Remote fire up of car heater

During the winter it’s great to have an additional car heater that can de-ice the car without the need to start the engine. This results in a nice and warm car when you’re about to leave.

These systems exist. But they are incredibly expensive. Basic heaters start at around 1000 euro and then they don’t have any internet connection. Those that do have it are easily 1500 euro. One aspect that makes them that expensive is that they usually heat up the cooling water which adds complexity, but in the early days those devices worked different and usually they were just air heating devices. But those things would still work for modern cars and because they are not very complex it’s easy to control them.

I happen to play around with classic cars, so here’s the plan: I’ll buy a vintage heater including an overhaul kit, overhaul it and add a node to it. A pretty straightforward arduino based solution with a few relays should do the trick. Budget: 200 euro.

No more scratching the windows :sunglasses:


Would be awesome if you could share this project as a group on

I agree with @wienke. I already bought a electric heater (on a seperated 12v battery which charges when driving) two years ago for a project like this, but never finished it. Now my LoRa enthusiasm will probably help me to restart the project.

What plans/ideas do you have to activate the heater? I think it could be several things:

  • Auto detection of iced windows
  • Activated by a (responsive) web app (button or schedule)
  • Activated by a LoRa key remote

@TijnOnlijn Let me know what you think, or if you need any cooperation in your project!


I’m currently focussing on the gateway, this is next in line. The idea is to use a web app to start it. I won’t consider ice detection, because first it increases the risk of starting it up too early, and second even if it is 1 degree above celcius a hot car is still very nice.

I think using a battery is insufficient. Unless it’s large but then the need for a daily charge would spoil the fun a bit. Some years ago I used a 220v heater for the same purpose and back then it took 0,5 hour of 2000w to warm it sufficiently. I may not have made it clear in the OP but that’s why I’ll use a gasoline heater. (I moved, and the car can’t be on the drive next to the 220 anymore). But if you’ve got reasons to believe it’ll be otherwise please let me know. It would simplify the project for sure.

Coorporation is always welcome. If you consider doing the same it might be good to buy the heater already since they are cheaper in spring than in autumn.

I’m also currently focussing on the gateway. It’s running, but need to attach the outdoor antenna which is already mounted on the roof of our office (same antenna as you using) and installing the system on a proper place. Next I want to do some coverage measurements and after that starting the heater project.

I didn’t try anything yet with the heater I own (I guess it is a 300Watt 12V heater), but my goal in this is only to de-ice the windows, not to make a comfortable temperature inside. Futhermore I don’t want to make any changes in my car, so a gosline heater (like a weaco system) isn’t a option.

Anyway, maybe the activation of the whole system could be the same? For example a button on a web app, which toggles a state machine on a server. The node is polling every minute to that server to check the current state and activitates the heater.


That might not work, see Limitations: data rate, packet size, 30 seconds/day Fair Access Policy, nodes per gateway. But in a car, the power source might allow for a Class B end-device, which I hope TTN will support one day.

After writing my post I realized I may work if you specifically heat the windshield. Is your heater of that type that is pointed directly at the window? Something like below (random image)? It could be a great solution since for sure it’s easier to build.

I guess most of the soft- and hardware would be the same. But I don’t think it’s needed to put the state machine on the server. If you put it on the node you’ll have the same result with a minimal amount of data to be transferred.

I know about the limitations, but since it can be only one byte (only need the response message from the server if the heater needs to go on for a predefined time), I tought maybe it is possible at a 1 minute rate?

It’s almost exactly the one from the image. I tought a online saved value (on/off) is neceserry because class A is the easiest (and probably only class available at TTN at the moment), so we have to send data to the car in the downlink message.