EnOcean + DALI + LinuxMCE, doable ?

[Crab]

Have been lurking on forums for some time while renovating my apartment. Now it has come time to set up lighting and some electrical stuff.
Problem is that planned lighting is not really feasible on my existing wiring (especially control wiring) and in order to get things connected bus type solution like DALI would be optimal (since some new power wiring is needed so DALI wiring to light fixtures would create little extra trouble).
So in order to avoid new wiring for lightswotches wireless solution would be optimal and after checking both zwave and EnOcean out I am leaning on latter.

Now I am wondering how well LinuxMCE, DALI and EnOcean would work together. Have any of you set up this kind of system and how easy/hard it was, is it working out etc.
Also the wiki indicates that EnOcean works with LinuxMCE, but any personal experiences with enocean system would be apreciated as well.

[grind]

Hey Crab,
i am using Enocean and wanted to use LinuxMCE.
There is simply NO documentation on how to implement Enocean in LinuxMCE. Only for the USB-Gateway and the rocker switches.
Beside that there is no support in this or other forums.

To summaries: I am going on using IPSymcon for my home as a quick solution. Thats a piece of Windows-Software which is very easy to use and supports nearly everything you can connect to a PC in any way (IR, RS232, USB, LAN, ...).
More information you can find on www.ipsymcon.de

The other option is to use Z-Wave which should be supported very well within LinuxMCE. Beside that it's much cheaper than Enocean.
But it also has disadvantages like batteries for example.

As soon as my IPSymcon environment is set up i am out of pressure and will start learning, developing and documenting for LinuxMCE.
My first goal is to fully implement Enocean and 1-Wire. As i have a lot of that stuff here all the requirements are fullfilled.

Regards,
Grind

[hari]

that should be doable by using EnOcean rocker switches and a KNX/Dali Gateway.

br Hari

grind: What documentation or support needs do you have regarding EnOcean? I did write the implementation but there were no users, you seem to be the first one :-)

[Crab]

Batteries did put me off from z-wave and EnOcean also seems more reliable, but that can be just better marketing  , also price is not big obstacle since apartment is only 3 rooms so I am looking to fit prolly about 10 switches or so, maybe some electric blinds/PIR later. Idea is that linuxMCE would control more 'advanced' functions through DALI.

After quick internet browsing found DALI master and enocean modules from Wago which should work:
http://www.wago.com/cps/rde/xchg/SID-41B192EC-3F263566/wago/style.xsl/fin-1903.htm
http://www.wago.com/cps/rde/xchg/SID-41B192EC-3F263566/wago/style.xsl/fin-1901.htm

And even after power supply + fieldbus the total price would be lower than most DALI controllers around here... however if it could work with just enocean modules and KNX-dali gateway it would be simpler, but are you sure there is no need for EnOcean gateway between KNX-dali gateway and EnOcean 'net' ?

Also is there any english sites about this IPSymcon... quick look around the site didn't reveal any english manuals ?

[totallymaxed]

Batteries did put me off from z-wave and EnOcean also seems more reliable, but that can be just better marketing  , also price is not big obstacle since apartment is only 3 rooms so I am looking to fit prolly about 10 switches or so, maybe some electric blinds/PIR later. Idea is that linuxMCE would control more 'advanced' functions through DALI.

After quick internet browsing found DALI master and enocean modules from Wago which should work:
http://www.wago.com/cps/rde/xchg/SID-41B192EC-3F263566/wago/style.xsl/fin-1903.htm
http://www.wago.com/cps/rde/xchg/SID-41B192EC-3F263566/wago/style.xsl/fin-1901.htm

And even after power supply + fieldbus the total price would be lower than most DALI controllers around here... however if it could work with just enocean modules and KNX-dali gateway it would be simpler, but are you sure there is no need for EnOcean gateway between KNX-dali gateway and EnOcean 'net' ?

Also is there any english sites about this IPSymcon... quick look around the site didn't reveal any english manuals ?

Can I ask why in a 3 Room installation you would not just go with a straight ZWave or ENocean based installation... we do a lot of this size/type of installation and to be frank ZWave is damn hard to beat in terms of simplicity of installations, price, flexibility and reliability. I think you could say much the same about ENocean with exception of 'price' ;-)

If KNX/DALI is of interest from a technical perspective for you personally then I can see having an installation to 'experiment' with would be good. But from all other perspectives it would seem a non optimal choice.

All the best


Andrew

PS Hari we're still looking for our first ENocean installation from one of our installers.... fingers crossed ;-)

[Crab]

Lighting is going to like this: general lighting provided with dimmable fluorescent lighting (indirect/direct fitting in cove spanning 1-2 walls in each room + kitchen) providing most of the lighting and then just few specific lights in workstations and so forth. This setup requires quite a few electronic ballasts (maybe about 10 or so) and there is not that much difference in 'regular' ballast vs. DALI ballast (about 50 euro a piece) and since new powerwiring is needed then there is little extra work to fit in the dali (just use 5 wire line instead of 3). In addition of ballasts the DALI controller is giving extra cost, but WAGO one seems reasonable.

With just EnOcean or ZWave setup there would be same amount of switches, but I would also require some sort of dimming actuators which cost around 80 euro or so (just quick look around net)... and since lighting setup would require about 4-5 or maybe even more lightgroups and the cost of these actuators would make the gap to DALI 'decent'. However, I am certainly no expert in home automation and might have understood whole thing wrong so any input is always apreciated.

But off course this whole setup is not really simple at all... if I wanted simple I would have just left the lighting as it was so there is indeed a strong 'experiment and tinker' attitude as well, which is why linuxmce also caught my eye.

[grind]

Hari, thank you for answering in my thread because of the Enocean issue.

Crab: Enocean is really a great technology and as soon as it's fully integrated in LinuxMCE it's the first choice. At least for me
Z-Wave is a good product too and it's getting more popular every day but i don't want to swap batteries...

Just one thing you should think of: Whats if your core crashes and there is an emergency in your home. I am sure you don't want to tell the medic that you first have to fix your server. So in my opinion every room should have a rocker switch which directly controls one light. You can also integrate those lights into LinuxMCE or whatever but that has to work independent as well.

Regards,
Grind

[totallymaxed]

Lighting is going to like this: general lighting provided with dimmable fluorescent lighting (indirect/direct fitting in cove spanning 1-2 walls in each room + kitchen) providing most of the lighting and then just few specific lights in workstations and so forth. This setup requires quite a few electronic ballasts (maybe about 10 or so) and there is not that much difference in 'regular' ballast vs. DALI ballast (about 50 euro a piece) and since new powerwiring is needed then there is little extra work to fit in the dali (just use 5 wire line instead of 3). In addition of ballasts the DALI controller is giving extra cost, but WAGO one seems reasonable.

With just EnOcean or ZWave setup there would be same amount of switches, but I would also require some sort of dimming actuators which cost around 80 euro or so (just quick look around net)... and since lighting setup would require about 4-5 or maybe even more lightgroups and the cost of these actuators would make the gap to DALI 'decent'. However, I am certainly no expert in home automation and might have understood whole thing wrong so any input is always apreciated.

But off course this whole setup is not really simple at all... if I wanted simple I would have just left the lighting as it was so there is indeed a strong 'experiment and tinker' attitude as well, which is why linuxmce also caught my eye.

So if I am understanding correctly you have say 5 circuits to control (some require dimming). So if you used Duwi load controllers for switching & dimming that would come to about 60 Euro's per circuit (including surround & paddles). So lets say 300 Euro's for the load controllers (inclusive of paddles etc) and then lest throw in some wireless ZWave switches that can be surface mounted anywhere you need them - the Duwi model is about 45 Euros (and either of these can act as the Primary ie for inclusion or secondary controller). So for less than 400 Euro's and some time with a screw driver you'd be done... no new wires to run/hide and no holes or other work to do (assuming all the actual lights are in place that is). If you dont have a coffee this could all be in and basic set up complete inside 60-75 mins... :-)

...and this would still offer you some 'play time' too of course ;-)

If you Core ever needs to be offline then your ZWave lighting will still work independently from the wall switches as long as there is mains power.

All the best


Andrew

PS Anyway whichever way you decide to go you'll have a lot of fun doing this I'm sure!

[nite_man]

Whats if your core crashes and there is an emergency in your home.

IMHO at least all important switches should have possibility to manage them without any smarthome system, i. e. it's better to use both: software control and manual as well. It isn't good to relay to any kind of controller because in any case it's just computer and it can crash or freeze. And it won't be funny if you won't open garage door to go to work morning because you didn't install simple switch to open door even without your smarthome system

[totallymaxed]

Whats if your core crashes and there is an emergency in your home.

IMHO at least all important switches should have possibility to manage them without any smarthome system, i. e. it's better to use both: software control and manual as well. It isn't good to relay to any kind of controller because in any case it's just computer and it can crash or freeze. And it won't be funny if you won't open garage door to go to work morning because you didn't install simple switch to open door even without your smarthome system

Well i think the above goes without saying. ZWave, ENocean & KNX will all provide basic control of lighting and som other functionality without the Core being up ie you press a light switch paddle and the lights fade up/down even when the Core's is off-line.

Andrew

[trentend]

I think it's a very important design consideration.  To the extent that I have designed my home with full independent autonomous systems.  I'm only using LinixMCE to integrate those systems together and to add functionality (....for example turn the lights on, if someone enters the room, and it's dark).  For me LinuxMCE is the icing on the cake. I would be uncomfortable with it being the whole cake.  It's not just the obvious that you need to consider.  How do you feel about not being able to make or receive phone calls if your core is down? What provision have you made for making emergency calls?

There are certain things that you need to keep working irrespective of the status of a server. It's a reason that I'm so pleased with KNX.  The equipment is of a high quality and has industrial level reliability.  For other systems I generally have one self contained subsystem capable of working without Linuxmce for each important function - except security/alarm which is not a big concern to me, but if it becomes so I have planned everything so that I can drop in an approved alarm box.

[totallymaxed]

I think it's a very important design consideration.  To the extent that I have designed my home with full independent autonomous systems.  I'm only using LinixMCE to integrate those systems together and to add functionality (....for example turn the lights on, if someone enters the room, and it's dark).  For me LinuxMCE is the icing on the cake. I would be uncomfortable with it being the whole cake.  It's not just the obvious that you need to consider.  How do you feel about not being able to make or receive phone calls if your core is down? What provision have you made for making emergency calls?

There are certain things that you need to keep working irrespective of the status of a server. It's a reason that I'm so pleased with KNX.  The equipment is of a high quality and has industrial level reliability.  For other systems I generally have one self contained subsystem capable of working without Linuxmce for each important function - except security/alarm which is not a big concern to me, but if it becomes so I have planned everything so that I can drop in an approved alarm box.

Well we certainly agree with that principle. However any hardware can fail... even KNX... and it can fail in such a way to cause a 'domino' effect as well. The approach we take with our Dianemo Core's & MD's is to use server grade components through out with ECC memory, processors & motherboards, dual redundant power supplies etc etc.

All the best

Andrew

[tschak909]

Andrew,

years ago, I saw the TRON demo house in Japan.

It ran on a matrix of operating systems based on the TRON microkernel (The central server ran preemptive CTRON, the PC workstations inside the house ran BTRON, the various touch panels, ran a micro TRON),

there are plenty of papers detailing the house (all the way down to the fact that each and every doorway was automatic)

all in all, it took 380 individual computer systems to run the entire house, and the feature set was roughly equivalent to Pluto, with late 1980s technology. The cost of the home and its electronics was in the double digit millions.

and the irony is..

they STILL couldn't account for every single possible failure, even with a family of hard-real-time OS with context switches measured in low microseconds. They spent five years debugging that house, and ultimately shelved it.

It has been revived several times, the latest one being in Taiwan (first shown in 2004) ...

but really, no matter how much you throw at it, you're going to hit the inescapable boundaries of a given design pattern's complexity. Which is why I think throwing more and more hardware at a problem and distributing it out can help, but you are trading all of this for the fact that you're maintaining a much larger set of systems.

 Most people simply can not do this.

-Thom

[totallymaxed]

Andrew,

years ago, I saw the TRON demo house in Japan.

It ran on a matrix of operating systems based on the TRON microkernel (The central server ran preemptive CTRON, the PC workstations inside the house ran BTRON, the various touch panels, ran a micro TRON),

there are plenty of papers detailing the house (all the way down to the fact that each and every doorway was automatic)

all in all, it took 380 individual computer systems to run the entire house, and the feature set was roughly equivalent to Pluto, with late 1980s technology. The cost of the home and its electronics was in the double digit millions.

and the irony is..

they STILL couldn't account for every single possible failure, even with a family of hard-real-time OS with context switches measured in low microseconds. They spent five years debugging that house, and ultimately shelved it.

It has been revived several times, the latest one being in Taiwan (first shown in 2004) ...

but really, no matter how much you throw at it, you're going to hit the inescapable boundaries of a given design pattern's complexity. Which is why I think throwing more and more hardware at a problem and distributing it out can help, but you are trading all of this for the fact that you're maintaining a much larger set of systems.

 Most people simply can not do this.

-Thom

There are many examples of these kinds of complex HA solutions (although they were certainly not called that at the time!)... the work carried out at the MIT Media lab and of course Xerox PARC too with its Ubiquitous Computing project all spring to mind. All of these systems, and I include the TRON house alongside these, were based on much more fragile hardware (and to some extend software too)...and all suffered relatively high technical failure rates.

However even with our earlier line of systems based on essentially consumer grade hardware can point to really very impressive reliability levels indeed - and of course we have i believe the largest body of continuously deployed & in-use Core's that have ever existed - and therefore we have very good data indeed on MTBF etc etc. We have many systems now that have operated without incident for over 20 months - 24-7 and these systems are heavily used too i can assure you!

However our move to server grade hardware was motivated not just by reliability but also equally by longevity of the production runs on these hardware items - we have a guarantee of production for from 4-6 years and after-life availability of another 4 years... and also a desire to use less complex hardware in fact. All of this adds up to stability and predictable hardware support. These in the end are far more important than anything else when you are trying to build a reliable product that people can build their family activities and life-styles around. Our newer hardware is in many no more, and in some ways far less, complex than the previous generation of hardware - server hardware is inevitably more conservative in design and often uses relatively older components for example - so we are definitely not 'throwing more and more hardware' at the problem at all.

All the best


Andrew

[Lexje]

solution like DALI would be optimal (since some new power wiring is needed so DALI wiring to light fixtures would create little extra trouble).

FYI: The DALI system is purposed for LARGE installations, and focuses only on lighting. It's purpose is to indeed have a single control cabling; each light device would have its own ADDRESS. This way, one could, at random and at any time sub-divide e.g. 300 light devices into, say, left alley-way, middle alley-way and right alley-way, or whatever grouping you'd see fit.
Basically this is really for hotels, government buildings, universities and such.

If you insist on a wired solution, and certainly if you're within EU, I would advise to go the KNX route.
You'd be following the same principle route, i.e. each light appliance will have its own ADDRESS. For KNX there is also only 1 signal cable which is included into the KNX 30V power supply line.
Depending on the cabling of your apartment, you could also use PL KNX (PowerLine), in which case you wouldn't even need the 30V KNX power supply cabling.

As Hari already stated, KNX is integrable with LMCE.
KNX may ultimately have other advantages such as ease of interconnection with ...

KNX may be a little more expensive as Zwave, prices are coming down these days.
It may also be worthwhile considering power monitoring..

Best regards,

Erwin