ᐅ Consultation for Smart Home New Construction Wireless Systems

teh_M schrieb:

Maybe run a bus cable to those two points as a "preparation"?

It doesn’t really make sense if you don’t connect the rest of the bus as well.

teh_M schrieb:

The Raspberry Pi, well. I’ve been burned by that too. Back then, I used the Pi with FHEM and EQ Max thermostats, plus the 433 MHz wireless sockets. That was (sorry) garbage. Then the SD card failed after about six months, and everything was lost. No, never again.

So I try to avoid it as much as possible. OpenHAB is already a “DIY solution” enough. I don’t want to cobble together five more interfaces with the Pi.

In case of emergency, it’s wired conventionally and I remove the wireless components. If the Synology or anything else crashes, I can still operate about 95% with switches. That was very important to me (or my wife ).

When I read stuff like this, I always wonder why DIY solutions are even considered in the first place?

Mycraft schrieb:

When I read things like this, I always wonder why DIY solutions are even considered?

What you understood and interpreted is not what I meant at all.
If something in the entire system doesn’t work, I can still use switches. Just as a fallback. You can’t do that if you don’t have a switch for the blinds, as was considered a few posts earlier.
Also, if a component fails, I can simply rewire that section and it keeps running—for the time being, at least. As a fallback.

I see the “DIY solution” more in using the Raspberry Pi to control the mechanical ventilation system or similar, and partly openHAB, yes. Once it’s running, it’s fine and generally stable, but during initial setup you definitely have to engage with it and solve some problems.
openHAB is certainly not a top-of-the-line solution like KNX; it’s somewhere between a DIY project and KNX. Hence the quotation marks...

Honestly, I haven’t studied KNX in enough detail to fully assess this part:

Mycraft schrieb:

It doesn’t really make sense if you don’t also connect the rest of the system to the bus.

With the (pre-installed) cable and KNX gateway, you could theoretically integrate it into a control unit of your choice if you need to.

I have understood it exactly as you mean. DIY solutions rarely have a fallback option. So why bother with them at all, spend money, and constantly live with the thought that the system will eventually (most likely completely) fail? Your example with the Pi clearly illustrates this, and you are by no means the only one with this experience.

teh_M schrieb:

With the (pre-installed) cable and KNX gateway, you could theoretically still integrate it into a central system of your choice if you need to.

KNX only becomes cost-effective if it is properly set up from the start. Building isolated segments here and there won’t help, as you end up bearing the full system costs anyway.

By the way, KNX and OpenHAB are not comparable, as they operate on different layers of the OSI model.

Mycraft schrieb:

KNX and Openhab cannot really be compared because they operate on different layers of the OSI model.

Yes, you’re right. OpenHab with Z-Wave or something similar would have been a more accurate example.

Mycraft schrieb:

DIY solutions rarely have a fallback level.

I agree with you there, at least not to the extent that you could really call it a fallback level. For me, it’s essential that the operation of blinds and lighting is always possible—that’s my fallback level. This is fulfilled with Z-Wave actuators.
If the logic controller, Z-Wave stick, or NAS fails, well, that’s inconvenient. But the lights and blinds still work.
For me, in a single-family home, it’s acceptable that then the motion sensor no longer switches the lights on.
If there’s no switch available, it’s annoying if kids have to sleep in the light.

“For a small budget” it’s sufficient as it is. Of course, I would like KNX, but I can’t—or perhaps don’t want to—afford it.

But even with KNX there are probably single points of failure… power supply for the bus breaks, the MDT glass switch fails, the logic server crashes, etc. But as I said, I don’t spend enough time dealing with that.
And this isn’t supposed to turn into a KNX vs. … thread again.

teh_M schrieb:

For me, it’s also fine in a single-family home if the motion sensor no longer turns on the light.
If there’s no switch nearby, it’s annoying when the kids have to sleep with the light on.

Regardless of what system you use, there must be a clear concept behind it. There are established experience and proven solutions for this. Home automation in single-family homes is by no means such a new trend as it is sometimes suggested by commercials.

teh_M schrieb:

For me, it’s also fine in a single-family home if the motion sensor no longer turns on the light.
If there’s no switch nearby, it’s annoying when the kids have to sleep with the light on.

If someone has to sleep with the light on, it simply means that the task set was only partially fulfilled.

This is also the root cause of many prejudices against home automation in single-family houses. “What do I do if the light no longer switches on automatically, or if the fingerprint sensor stops working?” In well-designed installations, none of this is a problem; there are always at least one or two alternative options.

DIY solutions usually don’t belong here. They “somehow work,” and in the event of a failure, the house becomes unusable. That certainly does not help build trust, and the WAF (Wife Acceptance Factor) is often very low.

And yes, this is not meant to turn into a KNX vs. thread, just a note to explain this to you. The only single point of failure is the power supply. However, these are designed with so much reserve capacity that failure is almost impossible. As an example: my main power supply is from 1999 and still runs quietly and without issues.

If you want to be extra safe, or a bit paranoid, you can keep a second power supply on the shelf or even install it directly into the system so that it immediately and almost seamlessly takes over in case of failure—just as emergency power generators do in hospitals, for example.

Everything else you mentioned can fail without major consequences (although this is an unlikely scenario); generally, only higher-level functions (logic, etc.) and manual controls at that point would stop working. The basic functions—lighting, roller shutters, heating, etc.—will continue to operate even if the server or sensor switches fail.

I use Openhab on a Raspberry Pi myself, and (so far exclusively) Fibaro Z-Wave actuators... Everything has worked perfectly up to now (lighting, blinds, irrigation system, outdoor awning, a few motion sensors—I’m not a big fan of those), no failures, no crashes, I’ve never experienced a missed signal, incredibly flexible with the future in mind... I can only recommend it. I’m definitely not an IT expert, but I found the whole setup quite straightforward and relatively inexpensive to implement.

Troubleshooting in case of problems is actually simple... Check 1: Is the Pi still running? Check 2: Decentralized actuator — super easy to find since it’s right where it’s needed.

Heating and controlled ventilation won’t be integrated into my smart home... that’s just going to cost efficiency without any real added value in my opinion.