Wiring for lighting in a new home?

I’m at the very beginning of my smart home journey. I’m living in my old place and we’re considering the idea of (re)building and incorporating at least the ability for smart home features. So far, I’ve played with wink (don’t like the cloud requirements), got a Pi3 (currently as a media client), have a ubuntu htpc (media server) with openHAB2 and a Zwave dongle. And a couple of GE Link lights. In other words, barely got my toe in the water.

1/ True or False? My impression is that a wired solution for lighting control would be more reliable/flexible than wireless or that there are some other advantages (what?) that outweigh disadvantages (primarily cost?)

2/ What are the basic essential components involved in a wired lighting system? e.g., a ‘simple’ scenario: I have a server room and I want to control a light on the ceiling in the entrance way: what are the components and where do they go?

I’ve got lots more questions (reading about DALI / DMX / KNX has so far only raised more questions) but maybe I’ll leave it there for now. Hard to know if this is even on topic for the off topic part of the forum.

edit: You could say, if he doesn’t know what a KNX Bus is (and I might not really) then he doesn’t have the ability to implement that as a DIY project, however this is an area of interest for me and I would like to devote my own time to keep things interesting (learning is fun!) and keeping costs down is good where possible.

Both yes and no… The freedom is often higher with wireless components, since you can make more on/off’s and dimmers - maybe only on a 5 wired cable…
I’ve myself has combined the wired/wireless… i’ve seen that wireless LED dimmers is easier to fint than wired - and easier to group…
But the cost can be a bit higher with wireless - denpending on system, Z-Wave is for OK prices…

You need a switch element (output module) for the on/off of the light - or a wired/wireless dimmer…
Then you also need a control panel, could be a phone running OH, or a wired/wireless switch - again depending on system you go with…

I think the first thing should be to define what system you wanna use, there are advantages/disadvantages at every system…

Myself made my house with KNX and Z-Wave… this works perfectly for me… i think DALI is for large areas with a lot of lamps… DMX i don’t know - think it’s more a US brand and pretty expensive

I agree with @Rasmus7700. Neither true nor false. Wireless will be easier to install and use in an existing house but they will be more expensive and possibly have networking problems. Wired can often be cheaper but requires running wires which could be expensive in and of itself. But wired typically does not have the same networking and interference problems one finds with wireless.

Depends on whether you want to go DIY or commercial. I’ve no experience with wired commercial technologies (KNX is not widely available in the US). But I would imagine the commercial stuff would work similar to the DIY.

You would have a controller probably located in your server room that all the wires connect to. The other end of the wire goes to the light switch for the bulb or at the bulb itself. At that location there is a relay wired to the mains power and the control wires. When the control wires go high or low it switches the mains power which turns on or off the light.

I wouldn’t. Everyone has to start somewhere. I’d just caution you not to bite off too much to learn all at once. That has caused a lot of people on this forum untold problems.

I would split the two sides of the system - mains, and controls.

Personally, I prefer the reliability and not-being-locked-in of hardwired…

On the mains side, take everything back to a ‘Node 0’, of if you’re in a big place and cable use would be excessive, have a number of mains nodes - one per floor, one per wing, whatever. In these nodes would live your joints, busbars, relays etc.

On the controls side, prior to jumping into any particular wireless camp, I’d flood the place with Cat5 cable. It’s so cheap, especially the CCA type - I have at least two Cat5’s to ever switch position in my place, sometimes five. Most wireless devices will need wired power supply anyway, it’s only the output which is wireless.

Even though I’m not yet on an active openHAB implementation, by using Cat5 I’m able to run simple pushbutton switches, of my own design, back to my controllers at node 0. The controllers (old Webbricks, old siemens PLCs and various discrete step relays) can handle scene control, timers, temperature monitoring, dimming… all in the one convenient location. And when the day comes that the Webbricks die, I can replace them with Sedona controllers, or xLogic PLCs, or whatever - without taking down lights, or removing sockets, or lifting floorboards, or rewriting config files…

For me, where openHAB comes in, is interfacing. The more functional side of automation, to my mind, is more ‘state’ driven than ‘event’ driven - openhab seems to be an ‘event’-focussed solution, so I use other hardware to hand the ‘stateful’ side of things. I intentionally use simple and proven hardware controllers. I (eventually) intend to use openHAB to interface these dumb controllers to the outside world, openHAB acting as a sort of gatekeeper, it being kept updated much more than a PLC controller (stuxnet anyone?). Interfacing with any BMS control is via Modbus, UDP, Bacnet, whatever. For example, geofencing is best implemented through openHAB, so too weather prediction; and yet daylight sensing and thresholds seems better handled by simple hardware.

There is far more upfront thinking required!

You would not ask these questions when buying a car?!

a) Why do you want to automate?
b) Do you understand the prime automation goals?
c) What is your risk appetite?
d) Are you worried about radiation in your house?
e) What is the cost and standards being used?

For a) Depending on what you want to do in a) determines your vision.
For b) Many realise that controlling lights is the least you will be interested in once you have an automated system.
For c) I am currently building and will not install any light switches; they are controlled by rules or voice – why would I need a physical switch? DIY and vendor systems fall under this category too.
For d) If you are, then wireless is not an option
For e) Once you know where you want to go, you will figure in case you do not have the electronic programming skills, that a consumer/commercial product will be the better approach.

E.g. My risk appetite is bigger than the average; given my background, I can envision and judge, whether the things I want to do are feasible /achievable. I do not have final solutions, but I know how I need to approach specific problems, and more so, that my planned infrastructure can handle it.

At the end, home-automation is about architecture, infrastructure, and vision. As you can see, I did not say a word about the technical detail, which is entirely secondary. If you get these three right, the rest will flow on, otherwise you end up with a mess of a system.

Good luck :slight_smile:

As the others stated, it depends on your requirements.

6 years ago I had the same question. We had not the financial ability to automate everything at once (e.g. change roller shutters to electrical drive) and also not the time to develop a working DIY system till we wanted to move in.

For Germany there are some standards to meet (e.g. VDE) for your mains cabling. Together with our electrician we found the LCN system very useful (www.lcn.eu) for our application:

  • starting with a slightly adapted cablig with Standard Switches/lights (230V mains)
  • room by room can be changed to LCN System, including changing shutter to motor driven ones
  • light switching/dimming, roller shutter, heating can be controlled within LCN System
  • intergration to openHAB is working and you can automate these things by rules
  • even if openHAB is not working (e.g. RasPi dies, configuration messed up,…) the light switches are working as normal
  • if we are selling the house, people without electrical/software skills can deal with it - with the help of the electrician of their choice OR can easy uninstall the automation and go back to standard mains wiring

The last point is also important in another situation: my wife and my four kids can deal with this system, even if I’m not be able to deal with this system (severe illness, divorce, death).

Read, read, read:

  • about the solution of others (like @Max_G with his switch-free Installation)
  • about the experience of others
  • make up your mind which things are important for you and your family and which things not


You can replace LCN with KNX or any other commercial vendor… for people looking at cost, such systems are wildly expensive (IMHO).

While I understand that my stance on light switches seems radical, I want to highlight that change is ongoing and sometimes disruptive and fast. E.g. ‘why would any private person need a computer in their home’. We know where this went. I have more than 10 in a two-person household.
The back-up or return what is standard today; would you have kept the horse carriage, because the car may break down.

Anyway, it was about lighting in a new home :slight_smile:

Ack… a few days away from email and a flood of responses. Where to start?

I’ll mention: I’m in Vancouver, Canada. So reading between the lines, that might direct some of my choices regarding choice of technology.

@Max_G my primary reason for wanting automation is to ‘work smarter, not harder’. There are a lot of tasks that we push down the list of priorities to the point where they don’t get done. In my life that will be watering trees/shrubs/garden when the season calls for it, selectively lighting up the outdoors at night or turning off lighting where it is not needed. Controlling indoor lighting would be for similar reason – why trouble myself to install beautiful lighting if I don’t take the trouble to turn on the appropriate lights and adjust them to appropriate settings for the desired ambiance? A smart home would take care of making sure that the lights were turned on to enjoy the carefully designed lighting in the new home, to make sure that the plants are watered, to make sure that blinds are opened/closed, that outdoor lighting enhances security and makes entrance/exit easier while not detracting from a desire for darkness.

In my case, it would be a hard sell to convince my wife to go without the switch. And really … it is an elegant solution, and very well tested. That said, I would be happy if the switches languished unused.

However, while I can see that my first question was vague enough to incite general advice, that brings me back to my question 2/. And to expand upon it, is a lighting system composed of the following?:

a controller (e.g., KNX bus) in a server room,
specialized wiring (or cat6?) to the entrance
connecting to a “switch” (which serves to report on the state of the lighting as well as switching the mains on/off/or as appropriate for brightness)
with regular mains electrical connection to the light fixtures?

@Rasmus7700 you answered this to some degree, but I guess my problem is that when I browse a hardware website I am confronted with so many options that I end up not knowing what I would choose if I were filling my shopping cart.

I suppose I was hoping for the generic/standard names of the components (if there is a standard name for the various items in the system) so that I could successfully figure out what I would need. And of course, I’d need an understanding of how they connect (if it’s not obvious). Are there any books out there? Or other web-related reading?

@imhofa, I think you may have given me the clue by saying, ‘together with our electrician’. I’ve had a difficult time finding anyone to talk to about the topic locally since the only people that I’ve seen so far doing home automation are the Control4 dealers. This hasn’t appealed to me since I likely won’t be going all-in right at the start, but would like to minimize costs/work down the road by wiring as appropriate.

@impvan: For some reason, this sort of layout is most intuitively comprehensible to me, I guess because it uses the old tried and true technology that I have heard of! At my current level of proficiency, it might be a quicker take-off. I’m going to see if I can find an electrician who’s willing to talk/interested in talking about this sort of thing (smart home wiring).

I am a big picture guy; you seem to look at details (e.g. entrance).
I have not solved all individual problems yet, but, from what I understand with, say OH, all I want to do can be solved with OH -> meaning, I need to get my infrastructure and architecture right first. For the detailed problems there will be ‘multiple ways to skin that cat’ as you can see to many of the solutions to a specific problem, right here on this forum.

The real issue is: which infrastructure can cater for all my communication requirements.
E.g. I settled on Ethernet for cabling; MQTT for messaging; OH for controlling / reporting / storing states and events; all open standards / source based… companies and proprietary solutions come and go (as the past has shown to many times).

The issues with books is the same as with say the ‘electrician’; as you’ve figured, for the latter: they either have no idea, provide shonky work, or rip you off, with the least like outcome of doing a good job… and for the former, the books show solutions that worked for the author or the author is familiar with.
This means, in order to arrive at a solution suitable for your needs, that you understand your requirements, beyond what you can envisage today -> a solid architecture and infrastructure are the foundation for future flexibility and expandability without headaches.

Also, you do not need to tell me / argue about any thing; you do what you like :slight_smile: my input was trying to instill a sense that there are more important things that need to be done before you even implement.
Like Rich said: do your homework upfront (which could take longer, most likely will) than the actual implementation.

As for proven technology: remember New York streets in 1906 (99% horse carts, 1% cars), and 1916 (99% cars, 1% horse carts); while sticking to ‘proven’, maybe still right today, it will certainly not last into the future. Building an infrastructure today is meant for the future, not today. (at least IMHO).

I would say: “True”…
I am a big fan of wired solutions, mainly because of reliability and autonomy. My walls are full of Cat6 :slight_smile:
I deployed a KNX system in my house for controlling the Lights + Roller Shutters + HVAC and I am always trying to avoid adding Wireless components (it’s a challenge… many cool new items are RF only).
There is a significant startup cost involved (commercial KNX Actuators & Push Button interfaces are expensive…) but I believe that their Wireless versions are even more expensive.
If you have any questions on the KNX Systems Architecture: Fire away :slight_smile: I can also link some reading material for you if you would like.

For a typical KNX System (wired):
Component 1: KNX Actuator (4 Channels/Outputs, General Loads) - Main Electrical Distribution Panel (DIN rail)
Component 2: KNX Line Power Supply - Main Electrical Distribution Panel (DIN rail)
Component 3: KNX Push-Button Interface - Wall Socket or Intermediate Distribution Panel
Component 4: Push-Button - Wall Socket
Component 5: KNX Bus Wires (2 wires - Twisted Pair)
Component 6: Mains Power Wires (to the Load)
Component 7: The Load itself (the Light Bulb)
Component 8: (optional) KNX/IP interface/router for external control (via OpenHab)

The Load (C7) is connected to one of the Actuator outputs (C1) by using regular power wiring (C6).
The Light switch (C4) is connected to the data interface (C3) by using bus wires (C5). Both units (C3+C4) are installed in the wall socket.
The data interface (C3) is connected to the Actuator by using bus wires (C5).
When you press the switch, a data telegram is generated and is received by the Actuator which then controls the electrical relay for the specific load.

I will design this stuff in a small image and upload it here to be able to see what goes where.


Mostly false. Wired is more reliable, that’s true, but meanwhile there’s also pretty reliable wireless solutions. But wired is not as flexible as is wireless: you need power and control lines everywhere, or have a fully centralized wiring. And you mustn’t think of lighting only, think home automation and long-term. Think of hard-or-costly-to-reach sensor locations such as door and window frames or locations suited for motion or smoke detectors, think of mobile lamps and household appliances in general. Not to mention the HVAC stuff, but that would lead off-topic.

  1. a transformer (if not 110/230VAC), 2) suitable wiring (depends on voltage, can be Cat 6 for 12/24VDC but not for 110VAC), 3) a sensor to attach the switch to, 4) an actuator to attach the light to, 5) a home automation server and 6) a control network to link 3) and 4) to 5).
    Depending on technology, 1) may not be needed and 2) is there. The server is somewhat optional. You will want to have one for advanced functionality (event based), but you will also want your basic lighting to still work if it’s unavailable. ZWave, LCN and KNX based lighting - if done right - will still work without a server. 3) and 4) are often combined in one device (mostly found with ZWave, LCN), but can be separated (often found in KNX).

Cost is very difficult to compare because it will depend on what you want to have in the end (and you obviously don’t know yet). Let me advise you that the most costly part is bad or changing plans: to run wires to one location, then finding out you need or want to change actuator or sensor or switch location means you need to re-wire. There’s many reasons for this: the motion sensor does not trigger, the lighting scenery doesn’t look the way you imagined, the temperature sensor is too close to other heat sources to be representative. The switch is not where your wife wants it to be etc. That’s all stuff you often don’t find out unless you believe you’re done - including wall decoration.
But now even a partial redesign means to run new wires and re-decorate the walls.
And that’s where the cost of a wired installation is hiding!
You need another electrician, another painter, material, possibly new devices …
For a similar thread, see

Also, believe me that even with good planning, you will want to re-design parts of your setup sooner or later, simply because your taste or usage is changing over time.
Note the determinant for a successful smart home installation is the WAF.

To give some guidance: First, you can’t cover all usages well using wired devices, so I’d prepare for a combination of wired and wireless.
Your intro is a little unclear whether you’ll be re-building everything from scratch. If so, I’d run wires to one central location wherever applicable (from switch and lights locations). Or at least to one (big enough !) hub per floor level. Have empty pipes to prepare for more wires to add in the future. Have enough cat 6 or 7, but don’t rely on it as a single cabling system for control and low power lighting, also add 110/230VAC capable wires.
To centralize cabling leaves you with the choice of installing a wired hub (multi-port actuator) or bus inside the wiring cabinet, or a set of wireless actuators. Enables you to physically access and exchange devices or even the whole system if needed.

Now as KNX is a mostly European thing, I would not recommend it to Canadians: limited choice of suppliers, import surcharge on top of already high cost, few electricians available to know how to install it - that’s all worse for you than is for the Europeans that run and recommend it.
Plus, my personal opinion is that KNX is expensive, it was fine in the 90’s, but now it is no longer state of the art but still as pricey.

On the other hand, I also wouldn’t go with a proprietary solution such as LCN, Insteon or Loxone. As someone already commented, vendors come and go. They also raise prices, sometimes have bad products but no alternatives in their portfolio, fail to provide software fixes, they stop product lines etc.
It’s fine to use these if you’re aware of the implications and prepared for a full exchange someday in the future.
Note, however, there’s many proprietary systems to NOT work with openHAB. I believe LCN and bticino/LeGrand do, HomeMatic Wired does, and Insteon hubs can be made to interwork, but that’s a whacky one.

A number of people use (old-style) PLCs or (new-style) RPis with I/O boards or connect relay banks to RPi’s GPIO pins to make them act as multi-port actuators and/or sensors. That’s the DIY variant. If you have the skills in terms of electrics, it can be made working. But you need to be aware of all its implications - it’s DIY stuff, noone will be able to implement and maintain it other than you.

So unfortunately, I believe we’re out of really good wired options.
At least I don’t know of a wired and standards-based system.

That’s IMHO leaving you with one of these options:

  1. KNX
  2. a proprietary but openHAB-supported system such as LCN
  3. DIY stuff such as relay banks and sensors attached to a RPi
  4. an all-wireless system such as ZWave or ZigBee

If you have the time, skills and patience, you could try the DIY route. But I’d stick with a standard system and keep DIY as a backup or hobby.
Now since ZigBee does not work in OH yet and you already have some ZWave stuff going, I’d continue exploring that path and continue gathering experience with it. Get some alternative devices (Fibaros are quite good), deploy dimmers, LED actuators w/ strips in a limited area (say your basement or entry hall), find matching bulbs, get motion detectors, implement your first openHAB rules. Find out usage patterns to work well or not. Find your own style before you re-build.

Thanks all. It was good to have a couple of more concrete examples to confirm/correct/guide my previous understanding.

I picked lighting because it’s a common thing and I hoped to have some people chime in with the way they had implemented it and what components they had used, so that I could understand what I might need if I did it in a similar fashion. It was just an example that I picked to try and keep the question from being too broad … which I see it probably was anyway.

Regardless, I’ll keep experimenting, reading and talking to people. And this won’t be my last plea for help I’m sure.

@Dim, I might indeed bug you for more info about KNX but as @mstormi mentions, it may be a poor choice because of my location (lack of supply and expertise).

I’ll have to give some more thought to how I can marry my own desire to learn and experiment with more polished solutions that please the other users of the house.

I am also building a new house. My priorities for light installation are:

  1. It has to work always (I travel a lot, and if it will stop then my wife will kill because only I will be able to fix it)
  2. It has to be cheap
  3. It has to consume no electricity in stable state
  4. I want to have also standard wall switches
  5. It has to be flexible

And my solution so far is:
Power cable is cheap and stays forever. So each light (CYKY 3x1.5mm), all wall sockets (CYKY 5x2.5mm = 3 wall sockets in the room), all wall switches (CYKY 2x1.5mm) will go to electrical cabinet (80x200 cm).

And because smart home is never done, and there is always something to improve, and I will not have time to play from the beginning (you need to finish all the small things in the house), there will be possible:

  1. Connect all sockets directly to power. (No automation) All lights will be possible to control because I will connect them to this relay: http://www.tme.eu/gb/details/ (For this type of installation its easier to pass electrical tests. At least here in Czech Rep. you cant just take some devices, put them together and use them in house installation legally)
  2. Then later, add some contactors for wall sockets, but only there where you need it (its expensive and high power relays are not bi-stable, so they consume electricity when ON). Use rPi (or similar) with IO expanders to sense state of relays and control them with tine relay or tyristors/transistors parallely connected to the switches.

And why I don’t want to buy some commercial solution? I work in industry (automation). And sometimes I have to replace old control system with new one. It is because after each 10-20 years (in home automation much more often) you are force to do it because you cant buy any spare parts.You have no support.They are not flexible.

I am open for any comments.
Best regards

1 Like

I hope you are aware this applies to everything else: car, fridge, TV, door locks, etc.
And I would argue that a Linux machines runs longer without fault than a car. Mind you that I have seen Linux machines up and running for 10 years!!

I reckon, what comes out of this thread and others:
a) run wiring that can be used for control and power
b) power cables can carry 12V for electronics, but can be re-purposed to run 230V as well.
c) if you want reliability to the n-th degree, you have to implement fault-tolerance to the n-th degree. And remember, every 9 behind the comma will cost 10 times as much.

While I agree with your stance on commercial / proprietary solutions, however, I also believe that a home automation system can be built with standard components, sensors, actuators and wiring, so that others (of course with the respective knowledge) can maintain them.

If something happens to Linux machine, its not a problem. Lights will “always” works, because they use standard bi-stable relay. Automation is connected only parallel to it.

You can use standard components (sensors, actuators and wiring) but there is nothing like standard controller.

To be clear: for safety reasons, don’t use Cat 6 for lighting if possible.
If you deploy wires, deploy 1,5mm² NYM 230VAC capable (or the equivalent US/Canada electrical standard, to be precise).
The lower the current, the higher the amperage, that is, the thicker the wire needs to be !
And you don’t know what the next-gen device or light will require in terms of wiring, no matter when.

The lower the tension, the higher the current :wink: or the lower the voltage, the higher the amperage…