That’s exactly what I discovered. I had to put my desire for (what i considered) super reliable wired VS cheaper wireless solutions.
Eventually it was obvious I could do a wireless solution at a fraction of the cost and it was infinitely more versatile.
I’ve lived it the house 6 months now and the only problem I had was after a power cut when I hadn’t set my VM home assistant to autoboot. Not sure being wired would have made any difference!
Thanks a lot for your input. I need to decide what to do with my house, and I have to wire it soon, so I don’t have a lot of time to decide if going with one or the other… wires sounds much more reliable, but a single KNX swich cost around 100 euros, when i can get a wifi one for 20. Since you are already doing the test i would love to have done, Can I ask what devices do you use?
I had a look at Loxone when we were building, I saw some users on a local forum that use it, and that are very happy about it, very easy to setup and change things around. However I found that the system is quite ‘closed’, so getting data out might be harder to do. 1 of the reasons to choose for KNX over Loxone for me was the vendor lockin, with KNX you can combine any manufacturer that makes KNX stuff, it’s a standard which isn’t going to go away anytime soon
Yeah - this was my experience when reading about it too. There is no “choice” in regards to switches or other things because you need them from the vendor. Are there any good resources to get started with KNX I can read to get a better understanding of what I need from the electrician and what I can do on my own?
HA KNX integration is working very good and stable (thanks @farmio )
For me this book was really helpful:
“Heimautomation mit KNX, DALI, 1-Wire und Co: Das umfassende Handbuch” Stefan Heinle (german)
I did almost everything myself: Wiring the house and the “control cabinet” and the programming.
I think my own costs were approximaly 10.000 € (including wires, switches, all the KNX stuff for light, cover, room temp. contol, dimmable lights almost everywhere, AFDD-decives …) + 4.500 € for the electrician (mainly grounding and “electric meter cabinet”).
If you have to pay the working time for installation and programming, then it’s much more expensive of course.
I’m also almost committed to do a Sternverkabelung for everything and CAT6 everywhere - including certain outside locations.
I’d like to automate almost everything (I already have where I live now) but I don’t think KNX is necessary.
Running every cable to the utility room gives you a lot of flexibility regarding which solution you end up using or may be using in 10-20 years.
Right now I really like Shelly and while I’m hesitant to put 50+ wifi shelly devices on a DIN Rail… I really look forward to the Shelly 4 PRO PLUS which should be available next month and is available with an LAN Port.
So far my Shelly 1PM and 2.5 devices have been pretty reliable.
Regarding WAF you should keep in mind, that if you wire each switch to the corresponding shelly, you can archive the basic functionality “out of the box” and without any central unit present. So if your HomeAssistant fails… reboots… does not matter… you would still have basic light on/off functionality everywhere!
I’m planning on using momentary switches (Taster) everywhere and not the typical on/off switches which would normally be installed.
Basic functionality would be available without reliance on any wireless technology.
Another plus for Shelly: Power metering … which I use in several automations and which enables you to make dumb household appliances smart. With an basic KNX install… even if you could power cycle your for example washing machine … it probably could not provide information about consumption and therefore status of your connected devices.
I also want to install motion sensors everywhere… and maybe they will be wireless zigbee devices. An wired option would be “dumb” PIR sensors + Shelly i3 (one shelly for three sensors). But if I stick to zigbee for this… I could use sensors which not only sense motion but also brightness and temperature. Also they would allow setting sensitivity and can be placed far easier than wired ones.
Sure, switch actuators with power measurement are available for ~30-40 € per channel (capable of using multiple lines a 16A). Pure KNX energy meters are also available.
I talked to a few of the contractors and most of them do not know anything about KNX If I run a “normal” cable for each switch/light down to the utility room into the utility box, can I start “KNX” from there (in a separate (or bigger) utility box)? Or do I have to run KNX cable from the utility box to the switches If I want to use it?
Yes and no…
You can run normal cables to lights, sockets etc. and have all actuators in the utility room. This is the preferred knx style.
But switches and presence detectors are connected directly to the knx bus. So they need the Green cable.
You could also use normal switches with 1,5 mm2 standard cable to the utility room and connect them to binary sensors there, but why then even bother using knx at all - a huge benefit of knx (at least for me) is nice looking, hard wired sensors (switches) with powerful applications.
For lights I’d use at least 5x1,5 (if you do 230V) so that you can always opt for DALI later. Copper is not that expensive.
I have the exact same thoughts. I bought couple of devices to test, and started the planning now. I still think knx will add some zeros to the math, and can’t really see the benefit
We talked about prices being high for KNX. As this is a standard more and more companies going into this field. Honeywell being one of them recently. At least for me its new. 120€ for an 8 channel actor. This is a new level.
I have not tested them nor any affiliation with the shop below. But this seems interesting.
I would recommend KNX as it is a global platform and the contributing manufacturers are increasing everyday. I have my house setup with KNX and all infrastructure (Switches, Actuators, Keypads, Touch panels, etc…) from Zennio, but you can choose from a wide variety of providers. I have 1Home Bridge for Google Assistant and Apple HomeKit but you can do this with Home Assistant as well.
By end of 2021 KNX ORG are planning to release a new version ETS 6 of their Programming tool ETS which is more IoT oriented, although they are compatible with IoT now.
Hi, I am in a very similar position and would love some advise from people.
We are in the process of buying a new build apartment here in Germany, which is already planned with a basic Gira knx install. However, now that it is time to put forward our “sonderwuenche” - special wishes, the cost is growing and I suspect part of the problem is how the lighting is being designed.
Currently, all switches are knx buttons. Heating is underfloor heating with knx control, and a Gira video doorbell - so far so good. The question is about lighting.
The default that is provided is simply switching of lights somehow with knx. Dimming is a significant extra cost of about 200EUR per circuit. I suspect that this is a simple 230V knx dimmer. We will generally want LED downlights in a drop ceiling, or some hanging lights, but ideally also LED based. My wish was to have everything dimmable, and I would love to have RGBW or at least colour temp adjustable.
So my question is which way is best to go technology wise for the lighting? I am happy to mix and match technologies a little if it makes sense. I already have a few z-wave sensors and Wifi lights that I have been playing with in the past but this would be on a bigger scale.
From my understanding, knx with dali for the lighting might be the right option here. I would love any advise.
Further info, WAF is always important, but my wife also loves tech, so maybe can forgive the odd issue. I already have the ETS6 license, as I figured I would be needing this and I made use of the black Friday specials.
I suspect the high pricing for the special wishes are due to the building company using expensive 230V dimmers from Gira. I also need to decide if I pay extra for the Gira X1 knx gateway, or if some alternative gateway would be better.
I second KNX for basic lightning (on/off) and DALI for all dimmable lights. 5x1.5mm in one or multiple rings for all DALI ensures simple cabling structure. Happy to exchange in a more detailed way. I built my house in 2006/7 and have mainly KNX/Dali and recently some Hue.
If cost is a concern, you can go with ESP, hue, or WLED and alike. Connection via Wifi.
I am not a big fan of having critical functionality (i consider most lights to be critical) relying on WiFi. Thats why I stick to KNX/DALI combination. But that’s my personal choice.
I’m going for standard wiring (Switzerland) to a central box. With the plan that in there I can do whatever I want.
Yes! this is a lot of extra wires and pipes, but looking around, most new stuffs I find around are not KNX, or if there is a KNX version, is several orders of magnitude more expensive for no reason (like presence detectors, smart devices like my dishwashers, garden watering systems, solar pannels, etc)
My case might be special, because I have enough space at home to put as many central fuse boxes as I want, with tons of wires centrally going there.
I am starting with my own PCB for lights controls (for several reasons, here the link if you’re interested), but my setup is flexible to install any device (like shelly or KNX) in the future if I want to.
In my case, trying to predict what I was going to need was not really working… too much to learn in too short period of time
Have you had a look at loxone.com as of yet?
+++ rock solid, cheaper than KNX, the Home Assistant Loxone integration works flawlessly which was important for me - I wanted the stability of a wired system but the nice user interface of HA
— People will tell you “you’re bound to 1 supplier only unlike with KNX” (as Loxone is the only supplier) → yes right, but they already have >>100,000 installations that they have to supply with hardware. I guess that’s a pretty good business case and if you want any KNX hardware you can always integrate it with their KNX interface.
The good thing is that even in your case where you start with a classic installation you can retrofit Wireless Loxone (Loxone Air) which is again rock solid. Unlike with my Zwave sensors I have not had any problem so far with the Loxone, at all.
I’ve purchased a few hardware items from them (I started with a Miniserver Go Air, some door sensors, wireless wall plugs etc.) just for learning. Based on this experience I will now build my new house with this system (wired, not wireless but in case I forget something - which I for sure will - I can add it anytime with their wireless hardware).
I have seen a lot of reports about Loxone being very not rock solid, burning through SD cards on a regular basis. But have not tried it myself - maybe that problem is long fixed.
There is no Loxone integration in HA core, so I guess this is only supported via a custom component.
For what its worth, Insteon had about 1.3 million users when they went out of business lately
I’d not go down that route as Loxone is known for their poor implementation of the KNX specifications. Their old KNX interface was not certified for KNX use and caused some troubles in KNX land due to bad communication. Maybe their new (~600€) KNX extension is better, but as far as I know, it is still not KNX certified.
Well I guess there is always a way to discredit something.
SD cards: Loxone recommends its own Industrial SD card and advertises up to 60 times longer life span than with conventional SD cards - it’s probably the wrong approach to pay 1€ instead of their 12€ as I suspect a few to do. Also using real time statistics in Loxone will increase the wear of an SD card incredibly due to the much higher data storage rates. The user has it in his own hands how much tear & wear the SD card is taking. But it’s anyway the same with HA - many users are basing HA on a RaspPi setup with a (mostly cheap) SD card.
KNX hardware: One of my best mates runs an electric installation firm, he’s implementing lots of Loxone installations, many people are going for KNX switches and push buttons instead of the Loxone ones (Loxone has only a few designs) through the KNX interface without paying the price for the KNX bus of the system backbone. While this one interface you quote (the 600€ which you chose to quote are only a part of the entire equation).
All in all:
If someone wants to go for KNX and also wants to afford it: do it - KNX is great!
My personal choice was to go for Loxone as the costs are significantly lower (ca. 40-50% depending on the use case and hardware used)
Plus: I’m marrying up Loxone with HA which works great for me - I have the stability of Loxone and the amazing user interface of HA (I’m hosting HA on a virtual machine of my NAS). And yes you’re right: the Loxone integration is a HACS one.