Is Smarthome KNX automation possible based on the floor plan?

: you haven't understood the basics of KNX yet ....

I'll do this now exemplarily for 1 lamp on/off and one roller shutter up/down

The active side, meaning 230V

Roller Shutter
A 5-core cable runs from the distribution board to the roller shutter. This cable is connected on the roller shutter side to the motor and on the distribution side to the actuator: 1 core for up, 1 core for down, and of course PE and neutral conductor

that's it
Light
Similar to the roller shutter. From the distribution board a 5-core cable to the lamp. 1 core is connected to the actuator and one to the lamp, PE and neutral as usual (the two free cores are just for the possibility to retrofit something later)

that's it. that's all there is. And with that you can also at any time, for example, switch a line to 24V since the line from the consumer goes into the distribution board.

If you now want to make a dimmable lamp out of the switched lamp, you simply connect the core to a dimmer in the distribution board. For the sockets it is analogous to this wiring. This way you can first 'cheaply' switch everything on and off and then later switch to dimming (dimmer actuators are more expensive per channel).

Basically, this corresponds to the classic installation but with different devices. Instead of the switch, you have the actuators.

Now comes the 'magic' KNX part. These actuators (switching relays) are connected to the bus and listen on the KNX bus and react to the commands parameterized for them.

The sensor technology (pushbuttons or the classic switch) is completely decoupled from the actuator technology. All sensors, whether PM, motion detectors, pushbuttons, home server, etc., are connected to the bus. If you now press a sensor/pushbutton or a sensor has detected movement or a thermoelement has reached a switching threshold, it sends the parameterized command onto the bus. For example, you have a switch in the living room programmed to send the command 'close all roller shutters.' Then all actuators that know this command react and execute this action.

You can now also give the actuator more than one command code. So you say to the actuator: you listen to 'close all roller shutters', 'close roller shutter kitchen 1', 'close all roller shutters kitchen'.

If you now program a button to 'close all roller shutters' and press it, it sends this command; you program the next button 'close all roller shutters kitchen', then it sends this to the bus and the corresponding actuators programmed for these commands carry it out.

That is the advantage of the system. You can now attach any number of sensors in any number of places and program them accordingly. No matter where in the house you press, the actuators respond. If you now have 4 equally programmed switches in a room, it does not matter which one you press, the same action will always occur.

That is the huge advantage of KNX: action and reaction are programmable.

For the bus there are now tons of commands: on/off, dimming, up/down, time, positions, etc., etc. And all devices, regardless of the manufacturer, adhere to these commands. That means if a B&J sends the up command, an ABB or MDT actuator also understands it.

Now to the server: it basically does the same. Software runs on it that can send and receive commands on the bus. And there you can create links (so-called logics). Many current actuators and sensors already have simple logics built-in so you can do a lot with them already.

If you do lighting and roller shutters in a star topology and then also run one or two lines for sockets per room and distribute them in the room, you can also later install the corresponding 'switch extension' in the distribution board. If you don't want to switch or monitor, the respective lines are simply bridged.

I hope this explanation could help you a little bit. Because if you haven't understood the basics, it doesn't make any sense to deal with details or specifics at all.

@Tom

A very well and clearly written explanation for laypeople. Thank you very much for that!


Construction expert on the go

: you have not yet understood the basics of KNX ....

Essentially, what you describe is my understanding of KNX. At least since yesterday...

Light
Analogous to the roller shutter. A 5-core cable from the distribution box to the lamp. One core is connected to the actuator and one to the lamp, PE and neutral as usual (the two free cores only serve as the possibility to retrofit something later)

From the fuse, one core goes to the actuator, and then the L line of the 5x1.5 NYM cable is connected to the actuator? The N line goes directly from the 5x1.5 NYM cable to the fuse? That is my understanding now?
So, L line: fuse -> actuator -> lamp
N line: fuse -> lamp
Plus: 2 free wires and PE

That's it. Nothing more than that. And with this you can at any time, for example, switch a line to 24V because the line from the consumer goes into the distribution box.

Should you then lay a cable with a larger cross-section for 24V? Because of voltage drop?

If you now want to make a switched lamp dimmable, you simply connect the core to a dimmer in the distribution box. The sockets are wired analogously to this. This way you can first make everything 'cheaply' switchable on and off and later switch to dimming (dimming actuators are more expensive per channel).

That is clear to me. At first, I could make all lamps switchable and do nothing with the sockets. And later make the lamps dimmable and use the switching actuator for the sockets. But if desired, I could also make everything that should be dimmable dimmable right away.

That is the huge advantage of KNX. Action and reaction are parametrizable.

That is why I said I would first parametrize roller shutter 1 up, roller shutter 1 down, lamp 1 on, lamp 1 off, etc., to the buttons. And when we then know how we want to do things, I would change that again and perhaps summarize something here, parametrize a scene there, etc.

For the bus, there are now plenty of commands, on/off, dimming, up/down, time, positions, etc., etc. And all devices, regardless of the manufacturer, comply with these commands. That means if a B&J sends the up command, an ABB or MDT actuator also understands it.

Now about the server, it basically does the same. It runs software that can send commands to the bus and receive commands. And you can then create links (so-called logics). Many current actuators and sensors already have simple logics inside so you can already do a lot with that.

Yep. And if I want to set the wake-up time in the evening via web interface (server in the intranet), that is presumably only possible with a server. On the other hand, that is nothing essential, so the basic functions (roller shutters, light) are directly parametrized and, for example, the wake-up time (light dims up slowly, roller shutter goes up, etc.) is controlled via the server.

question 1: the N conductor goes to the N potential of the RCD. only the phase goes to the fuse

question 2: it depends on the distance to the consumer and on the power, it wouldn't do any harm, it just costs more and if you don't need it, the money is wasted

question 3: yes

question 4: exactly, although I would already think about grouping. for example, in the bedroom it certainly doesn't make sense to control only one blind

question 5: no, there is the MDT glass control center for little money (something around 150€ I think). you could install it at 1.55 above the switch (if you like the design). you can easily set things like that on the display. you see, there is nothing that doesn't exist, also for little money there you can then trigger all sorts of things time- or event-controlled. there are also regular enclosures for the control cabinet that are also in this price range.

of course, you can also use something like edomi for little money. the software itself is freeware, the required PC costs about 100€. it is a really ingenious and super easy to use logic/visualization software based on a linux system.

I use it too and it is really great!!

additionally, I have fhem running on a raspi to bring my photovoltaic system to the knx bus. and another raspi with knxd as a router. but you need to have a 'tinkering passion' for that.

for the 'normal user' a simple IP interface from MDT for about 150€ is enough. you need that anyway for parameterizing.

question 1: the N conductor goes to the N potential of the RCD. only the phase goes to the fuse

And exactly that should be properly prepared by an electrician with terminal blocks and an N isolating strip, for example. Then "the builder could theoretically" wire the actuators himself. This way, later changes are also not an obstacle. At the moment, for example, most of the sockets intended for switching are bridged. I only need to remove the bridge and connect the actuator channel in between, then I can actually switch the socket.

: that's exactly how it's done. but since I assumed that the OP only needs it for 'understanding', I didn't go into detail. you actually need at least 2 RCDs today.

I have 6 installed. 1 RCD for lights, 1 RCD for kitchen refrigeration/freezer, 1 RCD for outdoor area, 1 RCD for sockets/roller shutter, 1 RCD for kitchen, 1 RCD for internal distribution cabinet