Overall Z-Wave impressions and reliability

[Phobophile]

A question for James and other Z-Wave users out there. At the moment, I'm thinking about the ways to implement a home automation system for a new apartment. For now I'm looking at Z-Wave, but I've heard about some reliability issues with it, like lost signal and something like that, especially in places with a lot of "wireless noise". Anyone noticed something like that? I could go with some wired bus since the apartment is without any finish yet, but all the other system I checked, in my opinion, have some downsides - price, limited choice of software and hardware, ugly software and hardware etc.
So, are you happy with Z-Wave? Any advice would be welcome! 

[jitterjames]

I can only share my personal experience which is somewhat limited, so hopefully other will have something to say on this topic as well.  So far I have found Z-Wave to be 100% reliable.  I have had some issues with Vera but they are not related to reliability, and for the price, and ease of installation compared to wired solutions, I have been pretty satisfied overall.

We live in a fully detached house in the country so the only wireless noise around here is what we generate ourselves, and our new "smart meter" that the power company installed.  Our house is not giant (about 2000 sq-ft on two floors), and we don't have that many Zwave devices installed.  Only 6 lights and 2 sensors.

Sometimes we can see a couple of our neighbours' routers showing up.

[aussie mate]

Phobophile,

while I have not used ZWave - I have used Insteon and x10 for HA.
x10 has it issues but with careful planning & a bit of luck it can be made to work reliable.

However IMHO for the HA system with the best reliability  - you cant beat Insteon - it uses a dual mesh system of both RF & PLC signals so every device can repeat and forward commands to each other. I have about 12 lights and 9 motion and 4 other types of senses that all work reliably.

[jitterjames]

Not that I want to appear negative, but I cannot recommend against X10 strongly enough!

[jitterjames]

If you are serious about making this decision I think you should look much further than the VC forum.  There are many forums out there that focus more on home automation.  You will likely read many good and bad things about all of the other options.  You will need to take some time and look at multiple sources and then just go with your gut.  Obviously if you are in a position to buy with the option to return stuff that is best so you can minimize your risk, but for most of us the home automation stuff is primarily mail order only.

Just doing a quick web search for Zwave vs Insteon brought me to this page: http://www.avsforum.com/t/1191332/z-wave-vs-insteon-and-others which has some interesting perspectives.

I would not take any one opinion too seriously, but looking at multiple threads where different people chime in can give you an idea.  The impression that I get is that more people have problems with insteon than with Zwave, but that is just an impression and I am probably biased by the fact that I have found Zwave reliable and after my fiasco with X10 I am maybe too easy to impress.

If you want a 100% reliable solution (well there is no such thing but...) then maybe you want to go with something that has dedicated wiring.  This is obviously more expensive and harder to install but if you are in a new place maybe this is a viable option for you.

[aussie mate]

I was not recommending x10 - just pointing out that it can made be work reliably - it depends on many factors.

regarding Insteon - I understand that old insteon "single band" devices did suffer from issues - however the dual band devices have sorted those issues and are now very reliable.

As jitterjames says - check out all the HA forums to work out what you think is best for you & want you want to achieve.

[garryjw]

Just a heads-up that a very senior representative of the Z-Wave alliance (might be the chairman but awaiting confirmation) is due to be interviewed on theDigitalLifestyle Show podcast next week which I co-host. If anyone has any specific questions I'll see if I can get them answered.

Cheers
Garry

[jitterjames]

Yes.  I have some.  I think I already know the answers, and I already don't like them but it doesn't hurt to ask them anyway.  We'll get back to you on this one.  Thanks!

[Phobophile]

Wow, thanks for alll the suggestions! Of course, I've checked other sources of information and there are holy wars all over again. 
I'll check further about Insteon.
X10 is not an option, too many bad reviews about it... oh, wait, there's another one up there! 

[jitterjames]

Just a heads-up that a very senior representative of the Z-Wave alliance (might be the chairman but awaiting confirmation) is due to be interviewed on theDigitalLifestyle Show podcast next week which I co-host. If anyone has any specific questions I'll see if I can get them answered.

Cheers
Garry

So maybe these are stupid questions but the first things that come to mind are:

1 - Why can't ZWave devices be priced more competitively, and is this because of manufacturer choices alone or is it caused by the high licensing costs associated with using the Zwave standard?  The lowest price for a wall switch (in Canada anyway) is $50.00  and while this is more or less in line with other standards like Insteon, I can't help but feel that this is too high for what these devices do.  I see no reason why you should pay more than $20 for a simple sensor (motion / flood / door window) but these start at $50 and up.

2 - Why doesn't the standard require devices to broadcast their status when it changes to avoid frequent unnecessary polling.  Example: If I turn my living room light on from the wall switch, my ZWave controller should be notified immediately, but with many switches this does not happen and the controller (in my case, a VeraLite) won't know about the status change until it does a poll of the device.  These makes frequent polling of many devices a necessity and it means a lot of traffic for nothing and slow updates.  

2B - I see similar problems with environmental sensors that report on temperature, humidity, and light levels.  These devices send instant messages for motion (which is good) but for the other information they use a silly method where they report to the controller that they have data available, then wait to be polled (must remain in an active battery draining mode for a period of time), and then give up the information iff the controller manages to poll them in time.  Again this seems like extra traffic, wasted battery life, and often failed updates.  Why is it done this way?

[AgileHumor]

1. IMHO, low volume (and the complexity, support, and returns).  HA just isn't mainstream yet.  Would be interesting to what he says about this.

2. My Leviton wall switches support instant status updates. Look for the feature "Supports Instant Status".
http://store.homeseer.com/store/HomeSeer-Z-Wave-Dimming-Wall-Switch-Matrix-W16C64.aspx
http://store.homeseer.com/store/HomeSeer-Z-Wave-OnOff-Wall-Switch-Comparison-W7.aspx

3.  It is odd that temp is pulled (aka polled) rather than pushed instantly (like motion tripped/untriped).    My guess is it's battery life.  To listen for the controller to poll you vs transmitting until the controller responds probably saves some juice.

Interesting note, the 'Poll this node at most once every' setting are irrelevant if the device is sleeping, it's skipped when it's turn.  It's the 'wakeup interval' that updates the temp/humidity/light, not the poll setting.  All my battery operated devices 'wakeup' at 1800 and 'poll' at 60 (which means the Vera tries often to see if the device is listening).   Inside motion sensors that are hardwired have a 'wakeup' of 60 seconds for more frequent temp updates (to bypass Z-Wave air vents).

[jitterjames]

2) Yes of course it is available as an option. I was asking why not make it a requirement as part of the standard.  That is why I put the word "many" in bold.  I can't imagine that it would cost any more to design a unit this way, so why doesn't everyone do it?

3) I think we both agree on the apparent sequence of events here and it seems clear to me that it is a flawed design.  It is not saving battery life, it is wasting battery life.  It makes sense for the device to sleep for long periods, and then broadcast its status.  If it did this it would only need to be awake for a very short time.  Wake up >> send data >> go back to sleep.  The controller is always on and I expect that it would usually be able to respond or buffer the data more or less instantly.  The way it seems to work (and as you describe it), it sleeps for a long time (while Vera is creating a bunch of wasted network traffic polling many  sensor nodes that don't respond), then the sensor wakes up on schedule, sends a message saying that it is awake, then the sensor waits (remaining in an active battery draining state) to be polled for a modest period of time, and hopefully if the controller gets back to it in time it then sends the information.  Only then can it go back to sleep.  I would guess that it is staying awake for significantly longer this way than if it just woke up and broadcasted.  This system seems quite ridiculous to me.  I'm wondering why it is done this way.  Perhaps there is a good reason, but I cannot fathom what it is, so I was hoping to get an informed answer from someone who should hopefully know.  I have two multi sensors from different manufacturers and they both seem to employ this convoluted and unreliable method.

I would recommend longer sleep periods for line powered environmental sensors than 60 seconds, especially if you have a lot of them.  It creates a lot of work for Vera and any clients such as VoxCommando and your mobile apps.  You should not need a temperature update more than once every 10 minutes under "normal" circumstances.  I seem to recall you having issues with delayed events and this would be a likely suspect.  If you have multiple sensors being polled that frequently then you probably will see endless chatter between Vera and her clients.

[AgileHumor]

#3 While the Vera is always on controller, most in the past have been not (i.e. handheld battery powered)...maybe that is why the sequence exists.

[jitterjames]

Interesting.  I didn't know that.  I thought that the handheld battery devices were just remotes, not full controllers that would keep track of the status of all devices the way a Vera does.  That could change a lot of things, but seems like a kind of "weakest link" design if it is true.  They must still have envisioned full fledged controllers back when they were dreaming up the protocols.

[garryjw]

Thanks for all the questions - hopefully I will try and get some answers.