It seems that there are a lot of things to consider before even buying the first smart device. How would you start when you would start over?
Are there any good beginner guides that helped you?
Important points for me are
- privacy (everything should be local, no Alexa-Karens in my home)
- use of open source/free software
- a good variety of smart things I can use (I don’t want to be tied Apple-like to only one company)
Is there a golden way to build a smart home with these factors in mind?
Okay!
Like everyone else “locally” managed devices are 1000x better, faster, more reliable, and less intrusive.
So I’d skip the cheap wifi devices and go straight to matter/thread as that new stuff is a better idea. It’s still in its infancy soooo tough call, for now I’m going “zig bee” and “zwave “ for new devices, but I hope to switch to “matter/thread” someday.
I hate the wall warts, but they’re useful in some places, it’s all about the light switches. When you replace them, everything still works the way you expect and you can automate on top. You don’t want things to “not work” the way people would normally expect them to, so smart bulbs and smart or remoted ceiling fans are a no-go for me.
The system is mostly novelty until you get to the third iteration.
First stage is just “Hey I can attach this to my phone and that’s cool”
Second stage is “voice automation and routines”. This stage you start to automate cool things so that it works more efficiently and the house can have some cool automation that makes things a little easier and allows for “hey siri” parlour tricks.
Third stage is where it gets fun but requires some thought and stability on your system. Now you’re into presence location and motion. I’m just at the beginning of this one now. This is where you start to forget that light switches exist. You no longer ask the voice assistant for almost anything because the lux and presence sensor in your kitchen knows that if there’s not enough light and you’re in the kitchen, the lights should just be on. The motion sensor on your “under cabinet” lighting knows that you’re working at the counter and it should just be on.
Stage 3 is the sweet spot because it just exists and makes sense and you find yourself forgetting that light switches exist into something breaks.
I found that I hate battery motion sensors.
That was a learning for me.
Most people don’t quite get there…. They allow the voice assistants to do some little work for them and then it just gets annoying because the day to day life hasn’t changed enough to warrant the effort that was put into it.
Being able to just forget that light switches are a thing is marvellous and you don’t really feel it until you’re at a friends house and mildly annoyed that the lights need to be touched.
…. And then you realize how little you’ve had to think about stupidities in your day to day because from a functional perspective, they no longer exist.
It’s about putting in the extra effort in the final steps and that’s hard to desire once the parkour tricks work.
Also…… the wall panels are SUPER cool, but mostly unnecessary once you hit iteration 3.
There’s nothing you need to touch!
So they’re great for “displays” or “cam views” at that point.
I’ve got a bulb based system, but I have switches that will operate without Internet. I don’t want to rely on my phone for everything if my Internet is out.
The bulb based system is painful when you hit a normal light switch and it doesn’t do what you expect it to do because the bulb is “intelligent”
Depends on scénario really, but yeah I’m getting away from “internet dependant”
I rent, so I can’t change switches. Luckily some of the switches I have cover traditional switches.
Fair enough!
I’ve also learned that I can’t stand battery operated motion sensors!
Have you found any plug-in motion sensors that you use instead? It looks like they are generally much more expensive than the battery powered ones, which doesn’t make any sense to me.
I have! The Tuya zigbee presence sensors.
They are WAY better because they detect presence rather than motion, so they stay on even if you’re not moving.
The m100 Check if you’re getting the 5ghz or 22ghz
Not sure which I recommend, I’m waiting on the 22’s now for a room that’s bigger than 6M
The 5ghz’s can’t sense more than 6 metres
These are using mmWave radar, so look for those type of devices if you want sensors that can detect whether you’re in the room while stationary. Most of the cheap sensors use PIR which only detects motion.
That’s correct! I actually work in the mmW radar field, it’s super fun!
I started with a bunch of cloud connected expensive wifi devices that would have all sorts of issues on the consumer crap wifi router I had at the time.
The amount of money I spent on those devices would have been much, much better spent on something non IP based like zigbee or zwave devices with a hardwired hub.
These days I finally have a more pro grade network setup that handles all the Wi-Fi IOT devices I have just fine, but it could have been so, so much simpler.
I always had the high end network, and I still hate them.
Wemo devices randomly stop working, tplink devices don’t give you access to all of the controls/data, myQ just decided to shit the bed, Tuya intermittently shits the bed, all the management apps and their separate credentials are a fucking pain, the integrations are painful and numerous and my zigbee/zwave just keep trucking along.
I started with a bunch of Tuya devices that are almost all flashed over to tasmota etc. by now.
What really got me to actually start with home assistant though, is the exterior blinds controllers I bought. They kept losing wifi after a few days until I rebooted them, but they were hardwired, so every day I would have to turn off a circuit in my house with everything connected to it.
I flashed them with openbeken OTA and now they almost always work and, even if they didnt, they have already rebooted and are working again.
I probably would have never started.
Everyone insisted that it was “easy”. I’ve come to realize this community is heavily developer-based and completely unaware of what they average person would consider easy.
My system breaks all the time and I have to spend hours trying to just get it back up and running again.
I was never able to get the app to work.
If I could have all my money and time back I would have just stuck with “dumb” electronics.
Everyone insisted that it was “easy”. I’ve come to realize this community is heavily developer-based and completely unaware of what they average person would consider easy.
This is applicable to every piece of advice I’ve gotten from lemmy and from reddit before it.
No, I can’t work from home.
No, I don’t have a raspberry pi and I don’t know how one works.
I don’t know what podcast you’re talking about.
I have to fight to prevent my eyes glazing over when you say apk.
If you link me to something and it’s just a list of the latest bugfixes without even a summary of what the fucking software IS, I won’t understand.
Poor people are not scary. Some of them wear glasses too.
I don’t even HAVE a PC, just a laptop that I carry with me.
You lost me at “terminal”.
I don’t know what any of those acronyms and abbreviations you’re using mean.
Or the “step-by-step” guides that are missing a step. Or assume a bunch of knowledge. If you don’t tell me I need to download a special compiler to install this thing, it’s not step-by-step.
I found one the other day that failed to mention that I had to put some code in config to make it work. But they had put a screenshot of the code they used, just hadn’t referenced it in the steps 🙄.
I’ve also seen one of the main devs respond to a user on GitHub saying that the bug they were seeing was not a bug because it was caused by the third-party system and “that’s just how it is”. Completely ignored the fact that the user could not achieve the intended behaviour from the integration. Was this information anywhere in the notes? Of course not.
Smaller steps. One at a time. Test.
I started off with HA, integrating all my devices i had at once, totally enthusiastic as I was. There were times where I installed several automations and integrations within just a few minutes, not giving them time to work for them selves, especially not in a „productive“ environment with wife, kids, neighbours, etc. all involved in the system.
Lots of automations broke or did not work as expected and I saw my self confronted fixing so many building sites at once.
I am through the roughest part. But should I start from scratch any time I will most surely start I’d small, test, iterate and move on.
I recently “started over” but it was more of a rebuild in place. I did room by room. Areas were not a thing when i started so i added everything to a new area. Deleted any automations that targeted that area. Renamed all devices/entities in the area to a standard naming scheme. Then creating a dashboard for that area. Rebuilt the automations one at a time. Letting them bake before starting a new automation. And finishing all automations for that room before starting a new area.
If you don’t mind me asking, what is your naming scheme?
I got into “home automation” during the initial rush. Everything was dependent on its own hub or cloud service. After using some of them, especially the voice controlled ones, I got interested in what all they can see/collect. I quickly dropped everything and have been in a dumb house for a few years.
What I’ve realized I need is cameras to watch my pets and a dumb keypad door lock to let people in when I’m not home.
What I want again is a thermostat I can control from my phone. I haven’t found one that doesn’t need cloud services.
If I wanted to get everything “smart” again I would isolate everything to its own VLAN with no internet access for the whole VLAN. I would have an electrician fix my no neutral drop situation and get wifi switches not bulbs. I would add temperature sensors in each room. Custom build a camera system server. Add network access to my Honeywell security system.
Switches not bulbs
Yep, having just moved and trying to setup the new/same network in the new house. Most of the new house is already fully led or had many bulbs on one circuit. Thus I need switches.
Other thing, save the instructions. I bought a new hub for the new house. Getting stuff to pair with the new hub was a disaster and I’m still using the old hub. Unpairing instructions would have helped a lot.
I am in the stage of completely stripping and rennovating my new house.
Here in belgium, pushbutton switches are becoming standard now instead of circuit-breaking switches. Every switch ia routed to the control board and every light circuit also with nothing in-between. Super handy for wiring and tracing wires. The only thing is, with electrical devices increasing in price so much, it is 20-40€ for a single
Since I am rewiring every single bit of wiring in my house, I am putting in a KNX system. A 16 channel smart switch output plus 32 channel binary input is almost the same price now as the “dumb” switches. They are potentialess, so nobody ever has to deal with shocking themselves on light switches like I did as a kid by being stupid.
Definitely recommend KNX and actually putting wires everywhere instead of wireless
Tuya stuff is cheap, but it’s not worth the hassle.
Their zigbee stuff is good
Amen. Same boat here. I have too much of it now and I want to start migrating to local control. I’m starting on Shelly devices myself.
My WiFi stuff is Shelly and Tasmota, and it’s working great so far. My zigbee stuff is excellent, but I struggle to get good coverage in my concrete bunker of a house.
I think you’re on the right track considering those factors and others (integrations with HA is another big one for me). IMO it’s always a balance though - a perfect example for me personally is I still use Nest Protect devices because they just freaking work, and work well. Same with a smart display I have and some minimal smart speaker usage. So sure, there’s a little bit of “lock in” there, but as long as your risk tolerance is ok with that (mine is) then you’re golden.
The factors you mentioned, among others like HA integrations (preferably official), company reputation (I’m personally ok with some minimal lock in if the cost/benefit calculation works out), etc. are my first considerations when looking at new products. Wherever possible local first control is a great plus and can be an absolutely huge factor to be sure.
everything using pyscript. again.
