect to see the resistance quickly drop as the LDR sees the burner. What happens then though? Does the resistance continue to drop as the boiler heats reaching a minimum when the temperature reaches maximum? If the LDR is detecting infrared then the resistance would drop since LDR resistance decreases with light intensity. However I have no idea what effect ambient temperature has on an LDR, maybe the data sheet will tell you. Importantly you need to watch what happens to resistance when the boiler turns off after reaching maximum temperature.
My money is on you seeing a quick increase in resistance as the LDR stops seeing the burner. It may not increase to the level you saw just before turning on, but if there is a quick increase that is all you need to detect turning off.
It sounds to me like you are using a single value of resistance to determine if the heater is on or off. I suspect that you should be looking for a sharp drop in resistance to detect turning on and a sharp increase in resistance to detect turning off and that those changes occur at different resistance values.
If you are using natural gas, or propane gas, and your burner is adjusted properly, the flame should be almost entirely blue. Therefor your need to be sure your LDR is made from Cds, cadmium sulfide, so it is most sensitive to blue light and not any other.
