# Thermometer type in electrical water heater

I'm working on my home monitoring system and already succeed with electricity consumption measurement and got closer to water system.

Measuring water consumption is almost solved, but as we are using boiler almost 365 days a year I would like to measure water temperature in a tank as well to have more info on it.

As the tank itself is tightly isolated I've got an idea to intrude into boiler and use it's temperature sensor.

Water heater has electronic (buttons) control (model ARISTON VELIS 50)

Under the hood it looks like this (2 tanks)

And temperature sensor for this model is the following

As it seen on the picture the left thermostat has 3 wires which I suppose are:

• GND
• VCC
• DATA

whereas the right tank has kinda 4 wires going out the same place.

I'm wondering:

• if its possible to know how to read the temperature data from 3 wires sensor or how to know the "protocol" its using for data transmission?
• is it possible to intercept this data not interfering with the boiler board?

Maybe there are some traditional/known approaches for such kind of tasks or this is regular temperature sensor, but I don't have a clue what to start with bec. organizing a workspace under the boiler not knowing what to do is quite uncomfortable.

Unfortunately I don't have an oscilloscope to read raw signal, but I have both Arduino and RaspberryPi3 and multimiter in my arsenal.

• Can you find any datasheets or other information on the sensor? – ambitiose_sed_ineptum Oct 13 '16 at 18:54
• That electronics does not look like it has isolation from mains. Do not mess with it, mains voltage is lethal. – Turbo J Oct 13 '16 at 19:01
• Is the yellow wire labeled DATA? Or is it your guess that the signal is digital? Or did you observe the signal with an oscilloscope and saw that it's digital? [In the meantime, I'd like to second @TurboJ 's point about isolation from mains.] – Nick Alexeev Oct 13 '16 at 19:04
• After much searching through the site mentioned in the image (A LINK WOULD HAVE BEEN NICE!!!) I finally found the page for it. There's a magic word in the name of it: "ARISTON sensor NTC VELIS 1500W" which tells me it's a Negative Temperature Coefficient thermistor. It may be wired as a half-bridge wheatstone. – Majenko Oct 13 '16 at 21:25
• Also they're nice and cheap. I'd suggest buying one to experiment with rather than risk blowing up your boiler. You could also examine it in more detail and even dismantle it. – Majenko Oct 13 '16 at 21:34

From your photo it appears that each sensor is inserted into a closed-ended tube that projects into the boiler, also known as a 'thermopocket' or 'thermowell'.

Depending on the size and shape of the thermowell, you may be able to insert your own sensor into it alongside the existing one. Because the boiler electronics are not isolated from mains, you then need to use isolated signal conditioning to measure the temperature from the sensor and transmit it to your home monitoring system. A type J or K thermocouple would be a reasonable choice of sensor: you could get better accuracy with a platinum RTD but you don't need it for this application and it'd cost more.

You could measure this with an analogue transmitter, which typically converts the output to a 4 - 20mA current signal. Typically these are loop-powered: you connect the device like this

simulate this circuit – Schematic created using CircuitLab

where the 'meter' could be a resistor across which you measure the voltage. $$\V_{supply}\$$ needs to be enough for the minimum voltage requirement of the transmitter plus the maximum voltage burden of the meter. Or you could use a data acquisition device from which you would read the measurements digitally - the type I've linked uses an RS-485 bus connection (and 'clones' of this type are widely available from different manufacturers) but you may be able to find other types.

As another answer notes, there are some important safety considerations here:

• you must make sure that your temperature sensor doesn't electrically contact or interfere with the existing sensor
• you must treat anything that could come into electrical contact with the boiler electronics, even by accident, as mains voltage wiring
• you must satisfy yourself that any signal conditioning unit is suitable for your purpose - the ones I've linked to are examples of what is available and not necessarily a recommendation
• you must make sure you understand what is isolated from what by the signal conditioning unit, for example is the power supply connection on the input side of the isolation or the output side

If you are unsure about any of those, don't do it.

• Thanks a lot for pointing out AT and ADAM-type devices. Now I know few more devices. Also thanks for pointing out all considerations. I indeed stopped thinking about tinkering into the boiler circuitry as it will definitely make it unreliable and it doesn't worth that for sure. – Paul T. Rawkeen May 15 at 20:11

I believe that @TurboJ is correct that there is no galvanic isolation. That would make the signals very difficult to deal with safely- the sensor appears to be tied to one side of the mains.

As it's clearly (well, after the good research done by user222030 and Majenko) some kind of dual NTC resistance probe it will be very nonlinear and referenced to some voltage such as 3.3V or 5V on the board. You would have to reverse engineer the circuit and then you'd still have an nonlinear nasty voltage connected to one side of the mains, ready to blow the living **** out of your home automation system, damage the boiler controls, or electrocute the hapless user.

While you could attempt to introduce another sensor into one of the protection tubes it would have to be insulated sufficiently for the mains to be safe.. which may be non-trivial and I don't think it would be responsible to try to make suggestions remotely on how to do that.

• I started thinking the same about intruding in circuitry after closer analysis. Good catch on adding additional probe inside existing well, didn't think about it actually. Not sure I will try it, but smth. to keep in mind for the future. Also likely non-linearity is what actually stopped me bec. I'm not sure I will be able to handle it correctly. What I will probably stop with is determine the characteristics of NTCs to call it a day. The efficiency I wanted to calculate basing on temperature can be calculated indirectly (not precise) and I will probably do it this way. Thanks for your time! – Paul T. Rawkeen May 15 at 20:23

No. Those aren't digital sensors. Those are two NTC resistors, with a common lead. I have somewhere resistance/temperature table. I can try to find it if you still need it.

• Would be very nice if you could find such data as it would give me possibility to compare smth. against. I paused this project quite a while ago, but will continue on any new solution/data I will gather. Appreciate your help. – Paul T. Rawkeen May 15 at 19:52