ᐅ Smart Home Planning – Is This Basic Approach Suitable?

Since my last thread, I have spent some more time studying the subject, and as the house planning is progressing faster than expected, I have already started making more concrete considerations regarding the "smart home" topic. I would like to share my ideas and plans, but also ask some fundamental questions to clarify my understanding.

The system will be implemented using KNX. For functions that I don’t want or “need” right now, I at least want to plan empty conduits (ducts) for future wiring. The system’s long-term expansion should then take place through these empty conduits or, if they are not available, via a wireless solution.

First, a few questions and statements about the basics, just so I don’t plan based on incorrect assumptions. Please correct me where I am wrong:

• A KNX system can be fully programmed on the actuators, fully through logic, or via a mix of both. However, the components always need to be parameterized, which can only be done with ETS.
• The logic engine runs on the home server, which communicates with the bus via an interface. The logic is based on the ETS project.
• Each controllable load or group is connected to its own power line running to the distribution board. This means, in the extreme case, one line from every socket and light fixture to the distribution board.
• Can bus cables be branched, or do I need a separate cable from each sensor to the distribution board? Is it enough to run one cable per room and lay empty conduits for possible future sensors, or do I have to pull cables right away?
• Does it make sense to control underfloor heating and a central mechanical ventilation with heat recovery system via KNX?

I want to carry out the setup and expansion of the system in multiple phases. Although I am technically and manually skilled, my knowledge of programming, electrical work, and electronics is not sufficient (yet) to handle such a project on my own. Therefore, the electrician/system integrator will perform the electrical and KNX installation and setup. Initially, only the basic functions will be implemented—for example, lights on/off or shutters up/down using specific pushbuttons—so the house can first be operated “manually.” The second phase will start after completion and include creating a visualization with smartphone control, programming scenes and automatic sequences using a logic engine (I am thinking of something like Edomi), integrating a central control panel (Android panel or touch display with Raspberry Pi), and a home server (NUC or thin client), as well as connecting other components such as an IP door intercom or remote controls. The final phase will be further system expansion as needed, for example with a weather station, outdoor and indoor cameras, humidity sensors and ventilation control, or integrating wireless components (e.g., window contacts).

The house will be a bungalow without a basement but with an attic, about 140sqm (1507 sq ft) of living space and an attached garage with a technical and storage room. The distribution board will be located in the garage to avoid clicking actuators inside the living areas. I have thought about what I want for the basic installation in the individual rooms:

Living/Dining/Kitchen:

• Multi-functional pushbuttons for lights and blinds; pushbutton in kitchen for under-cabinet lighting
• Presence detectors/light sensors in all three areas
• Several lighting groups in the living area, one group in dining, ceiling spots and cabinet underlighting in kitchen each as one lighting group—all dimmable
• “Multimedia sockets” switchable, possibly additional dimmable sockets in the living room
• Outdoor light sensor under roof overhang on south side

Entrance area/Guest WC/Hallway:

• Pushbuttons for lights; in WC multi-functional pushbutton for blinds and lights
• Presence detector/light sensor

Office and Children’s Rooms:

• Multi-functional pushbuttons for lights and blinds, lights dimmable
• “PC/multimedia sockets” switchable, rest possibly dimmable
• Presence detector/light sensor

Bedroom and Walk-in Closet:

• Bedroom multi-functional pushbuttons for lights and blinds, lights dimmable
• Sockets switchable and dimmable
• Presence detector/light sensor in both rooms

Bathroom:

• Multi-functional pushbuttons for lights and blinds
• Separate lighting groups above bathtub, shower, and sink—all dimmable
• Presence detector/light sensor
• Socket for mirror lighting switchable and dimmable

Utility Room:

• Pushbuttons for lights and blinds
• Presence detector/light sensor
• Water leak detector and shut-off valve
• Sockets switchable

Storage/Technical Room:

• Motion detector/light sensor

Garage:

• Multi-functional pushbuttons for lights and garage door
• Motion detector/light sensor

Additionally, preparations for a weather station, heating and ventilation control, wall panel, etc.

What would be a rough estimate of the cost range for this system, assuming average-quality components?

Quite a long post—I am sure something might be missing. But are there any major inconsistencies or nonsense here? Please correct if necessary.

HansDampf1311 schrieb:

• A KNX system can be fully programmed on the actuators, entirely via a logic controller, or through a combination of both. However, the components always need to be parameterized, which can only be done using ETS.

Yes. But be careful—there is a common misunderstanding: the logic runs either on the actuators themselves or on an external logic controller or server. Nowadays, every actuator comes with some built-in logic functions, such as gates, scenes, presets, etc. KNX is not programmed through logic, apart from being parameterized. In any case, ETS is always required.

HansDampf1311 schrieb:

• The logic engine runs on the home server, which communicates with the bus via an interface. The basis for the logic is the ETS project.

Yes. Although "home server" is a product name from Gira (not sure if that’s what you mean or if you generally mean a visualization/logic server).

HansDampf1311 schrieb:

• Every switchable load or group is connected to its own power circuit running back to the distribution board. In extreme cases, this means a separate cable for every outlet and light fixture going directly to the distribution board.

That’s correct. However, for multiple outlets, you can use cables with 5 or more cores. Alternatively, flush-mounted actuators can be used if cable runs become too long or complicated. Decentralized LED drivers, flush-mounted shutter actuators, or flush-mounted binary inputs are often preferred in these cases.

HansDampf1311 schrieb:

• Can bus cables be branched, or do I need a dedicated cable from each sensor to the distribution board? Is it sufficient to supply each room with one cable and only install empty conduits for possible future sensors, or do the cables have to be pulled right away?

You can absolutely do everything with the KNX bus cable (the green cable), except create a closed loop. KNX lines are usually installed as an open loop to provide a kind of "backup."

HansDampf1311 schrieb:

• Does controlling underfloor heating and a central mechanical ventilation system via KNX make sense?

These two systems don’t have much to do with each other initially. I assume you mean the valve actuators? Regarding underfloor heating, there are two schools of thought: 1. no direct regulation is necessary because a proper hydraulic balancing is sufficient, or 2. individual room control is desired regardless of whether it makes sense. Both approaches are valid. This has nothing to do with mechanical ventilation control, though. I would recommend letting the ventilation system regulate itself, except maybe for special modes like party or cooking mode, provided it can be controlled room by room.

HansDampf1311 schrieb:

Then also prepare for a weather station, heating and ventilation control, the wall panel, etc.

In my experience, "preparing" KNX never really works over time. You either do it right away or you don’t do it at all. Especially a weather station I would always implement from the start, because its functions are, in my opinion, essential. You usually don’t directly intervene in heating control; you mostly just monitor what the system is doing, same with ventilation. You can prepare the wall panel, but it mainly requires LAN and/or power supply rather than a bus cable (though having bus wiring never hurts).

In terms of cost, it’s hard to say much because every component is available in many variations. For example, pushbuttons range from about 50 Euro up to four-digit prices. It’s not quite that dramatic but similar for actuators and sensors. On top of that, you have parameterization costs, which you either pay for or do yourself. Expect five-digit figures overall. I don’t know your current market knowledge or KNX experience, but depending on that, consulting a system integrator or someone who has done it before is recommended. Otherwise, expensive mistakes can happen that cannot easily be corrected.

HansDampf1311 schrieb:

Switch for lights, multitaster (multi-button switch) for blinds and lights in the bathroom

In my experience, these are rarely used. Why bother when the lights turn on automatically and the blinds go up and down on their own? You can save money here. The same applies to the utility room.

Or only use them at the beginning until you get used to the automation and break old habits.

HansDampf1311 schrieb:

What would be the approximate cost of this system assuming we use average components?

That largely depends on the choice of components and the level of expertise involved. The more service you need to hire, the more expensive it will become.

It also depends on factors such as how the lighting should be dimmed and how many controllable channels you require.

Plan for a weather station from the start. Preparing for KNX works well in the attic, garden, and garage. Inside the house, you need to carefully decide now where everything should be installed and at least provide empty conduit boxes with bus cables and NYM cable at those locations, which will be connected later but should be installed during construction.

untergasse43 schrieb:

It’s hard to say anything about the price since you can have every component in many different versions. Switches, for example, range from 50 euros up to four-digit amounts. Not quite as bad, but it’s similar for actuators and sensors. On top of that comes the parameterization, which you either have to pay for or do yourself. It will probably end up in the five-digit range. I don’t know how extensive your market overview and KNX knowledge are, but depending on that, it’s advisable to consult a system integrator or someone who has done this before. Otherwise, things often go wrong and can’t be fixed later.

The complete planning, installation, and commissioning including parameterization will most likely be outsourced to the electrician/system integrator. I would basically trust myself to do it if I invested the time, but I simply don’t have it. Over the years I’ve also learned that even if you’re basically capable of doing something yourself, it often makes sense to pay a bit more and have an expert do it. The only reason I want to get relatively familiar with the topic is because I don’t like to have projects planned and contracts signed without a certain level of understanding myself.

Mycraft schrieb:

In my experience, those systems almost never get used. Why bother? The lights turn on automatically and the blinds go up and down by themselves. You could save money here. Same goes for the utility room.

At least only during the initial phase until you get used to the automation and break old habits.

Yes, you’re probably right. My idea was more about redundancy. Sitting on the toilet in the dark and having the presence detector malfunction would be annoying. Obviously, I wouldn’t install an expensive glass switch here.

Mycraft schrieb:

It strongly depends on the choice of components and also on experience. The more service you have to buy in, the more expensive it gets.

It also depends on how the lighting is dimmed, for example, and how many switchable channels you need.

The complete planning, installation, and commissioning including parameterization will most likely be outsourced to the electrician/system integrator. If we take components in the mid-price range, do you think we can manage with 10,000-15,000 (euros) or is it more likely to be over 20,000 (euros)?

Mycraft schrieb:
Plan the weather station right from the start. KNX pre-wiring works well in the attic, garden, garage. Inside the house, you need to carefully consider now where everything should be installed and then at least provide empty conduits with the bus and NYM cables to those locations, which will later be connected but already installed during the build.

Yes, I will definitely have the weather station prepared.
The house will have an accessible mezzanine floor as storage space. Would it be possible for retrofitting to run all cables there and then pull them down into the installation level from above to the intended endpoints?

HansDampf1311 schrieb:

And over the years, I’ve learned that just because you can do something yourself in principle, it’s often better to pay a bit more and have an expert do it.

That should be framed somewhere.

HansDampf1311 schrieb:

My thought was more about redundancy.

You can still have that without any switches/buttons in the rooms mentioned.

HansDampf1311 schrieb:

It’s annoying if the occupancy sensor malfunctions while you’re sitting on the toilet in the dark.

Of course, I could say that this almost never happens, but yes, there is a small probability of about 1:100,000 (or so). You’re still in the early stages and want to give up as little control as possible (like every beginner in this field). But that changes, and later you’ll wonder “what was the button originally planned for?” and “why did I spend the money?”

But for all the skeptics: do you still manually turn on your car’s ignition? Because that could malfunction too, right?

HansDampf1311 schrieb:

Logically, no expensive glass touch switch would fit here.

Definitely not. But every junction box, each insert, and every additional device costs money that could be spent elsewhere. Let’s be honest: after building the house, there’s usually nothing left over anyway.

HansDampf1311 schrieb:

If we take mid-range components, do you think we can manage it with 10-15k? Or will it rather be 20,000+?

How much does a mid-range car cost? Do you want three- or five-doors? I could calculate houses with comparable functions at 10k, 20k, or even 50k. There are just too many unknowns. Just the number of switches is obviously uncertain.

HansDampf1311 schrieb:

The house should have a walkable intermediate ceiling as a storage area.

How should that be imagined? Are you building an entire mezzanine floor as storage?

HansDampf1311 schrieb:

Would it not be possible to route all cables there for retrofitting and then pull them from above into the installation layer up to the target point?

It depends on what you want to retrofit.

Mycraft schrieb:

You can still have these without any switches/buttons in the mentioned rooms.

How exactly? Multiple sensors?

Mycraft schrieb:

Of course, I could say that this scenario almost never occurs, but yes, there is a small probability of about 1 in 100,000 (or so). You are still in the early stages and want to give up as little control as possible (like every beginner in this area). But that changes, and later you will wonder "what was that button even planned for?" and "why did I spend the money."

You’re probably right there. Changing the mindset isn’t easy, although I’m quite flexible in that regard. My partner reacted with a very skeptical and confused look to "Then we only have one central switch for everything in the living/dining/kitchen area" 😀

Mycraft schrieb:

How much does a mid-range car cost? And should it be a 3- or 5-door? I could calculate houses with comparable functions for $10K, $20K, or even $50K. There are just too many unknowns. Even the number of switches is obviously uncertain.

Ok ok, got it :p Then I’ll come back to that when I have more specific information.

Mycraft schrieb:

How do you imagine that? Are you building an entire mezzanine as a storage room?

Interesting idea but no haha As I wrote, it’s going to be a bungalow, by mezzanine I mean a solid ceiling (as opposed to a suspended ceiling or an open roof structure) that will serve as additional storage space.

Mycraft schrieb:

It depends on what you want to retrofit.

Well, switches and wall/ceiling sensors.