Heat powered Zigbee & BLE devices

Hi Everyone,

I just want to share a new Zigbee and BLE device that I have been working on. Some of you may have seen my work with BLE gateway with this module. The module that I use is capable to do more than becoming a BLE gateway.

This device is based on nrf52840. It is a multiprotocol module made by Nordic Semiconductor. It is one of the leading company in BLE solutions in their early products. Today, with their nrf52XX and nrf53XX chips, they are putting solid solution in other protocols such as Zigbee, Thread and BLE.

At this point, I am not making a final release device. It is more like a dev kit for myself. There will be variant of the device which will solve a specific problem or use scenario.

The demo is a combination of 3 demos in one.

In the demo, you will see thermal energy in action with BLE and Zigbee device. A couple demos are about BLE beacons that can help detecting your gas stove. The last one is Zigbee switch to control your fireplace. a Zigbee energy harvesting device is quite rare especially the non Green Power device. I am making an end device implementation of Zigbee device in this demo.

I want to take this opportunity to invite member here to share their input. It is most welcome. I am early in development. I can modify the design if there is compelling need to solve some problem.

There are a lot of energy around us. I am just scratching the surface of what is possible. There are energy sources from water, light and vibration. Combining them with the right MCU and software can solve a lot of issue for all of us.

I hope you enjoy the demo and thank you for watching.



Neat. Timely too. I see you commented in the thread where this was just being discussed. Can you monitor multiple thermocouples, or do you need one module per each?

The power source "thermophile" cannot be multiplex to a single module. The voltage generated is too small for any multiplex mechanism.

But, the module can be connected to a thermocouple to measure temperature. This can be connected to multiple probe. In fact, I am looking to make a device that can support up to 16 thermocouple device.

1 Like

Like any thermoelectric device this needs a temperature differential across the 2 sides of the plate in order for any current/voltage to be generated.

I don't want to sound negative as the specs (and price) of the SP1848 look pretty decent to power small devices given the right conditions but if people reading this are thinking that you can just stick one in a hot environment to power a device then they would be quite mistaken. Even if you stuck it on a hot plate at some point the increasing air temperature above the other side of the device would soon reduce the power output.

But I love the idea of harvesting energy (rf/motion/peizo/thermal/solar /wind) but I guess there's a reason why only a few are widespread :smiley:.

1 Like

This is a sound advice. 100% correct. There will be steady state the TEG device will be settling down to generate small power. At this point, the "dance" between the TEG and MCU with the software choreographing the move between all the component is the key. Arduino strategy such as loop() { do something; delay(100); } won't get this running.

The demo is just a short run. I have to confess that I have started the project quite sometime ago. I have run multiple days/weeks using TEG running BLE and Zigbee stack at 40 degrees Celsius plate with room temperature of 20 degree. The device does not skip a beat.

My plan is to use the TEG on my Smokey Mountain smoker. I will mount the TEG on the door with some heat sink. The door get the heat from the internal chamber which we will try to keep around 107 degree for long smoke session. In winter, fall and spring here, I am not concern. There will be delta temperature between the hot and cold surface. In summer, the outdoor temperature can reach 35 degree in bad day. I am quite optimistic at even at that point there will be some heat to harvest.

Anyway, thanks for bringing this point up.

On the other hand, thermopile generate energy as long as it is exposed to a flame.

1 Like

Very nice. But what about summer when ambient is way higher :grin: :grin:

How much voltage/current can you realistically get?

1 Like

Any of the TEG and Thermopile generate not more that 1 volt so far. In the demo, the thermopile reach .65 volt. This is .1 volt off the spec. TEG is much more powerful when you have high temperature delta. But, one will be mistaken to be able to count on this as you mention. To use TEG long term, one need to wait until it get into steady state. Then, plan the energy consumption base at this state. I would say actually one will get approximately similar energy input from thermopile.

I do not have current measuring tools at this voltage. But, during development, on the MCU side, I would not count on consuming 1 mA continuously. On average, I am aiming less than 50 uA at 3.1 V consumption just to give you an idea. This is why the dance is so important. Sometime, you have to use 20mA power. Then, you have to recover the loss while the energy harvested is coming in.

In BLE case, you may be able to get your power primarily from the harvester hence smaller capacitor. In Zigbee, this is not possible.

1 Like

Cool. Looking forward to seeing more results of your development work!

1 Like

Hello Everyone,

Just want to update everyone here on the Zigbee Heat powered switch. This is the first module that I am concentrating that will use heat as its main power source. The switch is design to control millivolt valve for your fireplace. Just FYI, I will still make others modules that will use heat as power. I have boards for detecting gas stove and temperature probe for your BBQ grill that is powered by heat.

Today, I just want to share that the fireplace switch is almost ready. I will have one more board iteration to clean up and adjust some of the board configuration issue. The firmware is ready and being tested by me at this moment.

Google Photos

Here are some the highlight of the switch that make it different from some that is available in the market.

  1. The switch run from the heat powered by the thermopile from the fireplace.
  2. The switch will operate during power outage assuming the hub is still running.
  3. The switch will have an optional setting to disable "on" command from the hub(Zigbee switch on command). This logic is protected in the module itself (not from the DTH or hub). I am aware and agree with some of us who are concern with being able to switch "ON" fireplace remotely. We can still switch off the fireplace remotely. This will allow us to automate the fireplace with other devices by observing the status of the switch. There is a physically wall switch that is functional to turn on/off your fireplace. This will force someone to toggle a physical switch to turn on your fireplace. Yet, you can power off your fireplace with anywhere, anytime and any how.
  4. I am testing the switch with multiple hubs such as Hubitat, Smartthings and Home Assistant. The switch should be compatible with any hub that support Zigbee 3.0.
  5. There is an optional custom Zigbee command called "timed on". The idea is to turn on the fireplace with an interval information. The switch will turn off the fireplace after the interval elapsed. We can eliminate the need for the hub to send switch off command. In the case where the hub is down, the fireplace has a way to switch off in this case.

Here, how it looked like installed on my fireplace. I am thinking to update with YouTube demo.

Google Photos

There are a couple wire patches that need be done at the current iteration of the board. The next iteration of the board have all those fixes. At this point, I should be able to share them with anyone who are interested. In the meantime, if you have any feature that you would like to see, please feel free to share it here.

Would it be cool for us as the community to have have modules that harness power from its surrounding? I feel that it is cool. Yet, more importantly, lets us take a leap on the technological front. I feel that it is time we should make a better switch that depend less on conventional power such as battery and power grid when possible. It will reduce the cost physically and mentally maintaining our devices.

On separated note, I am just a hobbyist. I funded this project from the proceed that I got by selling my other Zigbee module that I have developed. If you want to see this project to grow and looking for a Zigbee Environmental Sensor, please consider give the Environment Sensor a consideration.

I apologize for the long update. I want to thank you for your time and interest. I also want to thank everyone who had supported my other projects.



Just small update. Here is the latest board for the Zigbee fireplace remote control.

I finalized the design with a case in mind.

There are 3 wires from the gas valve and 2 wire for the wall switch.

Google Photos

Google Photos

There will be more updates for details coming soon.


Hi Everyone,

I have a quick update. Here is another device that I am working on that harvest energy from heat as its power.

As you may have seen on the original video, I made a BLE sensor that can be used to detect whether your Gas stove is on or off. I have consider many pro and cons. I decided to support Zigbee rather than BLE. Going this route require me to design a hybird power source rather than pure autonomous sensor that is powered from the heat. However, for Hubitat users (all of us here), it is just much more convenient as ZigBee is natively supported.

Here are some pictures of the board.

Google Photos

Google Photos

Here is my test setup. I plan to hide the module under the stove in the future.

Google Photos

One thing that it can do is to do rough measurement to the heat in voltage. It is not precise but we would be able to tell how high is the burner is set to. The sensor use 2 AA battery. Since I will try to hide it and it may not be convenient to replace the battery every couple of years if I use smaller battery, I would go all the way with much higher capacity battery for long battery life. I expect that it should last 10+ years with the battery alone. The harvester will add to the battery life especially if the stove is use frequently.

One of the thing that I can do with Hubitat is to have it announce when the stove is in use on google home.