After watching this video on YouTube: (wich compares the differences between WiFi and zigbee)
I was wondering; are there any good brands that makes good and reliable sensors using WiFi that are compatible with homey?
Because of all the zigbee related issues I have with my homey, I’m considering selling all my zigbee stuff and try something different. (offcourse I will wait with selling because of the zigbee rewrite, but untill that time I like to experiment with some WiFi gear)
Zigbee is on paper a nice protocol, but it’s not doing that much differently compared to WiFi. Range is not an issue as I’m having good WiFi in my complete house.
But I cannot find any app in the homey appstore with compatibility of WiFi sensors.
Yes, it does: Zigbee is low bandwidth/low power and it allows you to create cheap meshes. WiFi is high bandwidth/high power, and creating a mesh is (relatively) expensive.
As an example: the Xiaomi Zigbee motion sensors run off a cheap CR2032 battery. Mine have been in use for more than a year, and I never have had to change batteries. The recently introduced LSC motion detector (sold by Action) is a WiFi device, needs two fat CR123 batteries, and will likely run out of power within 3 months, if not sooner. And to save battery, it has a ridiculously large grace period of 5 minutes which can’t be modified.
Besides that, WiFi isn’t great if you have dozens of devices connected to it. For instance, you might run into issues with your router’s DHCP pool being too small (because each of those devices needs an IP-address). It’s also a PITA to add WiFi-devices to your network (unless they support WPS, which they almost never do), and WiFi is much more a “transport only” protocol, which means that each brand of device will re-invent the wheel and use their own data format to encapsulate sensor data. At least with Zigbee, some effort to standardization exists.
As far as I understand (I haven’t seen the video) one of his main objections is that Zigbee-devices are expensive. Xiaomi, IKEA and Samsung all make cheap Zigbee devices that work well (perhaps not with Homey, but that’s a different matter).
I do understand the mesh thing of zigbee… But when I have full coverage of WiFi, and every device can be reached… What is the point of having a mesh network?
About the DHCP pool, al my WiFi devices already having a static ip address without a problem. If I need to give all my sensors a static ip adress as well, I think this is no problem. I’m having a good router and can give all my devices static ip’s. Why should that be a problem? They are using just a little bit traffic, and regarding Paul’s video, this should not be a problem…
The power usage is offcourse an interesting thing… Paul is talking in his video about WiFi battery powered sensors wich can runs for more than a year.
Maybe they are using some more power. But with the costs of an AA or AAA battery, I don’t mind to change them every year. The batteries of IKEA are really good and super cheap.
But like you said, the LCS WiFi devices are really poor. Especially the motion sensor. Its useless. But maybe there are others wich working better. That’s the question.
Maybe you like to check the whole video. I like to know your thoughts on this.
Maybe i missed something, but why not use Z-Wave devices? The Z-Wave part of Homey works fine, or doesn’t it?
About the IP address part: i have already 50+ devices using a WiFi address and 75 devices on Z-Wave and Zigbee (that would occupy an IP address if they were WiFi). Since you normally have max 254 IP addresses to spare, that could become a problem in the future…
Z-Wave is an alternative indeed, but; I had some issues with Z-Wave in the past aswell (a couple of NEO Coolcam sockets who where responding very bad and I needed to re-add them every couple of weeks. So I sold them and replaced them by INNR sockets.
Maybe this is better now, I don’t know…
But for the rest of it, Z-Wave is a lot more expensive than most WiFi gear I have seen so far.
Thanks for claryfying the ip address part. I think, 254 devices including all sensors would be plenty enough. I’m not living in a castle. (not yet)
What exact devices / sensors would you like to have?
For Power Consumption i really like the TP-Link HS (Kasa) series, there is a working app on Homey and it shows status, Power consumption (if used with HS110 device). That count as a sensor.
Only having 254 IP’s in an “normal” home can be a case, but when you know how to assign Dedicated IPs you also should know how to extend the amount without any issues.
I think the main case why there are not much sensors for wifi is the Power consumption, Network sensors that are wired (PoE or power and WiFi) are other stuff and are not very user Friendly (at least you need to have one cable to each device) Battery and WiFi doesn’t fit good enough (at least in my opinion) therefor the WiFi Protocoll was also not invented, if was and will be a high troughput protocol and not for small (e.g. Temp, Humidity,…) information being transferred.
Nerverless, whereever it is possible a wired solution is in my case allways the better one. Don’t rely on Batteries will also be better for the environment and saves money ofer the time beeing.
A few wifi device is ok, but I have seen a local install running full wifi device and the latency can be very bad when the router is overloaded with all kinds of traffic…
Hi,
Unless you create a dedicated vlan, isolated from the rest of your network, I wouldn’t go there. Every WiFi device in your network has the potential to be a security risk.
A normal / common a-brand vendor in this time (2019) updates their hardware firmwares about 2-3 years if you’re lucky. A b-brand about 6 months if you’re lucky. (exceptions are there of course) There are also a lot of vendors that never update their firmwares, because that costs them money.
Any unpatched piece of hardware in your network can have vulnerabilities and become a stepping stone in to your network, become a spy for a buglar or become part of a bot-net.
This is exactly why I’m against all the iot devices that function via an Internet connection to operate (even in your own lan). An iot device that doesn’t need an Internet connection (let’s say a fridge, air purifier or thermostat) should never require an Internet connection. It should be usable without, when you’re in your own network. Operation from the Internet should always be optional.
This especially counts for us all, since we as homey users can operate our iot devices via the homey sphere, which only creates one potential risk rather than ten or more.
Also, the homey sphere is update regularly. This doesn’t provide guarantees of course, but in general, it’s a good thing.
It’s very hard to do that without aggressive power management, where the device (or at least its WiFi radio) is asleep most of the time. That’s why the LSC motion sensor has such a large grace period. And I don’t think it’s that bad in terms of hardware, it’s bad because it’s a WiFi device.
I cobbled up together WiFi sensor devices using ESP8266’s, using its deep sleep feature (where the entire device is sleeping, apart from the realtime clock, and it gets woken up every 15 minutes or so to take a reading) and my experience is that they will drain batteries too quickly for my liking. A few AA or AAA batteries really won’t cut it.
Just for you, I started watching the video, and to be honest, it’s pretty cringey:
“Zigbee is using on a separate frequency band from your WiFi router”. Both Zigbee and WiFi frequency bands overlap, and that can actually be a cause for interference as some Homey users have experienced.
He suggests that both with Zigbee and WiFi devices talk to each other. That’s not correct: for regular WiFi, you always need a WiFi base station as a middle man.
Even if WiFi devices were able to talk directly, they wouldn’t necessarily understand each other, because, like I said, WiFi is a data transfer protocol. It doesn’t define the format of the data being sent over it.
The price of Zigbee devices isn’t “at least twice the cost” of similar WiFi devices. Perhaps it is in the UK, or if you compare apples (Hue) with oranges (Chinesium).
The video quickly becomes a commercial for Ubiquiti’s Amplifi…
I really don’t understand the idea behind measuring the power usage of two 240V-powered lightbulbs to prove that Zigbee isn’t more energy-efficient.
That’s where I stopped the video, because to be honest, it’s not very good.
Youtubers rely on viewer reactions, you don’t get much of that by being nuanced. I watched this video the other night, and I consider his statements to be mainly correct, but not necessarily relevant. And they don’t change my plan to go for ZigBee/Zwave where I can. Initially I was tempted by the attractively priced Shelly switches and similar WiFi stuff, but after giving it more thought I warmed to the dedicated protocols.
The reasons have already been mentioned in the thread, security, mesh networking, power efficiency etc.
Now, if only ZigBee would catch up a bit. Small in-wall switches, motion detectors, led-drivers etc are hard to find, and of those, few are supported by Homey at the moment.
I did a year ago some experiments with wifi motion sensor “Neo coolcam PD02W”.
It Makes a wifi connection when activated.
There is no Homey app for it so i did it with the app for my unifi wifi access point.
There was a huge delay, not useable for switching direct a lamp.
The sensor is powered with a CR123 battery.
After 3 weeks empty battery. afterwards every week.
I stopped the experiment.
Now there are 2 unusing PD02W motion sensors in my kast.
IMHO I’m tired of zigbee devices disappearing every now and then and having to press a switch to exclude/include for it to work again… so I’m currently testing a few Shelly WiFi devices.
I have perfect WiFi coverage with several Unifi APs and have all IOT WiFi devices on its own subnet, works fine so far… Shelly have a few sensors as well
Thats exactly why I was looking for an alternative… Interesting to see that Shelly have some sensors aswell. Didn’t know that. However, the most important sensors for me (motion and door/window) are not there yet as far as I can see…
They have a motion sensor, but I haven’t tried it yet. Have just ordered Flood and H&T sensor for testing there claim of 18 month battery life. But I would be surprised if i’t over a year, so I plan to use 3 CR123 for each Flood sensor or a CR123 to USB adaptor where possible. Small door/window wifi sensors will probably not work very well, or need frequent battery changes…
The general comment here is “you shouldn’t … etc” but if you want just go for it.
I’m in the same boat as you moved from zwave switches to wifi and now slowly to Zigbee. there is a lot to be said for wifi devices when they make sense.
For example I could see a “BRUH Multisensor” in a room, alternatively a RPI with Kodi and some sensors reporting things through MQTT, the device is there why not expand it and use it more frequent.
The thing is you will need to build something yourself, the thing off-the-shelf is quite limited
