I have been lurking around in this forum for quite some time and used openHAB 2 and for a few months now OH3. Currently I am working with a few colleagues on designing a class of battery free devices (Think RFID, but as Sensor). Currently this is mostly done for industry applications. Passive sensors are used wherever electronic/batteries do not survive long enough or add to much weight.
Main idea for me is that I in the past kind of disliked the reliability of battery driven SmartHome devices, while at the same time cabling everything is not an option to me in a rented flat. I believe reliability has improved a lot over the last years (ZigBee ftw), but I kind of dread the idea to put sensors on every door and Window, as it would increase maintenance of my system.
So our idea is to put one cabled reader device in the middle of the room, i.e. like next to the light where your smart lightbulps are and where there is electricity and use this to read out passive sensors (door/window-contacts, buttons, temperature).
I am meanwhile wondering if this is just me who wants that/sees value in it or if there is more people out there who think the same?
Also, if yes: What would you be willing to pay to exchange all sensors in one room to be passive and have just one active cabled device? (200âŹ,500âŹ,800âŹ?)
Is there features you think this device would definately need. e.g. (Wifi+mqtt,REST api, or Thread+Matter support)
I donât think if i would spend more the 200⏠on this per room, also i donât want one âgateway/readerâ per room.
Most of my sensors are Zigbee sensors from IKEA or XIAOMI.
So i have 1 âGatewayâ (70âŹ) for the whole house, ~2 window contacts per room (10⏠each), 1 temperature sensor per room (10âŹ) and ~1 button per room (10âŹ).
So in a House with 7 rooms, i have overall costs of 50⏠per room (only for the door/window-contacts, buttons, temperature + gateway).
I have to swap the batteries once a year, some devices will last for 2 years, and this batteries a really cheap so no problem for me.
But you should look at the environment, so battery-free devices would be the best solution.
So i like your idea.
But how do you read RFID over several meters ?
And how are these battery free door/window sensors are triggered/when do they send data?
The gateway should use Wifi+MQTT to support multiple smart-home-systems (not only OH).
Also an API for future binding development would be nice, but not important as MQTT.
It is not RFID sensors, but SAW sensors. Basically they are though working the same. Just analog. You send in energy with an antenna and receive the energy back again, but much smaller. i.e. these can not send data on their own, but you could for example read out a window sensor every second and check if it is still closed.
You battery strategy is to have a battery change day once every year so you donât run into trouble of having to search for the one device whose battery died this week? (If I count roughly 35 Battery devices here?)
Its a question of time how long this pulling takesâŠ
i.e. think of this pulling taking a microsecond. So it is not that âoftenâ that the sensor becomes active. Even if you would check the sensor once a second with 5 Sensors everything is âmostly offâ or on for 5 out of 1000000 microseconds.
Conceivably you could poll less frequently but store energy for unsolicited reports?
There are venerable technologies like Modbus or even mouse that rely on polling, but in general latency is bad. An extra second delay between opening a door and the lights coming on is annoyingly noticeable to humans.
Variable poll rates are helpful, allowing sharing of available bandwidth between âimportantâ door openings and less time sensitive temperature readings.
To paraphrase @rossko57, latency sucks, and one second is a long time. One second is fine for temperature sensors, but makes a button almost useless to me. Same could be said for a door contact if you want it to turn on the light when you walk into a room. You also have to consider the additional latency from sending reports to the system.
I donât think this question can really be answered without more specifics, as each personâs space and needs will be different. You need to say how much range each reader has, and if range is impacted by walls/appliances. Only then can people determine whether it fits in their space.
If youâre concerned about renters, then your reader has to be something that can be plugged into a wall outlet and not require messing around with a ceiling fixture. Also, not all rooms have ceiling fixtures.
Iâd go in a slightly different direction. Instead of emphasizing âone reader for lots of sensorsâ, focus on âan AC-powered, plug-in reader for cases where you donât want battery-powered sensorsâ. As an example, a lux sensor placed outside, or a temperature sensor in a fish tank or high up a wall. I think itâs better to give it a very specific intention/purpose than to try being all things to everyone.
Does it have to be the middle? Iâm considering all my rooms and the middle of the room is not a good place to put anything. Consider a bedroom. The middle of the room is the bed. In my office the middle of the room is on the floor. As I consider it, I donât think I have a single room where I could but a device in the middle of it.
I canât be unique in this. It would be easier to put a device like this near the center of a given floor to cover multiple rooms.
I definitely see the advantage of wireless sensors that donât need batteries, especially for things like contacts on windows and doors.
How many sensors would this come with? Itâs the sensors I care about, not the hub. In any case, 200⏠is probably more than Iâd pay for most of my rooms. I can get enough relatively expensive Zigbee/Zwave sensors to cover all in one room that I need for less. I may be unique but Iâm not willing to pay that much of a premium to not have to change batteries that last 6+ months anyway. But Iâm also not someone who has a strong interest in this anyway. Most of my windows have no sensors and Iâve no plans to add sensors to them.
MQTT support would make it instantly supported by almost all the open source home automation hubs. If you use the Homie and/or Home Assistant MQTT standard OH (and others) could even automatically discover the devices.
Thread+Matter would give you support with all the big companyâs hubs and such but not openHAB at this time.
Most battery devices will report the current level. You can easily see the current levels of all your devices and change the batteries as needed.
It depends on the device and the amount of latency. Does it really matter if you learn that the window was opened one second after it actually opened? Maybe yes, maybe no. For something like temp and humidity itâs probably definitely no. Those change slowly enough that once a minute is probably overkill. But if you have a motion sensors that one second is way too long. Button presses a one second latency is too long as well. Even half a second is noticeable.
Ultimately almost everything is polling all the time. Itâs usually just that the polling is taking place on the microcontroller on milliseconds timescale. Push is usually an illusion, a very helpful and convenient illusion. So I donât have too much of an issue with a one second polling for something like this, for most sensors at least.
But I totally agree, it makes no sense to report every sensor every second. I would expect the hub to filter most of the reports out and only publish significant changes or on a longer timeframe.
Hah. I interpreted middle of the room to be a ceiling fixture, because I subconsciously ignored the idea of having a plug-in device on the floor. Now I realize that probably isnât whatâs being suggested.
The reach of passive sensors is very limited. Reading a sensor beyond 4 meters of distance becomes troublesome. Also a line of sight is necessary at this distance. Therefore an elevated position in the room with direct sight to buttons/doors/windows and being in range to windows in all directions makes the ceiling in the middle of the room the place to be. A wall-plug will just not do for this.
I also agree that depending on application different polling times are apropriate⊠i.e. for a button you might want a tenth of a second while the door may be fine with 2 tenths and the window can probably be slower than one poll per second.
I was also wondering what I would want from such a device, other than MQTT connectivity.
If I screwed something like this to the ceiling, for me, I thought adding some microphone to it to get simple voice command (e.g. 5 simple voice commands to run scenes in addition with the sensors âpoor manâs Alexaâ without any cloud connectivity is doable with a larger microcontroller).
My question here was also targeting at: Maybe someone has more/better ideas like thisâŠ
This severely limits the practicality of this design. None of my bedrooms have light fixtures on the ceiling. Most of my living spaces have chandeliers, track lighting, or recessed LED lighting. In the few rooms where I do have a light fixture in the center of the room, unless you somehow power the reader using an A19 or E17 light bulb socket while still allowing the light bulb, it will be of even further limited utility. Thatâs a challenge as most ceiling light fixtures donât have room to add extra length to the light bulb. And even if that works that means youâd have to keep the light switch on at all times pretty much requiring users to deploy and use a smart bulb(s) to that fixture.
I think in order for this to be practical for enough users to make it worth while, youâll have to figure out how to make the reader either work from a wall outlet or make the reader itself battery powered which kind of defeats the purpose, or otherwise find a solution for how to power the device. Maybe just the antenna needs to be on the ceiling and the device can be near a wall and a power plug, but that means running a wire up the ceiling which also defeats the point.
Accessing the cable the same place as your Chandelier would have been my idea for installation. And yes this forces you to have the original lightswitch all the time on while controlling the smartbulp with a switch/button. I thought this kind of setup was already pretty much standard for most rooms in the smart home. Lights are kept on all the time for smart bulps to function and being switchable.
Many of us specifically avoid smart bulbs expressly for this reason. I wonât install smart bulbs because IMO itâs far less convenient to have to interact with some app as the only way to control the lights. With smart switches I donât have to train family and house guests on how to use the house. I donât have to worry about someone accidentally turning off a switch. I donât need to create some fancy tablet to allow someone to tab a screen to turn on a light.
Unless one really wants to control the color, I just donât see the point of smart bulbs. They make things more complicated. Home automation should be about making things less complicated. Smart bulbs are a huge reason why when people bring up home automation many/most people roll their eyes.
And while I have a strong opinion on this, itâs not a unique one. There have been tomes worth written about home automation design and why to use smart switches instead of smart bulbs.
I am re-evaluating my mental image of Palais de Koshak.
On-topic
Smart bulbs? Toys for playing with, not for primary lighting. Want something disaster proof as well as visitor friendly. âPlease find your phone in the dark while SuperSystem restores a corrupt file.â
A sensor âhubâ might usefully add temperature and humidity sensing for little effort. Microphone and perhaps ambient light also have applications in presence sensing as well as the obvious.
Wire-free door and window sensors remain the primary attraction.
I think thatâs a non-starter. Itâs far more logical for a renter to attach a battery-powered sensor to a wall/window/door using 3M strips than to wire something into a ceiling light. More importantly, most renters shouldnât be messing with high-voltage wiring.
It feels lilke youâve come up with a technical solution, and now youâre trying to justify it for consumers. However, the only real advantage I see is that it doesnât require batteries. Other than that, itâs not better than a Z-Wave or Zigbee sensor that has fewer physical limitations.
Iâd also keep in mind that the home-automation market is small, and smaller still if you remove the people who just want to plug something into a wall and connect it to their WiFi (which is most of them).
I can see uses in commercial/industrial settings; pop a âhubâ in the overhead trunking or ceiling void, powered by PoE.
On the other hand, in such settings youâve probably already got a burglar alarm that can share its door/window events.
I have thought so too ⊠In practice though they seem to work really well even with visitors.
i.e. They default to power on after power loss.
Meaning if a visitor switches your normal lightswitch off the light stays off⊠he/she will automatically press again and then the lights will go on.
Also placing switches near the normal lightswitches helpsâŠ
Its not perfect, but the smart bulps work better than I have ever expected/would not want to miss them for primary lighting.
And as with the rest of the smart home⊠anything that is essential is not controlled by smartphone but a physical button.
if you are looking for wireless sensors, buttons, ⊠without batteries you may have a look at the Enocean alliance. That devices using energy harvesting and ultra low power electronics. You find that products build by different companies that licensed that technology, e.g. Eltako, Wago, Jung, Honeywell, âŠ
I think this is a solution with way better acceptance as the passive sensor approach. Iâve nearly changed all lights to smart lights (with different technologies from starting from Philips Hue over Tasmota-flashed WiFi-Bulbs and H801-stripe controllers and some milight-lamps to self-build pixel-led controllers) but I would never use a gateway beside every lamp on the ceiling. All doors and windows are equipped with contacts, even all my household appliances like oven, hob, dishwasher, coffeemaker, refrigerator, ⊠are smart (Bosch/Siemens HomeConnect) so I would classify myself as an âSmartHome-Nerdâ, but a classical light-switch is a must have. All of my lights are controllable with wall-switches (most of them are from HomeMatic some from EnOcean). And for all âcore lightâ (essential lights, not only ambiance) I have a fallback control in cases the complete SmartHome-System fails, no matter due to an software issue, gateway failure, âŠ
Just replacing all lamps with smart lamps, no matter which technology, is not a good idea. It can always happens that a gateway fails, the wifi-AP/-Router dies, ⊠you need to think about all contingencies.