Thanks for pointing this out! Let me connect with @EricM_Inovelli and we’ll get it updated
Also, thanks @dvellek for writing these out, very helpful!
Thanks for pointing this out! Let me connect with @EricM_Inovelli and we’ll get it updated
Also, thanks @dvellek for writing these out, very helpful!
Documentation at https://support.inovelli.com/portal/en/kb/articles/can-i-install-multiple-smart-switches-in-a-3-way-4-way-etc#Multi-Smart_Switch_Limitations still says:
The main drawback is probably that the “slave” (non-load) switch doesn’t show the other switches level, but you can make that less annoying by turning the switch off or to 99% and disabling the relay.
I really want to pull the trigger on these, but I’m waiting for official confirmation in your documentation that I can correctly sync the LEDs.
Looking at your notes, it sounds like you only set up the associations in one direction. You need to create the associations in both directions in order for things to stay in sync.
By the way, there’s no need to use group 2 associations at all – those are for the ‘basic set’ which is something of a legacy type of association at this point. Group 3 and 4 are all that are needed.
Despite the lack of updates to the documentation it definitely works as expected with LZW-31/LZW-31SN. At this point I have four different circuits setup with associations and everything syncs as expected.
I just spent a ton of time today getting this to work with SmartThings. Writeup below.
TL;DR: It works!
78839f6a-0881-41c3-8c3e-cbe2e1a1b927 12:14:05 PM: error error executing uninstalled method on Living Room - Load Association Group 3: java.lang.IllegalArgumentException: Command 'setAssociationGroup' is not supported by device 9bebc330-cf1c-44e0-a1d1-83fb0528f60d of type 'Inovelli Dimmer'. Supported commands: [on, off, configure, setLevel, refresh, resetEnergyMeter, ping]
78839f6a-0881-41c3-8c3e-cbe2e1a1b927 12:14:05 PM: error java.lang.IllegalArgumentException: Command 'setAssociationGroup' is not supported by device 9bebc330-cf1c-44e0-a1d1-83fb0528f60d of type 'Inovelli Dimmer'. Supported commands: [on, off, configure, setLevel, refresh, resetEnergyMeter, ping] @line 72 (initialize)
Device | Setting | Value | Notes |
---|---|---|---|
Living Room - Line | Association Behavior | 11 | Prevents race condition/loop behavior where switches “fight” each other. |
Living Room - Line | Switch Type | 3-way Momentary | May be unnecessary? |
Living Room - Load | Switch Type | 3-way Momentary |
Source | Destination | Group |
---|---|---|
Living Room - Line | Living Room - Load | 3 |
Living Room - Line | Living Room - Load | 4 |
Living Room - Load | Living Room - Line | 3 |
Living Room - Load | Living Room - Line | 4 |
Setting | Value |
---|---|
Dimming Speed | 1 |
Ramp Rate | 0 |
Ramp Rate (From Switch) | 0 |
Disable Physical On/Off Delay | Yes |
Awesome, thanks for testing this! Is there a way you could shoot a quick video? I’d like to link to it in this thread if possible.
In your use-case, have you found it to be reliable?
I just got a 4-way setup working in Hubitat today as well! Copying @xuor’s great writeup here is what I did:
Device | Setting | Value | Notes |
---|---|---|---|
ALL | Association Behavior | 11 | Prevents race condition/loop behavior where switches “fight” each other. |
Capability is Switch / Dimmer / Bulb for all of these.
Source | Destination(s) | Group(s) |
---|---|---|
Basement Hall Light | Basement Hall Light - 4-way middle, Basement Hall Light - 4-way end | 2, 3, 4 |
Basement Hall Light - 4-way middle | Basement Hall Light, Basement Hall Light - 4-way end | 2, 3, 4 |
Basement Hall Light - 4-way end | Basement Hall Light - 4-way middle, Basement Hall Light | 2, 3, 4 |
This ends up with 9 total entries in the associations app. Maybe someday when I get time a fork of this app that can handle this specific case without having to configure N^2 child apps would be handy. I have A LOT of 3-way and 4-way switches in my house and it is a lot of clicking right now.
Setting | Value |
---|---|
Dimming Speed | 1 |
Ramp Rate | 0 |
Ramp Rate (From Switch) | 0 |
Disable Physical On/Off Delay | Yes |
I did NOT set Switch Type
on any of the switches. I don’t really understand what it does in this setup or how it changes the behavior of the switches.
It seems reliable but slow. There is a noticeable 1-2 second delay between pressing on/off on one of the non-load switches and the load switch reacting. Holding to dim is reactive enough to be useful. I’m going to let this setup sit for a few weeks before doing any more N-way replacements to see how the family deals with it.
I do wish that Inovelli had a hard-wired non-load dimmer similar to the Leviton DD00R-DL which appears to use the traveler as some sort of communication line between all the non-load dimmers and the main load dimmer.
Maybe I’m misunderstanding, and they aren’t Inovelli branded, but are the supported Auxs different than what you’re suggesting? They can be configured both load and non-load and use a single traveler for communication.
Haven’t had a chance to take a video yet, but here are my observations on the 3-way setup:
Yes those are an option and I have a set of the GE momentary switches on the way to compare.
However they will look and feel different and I believe they don’t have any sort of sync’d dimmer bar. A lot of my N-Way switches are wired such that the load switch is not in the same place that I want to take advantage of the wonderfully controllable LED strip on the Inovelli switches.
Like if there would be any way to use a LZW31-SN as the hardwired non-load switch such that the LZW31-SNs were wired together with the traveler and used that to sync on/off/dim state while the LED strips were exposed as child devices as the are now.
Yep, the Auxs have no LED bar. You can use two Inovellis but they won’t communicate via traveler; direct association only. There have been mixed reports over the syncing of the LED bar when connected via direct association. I agree that two Inovellis communicating via traveler would be great as it would likely eliminate a lot of issues.
Does “everything:” work as expected? I just set this up recently and it mostly works as expected. But when I use the Dashboard to turn the load switch on/off, I don’t see the aux switch staying in sync with the primary load switch.
In other words, it does NOT work from a Dashboard or Rules/Automations from the smart hub. It works with manual switch toggling, but the LED bars do NOT stay in sync when sending ZWave commands from the hub.
So I’m wondering if you folks who say its all working have verified they all stay in sync when using zwave command and/or the Dashboard to turn the lights on and off?
I tested this pretty thoroughly when I originally posted about it, but just retested to confirm that I have no synchronization issues between the physical LEDs on my dimmers that are associated with each other, no matter how I initiate the dimming / toggling.
It sounds like both of you are may actually be exposing your ‘slave’ dimmers in your dashboard in addition to the ‘master’ which is located in the box with the load. You should only be actually exposing the ‘master’ within your dashboard / google home. As an example, all of my ‘slave’ dimmers have the associated device in Home Assistant set as disabled:
This ensures that:
Additionally, you should ensure that your ‘master’ dimmer is not the one that has parameter 12 set to 11. In other words, you should only be changing parameter 12 to 11 on your ‘slave’ dimmers.
I agree with you that only the ‘master’ should be controlled from any dashboard / alexa / google home. I do not have any ‘slave’ dimmers exposed in my dashboards. Only the ‘master’ appears on my dashboards and only the ‘master’ is exposed to my Alexa skill. Its a good point for others to check, but was not the problem in my case
THAT is the key to making it fully work with z-wave commands to the master dimmer.
Some of the previous comments implied that all associated devices needed to have parameter12 set to 11. But I have learned that the ‘master’ dimmer (the one with the actual load connected) still needs to have parameter 12 set to 15 (15 is default) so it includes z-wave commands. The other (slaves) need to have parameter12 set to 11 so they exclude z-wave commands. Having the ‘master’ parameter12 set to 11 is what disabled the synchronization when z-wave commands were sent to the master.
I don’t completely agree. Its true that the LZW31 dimmers don’t generate ‘BasicSet’ commands from LOCAL button pushes and therefore don’t need Group 2 association to sync LOCAL button events. However, with z-wave commands enabled on the master (parameter12=15) it is possible for hub-based automations to send BasicSet commands to the master to turn it on/off and you need Association Group 2 enabled for those events to keep the on/off state in sync on the slaves when an automation sends a BasicSet command to the master.
Not necessarily! If you want to use a LZW31-SN to control a LZW30-SN (or other on/off switch), you definitely need group 2.
The above 10+ posts should be put in an N-way wiki and summarized.
Great info, thanks everyone!
Can you provide a screen shot of you Hubitat settings for the associations?
Is there an equivalent to parameter 12 for the LZW36?
I finally grabbed a video of the delay I see with the setup I described in reply 24.
The setup has been very reliable but the delay when using the secondary switches is enough to be annoying.
Take a look at: https://photos.app.goo.gl/W1gVqaaF2rJbgXHb6
I don’t have a great way to measure but a rough estimate is between 200 and 700ms of latency when I interact with a secondary switch before the primary reacts. It seems to be longer if the switches have been untouched for a several minutes.
It is enough of a delay that I’m going to swap it out for the GE Enbrighten Add-On Switches, which have no delay but don’t look nearly as nice:
@Eric_Inovelli I would LOVE it if the LZW31-SNs could use the traveler to communicate or if you were able to get an add-on with the LED strip available. We have 7 multi-way lighting circuits and they are some of the most heavily used. I’d love to have the LED indicator on all of them vs whichever junction box happens to have the Load wire in it.