# Water detector for heating element?

I want to run a 300W heating element in a water bath, but I'd like to make sure that it isn't turned on by mistake when it's out of water, so I thought of connecting it to a controller and a sensor, though I'm not quite sure about the best way to sense this.

I thought of two options:

1. Detect water with an open circuit - I'm a bit worried about electrolytic corrosion though. Is this actually an issue?

2. Detect the temperature with a thermocouple attached to the heating element. This has the difficulty of attaching a thermistor without it melting.

What would you suggest?

• Why not use a float switch? – Ignacio Vazquez-Abrams Aug 29 '13 at 4:13
• @IgnacioVazquez-Abrams - because I'd prefer not having a moving part, and I'd like the heater (sensor included) to be as compact as possible. – nbubis Aug 29 '13 at 4:15
• Some sort of optical sensor then. Use the refractive property of water to detect presence. – Ignacio Vazquez-Abrams Aug 29 '13 at 4:16
• @IgnacioVazquez-Abrams - sounds like a good idea. Any waterproof suggestions? – nbubis Aug 29 '13 at 4:20
• See Kortuk's comment: allowing a suspended user to continue posting under spawned accounts is not really a productive environment. I can tell you that it is not as black and white as it may seem, but a suspension does not leave a stain on someone's record forever. The only way a users' suspension can be discussed publicly is if they (the user) disclose the information. We do always go out of our way to resolve situations; suspensions are not ideal – Connor Wolf Aug 29 '13 at 8:55

Water will diminish transmitted infrared light very well, by altering the lens characteristics and thus dispersion (diffusion, refraction), as well as by absorption.

See this answer for the absorption graph: The red end of the visible spectrum (and more so the Infrared) is absorbed by water about 100 times as strongly as violet/blue:

Hence, a low tech approach would be to use an infrared interruptor switch, such as the GP1A57HRJ00F:

(source: SparkFun)

These devices work by emitting infrared light from an IR LED on one side and sensing at the other side. The gap for this particular device is 10 mm, which should be plenty for the purpose.

Adjust the IR LED's current to where the output just barely but reliably switches on in air. When the device is immersed in water, the increased attenuation of IR by water will cause the output to go low.

To make the assembly waterproof, coat the pins, soldered connecting wires and any other exposed metal with some suitable high temperature, waterproof epoxy. For production volumes, you may be able to find assemblers who can do the epoxy coating and insulation immersion test for you in bulk. In any case, the voltages involved are so low that there is no electrocution or catastrophic current flow risk if the insulation fails.

You might also consider an installation design such that only the extremities holding the IR LED and sensor would be immersed in water, the contacts would be above water level. Insulation is still required to survive splashes, but the criticality reduces.

• Hmm.. Won't the hot water emit a lot of IR, thus turning off the heater? – nbubis Aug 29 '13 at 6:08
• @nbubis The IR sensor in this device is most sensitive at 900 nM, of the sensor's 400 to 1200 nM range. The IR emitter works at 950 nM. Firstly hot water will have little emission in that region. Secondly, what little there is will be very low intensity compared to the LED's output even at very low drive current. To be doubly sure, you can add a pulsed modulation to the emitter drive, and filter for that modulation on the output signal, as is done in IR remotes. This is a simple enough device to try out an experiment and see for yourself whether it works for your requirement. – Anindo Ghosh Aug 29 '13 at 6:18
• And now we run into the last problem which is to find a sensor that works above 85C (say to 110C) :) – nbubis Aug 29 '13 at 6:51
• @nbubis Hmm, that's a tougher problem. I have just now tried illuminating LEDs while they are immersed in boiling water for extended periods, and they work fine, but I doubt the plastic casing or the sensor would work acceptably. The sensor has a BJT inside that might go nuts at high temperatures. Perhaps someone can figure out a non-contact approach. – Anindo Ghosh Aug 29 '13 at 8:02

I'd suggest looking at what's done in commercial / industrial products as this is not a new or rare problem to solve. Everything from £5 kettles to multi-million pound industrial process machines tackle this issue in a variety of ways.

Kettles, steamers, washing machines, dishwashers, cookers, irons, steam cleaners, slow cookers, fryers, etc. all have excellent mechanisms for handling this which are cheap, reliable, and by law must be incredibly idiot-proof.

Your options are many, off the top of my head and by no means exhaustive:

• A PTC element that can't burn itself out / overheat if there's not water present
• Thermal cutoff switch or thermally coupled switch
• Float switch
• Pressure switch (as used in washing machines to detect water level, hence common, robust, reliable)
• Optical switch (which would require a driver/control circuit)
• Capacitive or ultrasonic transducers (require even more control circuitry)
• Old-fashioned and almost-never-used-in-real-products conduction type sensor you mention

If your goal is to maintain a particular water temperature, I'd suggest having a temperature sensor a small distance from the heater (perhaps more than one), and watch it/them to ensure that their behavior is in line with what should be expected in the presence of water. One should have a fail-safe to ensure that no hazardous condition will occur if the CPU turns the heater full-on when it's dry, but it should be pretty easy for the CPU to identify such conditions. Note that if you use two sensors you may be able to achieve much better regulation than when using one (use a sensor located away from the heater to determine a target temperature for water near the heater).