Thanks for your note. I'm not sure if I should feel better (perhaps I am not an idiot) or worse (so far my 'home automation' experience feels like building a car with a duck tape). I don't know how much time I should be wasting trying different drivers, repeaters, unpairing, re-pairing, etc. before I toss them into the bin and buy another dodgy set of crappy AliExpress junk by another Chinese company of questionable quality. I think I need to start buying 3x fewer devices, but 3x as expensive and hope they will work (and keep working even if I move my night stand, open a window, or turn on my oven).
This is a good philosophy. While their design aesthetic is pleasing, and they are low in cost, Xiaomi devices are not bog-standard ZHA1.2, and are known to be problematic even on zigbee networks where all the repeaters are Xiaomi-compatible.
I know - I had ~60 of these on Hubitat. I still use them (currently via two Mijia bridges, but probably zigbee2mqtt in the future). I would have been better served pitching them in favor of devices that work natively with Hubitat.
Makes sense. I've learned my lesson -- do better research before buying anything. I know it's difficult to generalise, but what brand would you say offers good value for money for buttons, switches, motion sensors, etc. so the whole portfolio? I've tried to find a thread that would cover that but feels like everyone has different answer to it and it's almost impossible to get a consensus.
In general, while they are no longer available, people here have had excellent experience with Iris v2 and v3 zigbee sensors. I have a drawer full of Iris v3 sensors, but never got around to replacing my Xiaomi gear with them.
Iris v2 products were made by Centralite, who still exist.
Hmm.. Not to rub it in, but they are working rock solid here..
I have 15+ buttons (different models; round, square, one and two button models, D1's and a Opple 2-button) along with motion- and door sensors in most rooms..
For repeaters i use a couple of IKEA outlets and repeaters, some Sonoff Zigbee ZBR3 switches and Aqara Zigbee relay switches..
Which driver are using..?
I use the drivers from marcus via HPM with great success..
Also make sure to set your Zigbee and 2.4 Wi-Fi channels far apart from each other.
I use 20 as Zigbee and 1 as Wi-Fi.
Zigbee and Wi-Fi are both communicating via radio signals..
They both operate in the same area (2.4GHz) and can interfere with each other, in some cases..
Luckily you can often choose the channel used for your Wi-Fi in your internet router, and set the Zigbee channel in HE (Hub settings -> Zigbee details).
If you search this great community site, you will get a lot of good info - see these threads as an example for more info:
As mentioned, some (like me) can obtain stability, but itās a balancing act. That said I donāt personally feel the Xiaomi stuff is junk. Itās inexpensive and shows it in some ways, but they have clearly created a proprietary version of Zigbee that doesnāt cooperate with the standard OTB Zigbee.
When I join their devices to their gateway ( which is where most of my Xiaomi devices live) they are rock solid. They pair easily and never drop. So people saying they are junk are in my opinion, misinformed about what they bought. Youāll find Zigbee written on their stuff, but Iāve never seen any claim of full Zigbee 1.2 compliance on any of their marketing or technical information.
@SmartHomePrimer - yes, you could say that I was misinformed about what I bought. But then I feel like any person starting out with this stuff is most likely misinformed. I donāt know how many hours of research I would have to do to understand all the technical nuances about what works with what, and what doesnāt, what works sometimes, what works when other things are present or not present, etc. before actually buying anything and trying out. Letās face it, and that much I have already realised, none of this is really plug & play. All of this is really a DYI project.
So back to my issue - would getting a dedicated Xiaomi bridge and connecting all of these sensors to it, and then connecting the bridge to HE do the trick? I.e. does the Xiaomi bridge do the Zigbee ātranslationā between the Xiaomi Zigbee protocol to the standard 1.2?
[I suspect the answer to this question will be - it works for some, for others it doesnāt work...]
The first two of these work and can be stable, the third only exists in a dream. Depending on what you have, the cost for this can be as low as 20 quid for just a Mijia Multimode Gateway, although I would opt for the Aqara M1S instead.
I bridge Xiaomi Aqara and Mijia devices with HomeKit automations, so yes you need two additional hubs, and a RPi or computer to run Homebridge or Hoobs, but in my case I watch television on that Apple TV 4, and weāre an iOS family, and Mac users, so that works out for us. I also already had Homebridge and other node.js applications running on a separate computer long before I got my first HE hub.
If you are an Android wielding family, then it can be done with HA and a few simple automations to sync back to HE, where you can use them as if they were directly connected (almost). Just as my method with HomeKit automations excludes variables data from devices like temperature sensors and vibration sensors, so too does this. But I have found those two Aqara device types to actually stay paired to Hubitat quite well when the repeaters are in the right place. One Xiaomi hub and a RPi 4 or a spare computer, and patience will do the trick.
If I have my second wish granted (My first arrives today at 11:59am EST), then it will be to have a HomeKit Controller type of app in HE so anything that can connect via HomeKit (which is just about everything) can bring its devices into HE.
No there isn't a lot of cohesion yet. It's moving in that direction, but slowly.
That's great -- thanks for taking your time to clearly outline all these options, really helpful.
OK, so to summarise, I would need the following:
OPTION 1 (Apple):
Mijia / Aqara gateway.
HomeKit 'gateway' -- so this could be one of the qualifying Apple devices, e.g. HomePod, Apple TV, etc. (always staying at home, so no iPad that is taken out of house on a regular basis).
RPi / computer to run Homebridge / Hoobs that is always on probably a dedicated machine rather than a computer that we rely on for normal usage.
OPTION 2 (Androind):
Mijia / Aqara gateway.
RPi / computer to run HomeAssistant that is always on probably a dedicated machine rather than a computer that we rely on for normal usage.
We are a mixed device household - I am Apple (MacBook, iPhone, & Watch), my wife refuses to use Apple, so she is Win / Android. For entertainment we use Amazon devices.
Going through any of the two options above feels daunting, but I suspect this would open up a lot of other opportunities beyond Xiaomi integration, e.g. other non-standard protocols could connect to HE..?
But I think it may be easier for me to stick to devices that definitely (is there even such thing?) connect to Hubitat and avoid all these 'middle men' steps. My gut feeling is also that the more 'middle men' integrations there are, the more things that can break down and the more things that need to be able to connect & talk to each other, and remain stable.
My main reason for getting HE was because I though it achieved two main objectives:
1/ Speaks different languages so can talk between the different technologies directly, which means I don't need all these brand-specific bridges.
2/ Is able to bring all of these ecosystems into one, so opens up tons of opportunities for cross-device automations.
What I am now realising is that while it does (2), it's not really what I expected for (1) and I still need a lot of 'translators'. Basically, I thought HE would speak half of UN's languages, but really it only speaks English and maybe some Spanish. On a good day it can understand some 'Spanglish'.
@AnubizDK - this might be a silly question, but I guess I am not only looking for no overlap between my household's WiFi and Zigbee channels, but also my neighbours' WiFi channels that reach my house? I guess radio interference doesn't discriminate against neighbours' WiFi, does it?
Below is a summary of WiFi networks that can be picked up from my house (we live in a town house, so share two house walls with neighbours). It looks like the channel with the lowest load is 6, so I could possibly move my WiFi to 1 and keep my Zigbee at 20 so that it interferes with as few WiFi networks as possible. Or would you suggest a different combination?
I have 2 of these. They're in the same place they have been for about 2 years and solved my problems, originally, along with.....
LOL, I use the exact same. Switched along with adding the repeaters to get them stable.
Orignally veeceeoh's and then Markus' when I was having issues.
I have 3 Xiaomi cubes that the channel change and repeaters solved dropping off issues for as well. They, however, are still up and running without issue.
Correct. You will need to install the Maker API plugin on HE, but that is simple and Hoobs is a very easy route to take for the plugin needed, plus the instructions for HE are all right there. Android users are no more or less included in this option, but it requires either an iOS or MacOS device with the latest OS to create the linking HomeKit Automations.
You also need a method to bring HE virtual devices to HA so they can sync with Xiaomi devices or other devices in HA. This can be accomplished with a community instegration, or with Homebridge/Hoobs running on the same computer with HA. I would tend to choose Homebridge/Hoobs over the community integration for maximum flexibility. There are a lot of Homebridge/Hoobs plugins, which broadens one's capability to sync devices with HE that are currently unsupported, instead of limiting them. Again, you also need to install the Maker API plugin on HE, but that is simple and Hoobs is a very easy route to take for the plugin needed, and the instructions for HE are all right there. You don't exclude iOS users with this method, but you create the linking automations on HA, not HomeKit. iOS users can still access HomeKit, but remote HomeKit access is unavailable.
Correct. HA is fully open source, so there are no business concerns. On the flip side, there are no customers of the platform per se, so meeting requests isn't a top level priority. That isn't to slight any of the efforts made on that platform, but it's the reality. Hubitat is designed to be different, yet familiar. HA is very different in their approach, and their history drives their direction. Hubitat didn't have a history to contend with, and it has allowed them to create a friendlier user experience that empowers owners like me to create complicated automations, without knowing coding syntax. HA requires other systems to accomplish that with any ease, and they're not without their own efforts to add, learn and maintain.
Very true. I like to build and experiment. I don't like overly complicated interfaces and I'm not a developer. That doesn't mean I haven't introduced a level of complexity in my setup, but it's not as great as some.
This is true in that HE does actually talk to a lot of devices. The company is very nearly 3 years old and they have come a great distance in that time. On the other hand HA first came into existence 7 years ago. Their transformation of the UI has not evolved as quickly or as much at HE has in less than half the time. 7 years time allows for a lot of integrations to be added. There's no denying the number available, but the interface and ease of use I personally find lacking, and the need to flip to Yaml to create automations directly on their platform arrives pretty quickly, and I find it unpleasant to use.
If any of the devices on HA doesn't work 100%, you can report, and you can ask, but you are forced to keep expectations low. You also have a lot of cloud integrations on HA, instead of local. On the surface it looks like their local integrations are vast, but then when you start investigating them, you find a large number of them are merely linking to cloud APIs. The focus with Hubitat Elevation is always local first.
Hubitat Elevation speaks some other languages, but English is their mother tongue so to speak. I think that is OK. I speak English, as do you and the majority of the world population. To me personally, HomeKit has become what I didn't actually expect from the start, but if you can "speak to HomeKit" as the HomeKit Controller integration does on HA, then you essentially have a kind of universal translator that is 100% local. That is a golden egg I really want for Hubitat.
Today I have a new President. I'm hopeful for the future of my great nation, and looking forward to mending the relationships with our nation friends. I want this for Home Automation too. The more unity we have, the more we can accomplish together to achieve our goals, regardless of whether they are the same. So long as they are honest goals that raise people up instead of pushing them down, they are worth pursuing.
