I came across the Inovelli Red Series Dimmer switches and they immediately caught my attention due to the advertised ability to work the smart bulbs by disabling the internal relay. This would be a better option than bypassing the switch in the gang box and replacing it with an upgraded and highly functional device. The only other device I was considering was one from HomeSeer at around a 20% price premium but it did not have the relay disable. I was glad I hesitated and bought these instead.
That said, the devil is in the details and after wrestling with installing 3 of these devices over the last week or so (mostly due to confirming the switches functionality and then have to do some workarounds in my automation platform) , I wanted to report some findings.
First my environment…
OpenHAB 2.5.1 running on a Beaglebone Black with an Aeotec Z-Stick (USB Z-Wave controller). According to OpenHAB’s HABmin client all the dimmers are running firmware 1.35
1 Dimmer connected in a 1-way deployment on a former 3-way (fixture location)
1 Dimmer connected in a 1-way deployment on a former 3-way (remote location)
1 Dimmer connected in a 1-way deployment
A quick word on my “former 3-way” deployments. I’m in a very old house and switch placement was not well thought out. As part of this project it occurred to me that the traditional physical concept of 3-way controls is made obsolete when you have switches that are scene controllers are the automation platform is highly programmable. So, not all switch boxes needed switches and I’m much happier with the reduced clutter and a more logical layout. 2 of the switches primarily control TP-Link Kasa RGBW (KL130) bulbs and the 3rd controls an Inovelli ilumin RGBW bulb.
Observations
- Disable Relay Function - I simply could not get this to work. Before I added the switches all the bulbs worked fine via software controls. Once deployed, 2 of the bulbs (the 1-way and the former 3-way at the fixed location) would go offline unless the dimmer was all the way on. Testing all 3 switches, I noticed that the load terminal and line terminal were NOT showing the same voltage relative to the neutral (about 120V here in the US). I tested with the dimmers off, 50% and 100% (obviously this would show the full voltage though). Unless the dimmer was at 100%, the voltage was less than 120V. This was the case whether the relay was enabled or disabled. With the relay disabled, I would expect that the load terminal to show the full line voltage relative to the neutral and the voltage from the load to the line to around 0 if they were electrically connected. I might be missing something but I don’t see how else a smart bulb could stay powered if this was not the case.
- Dimmer Channel only reports when relay is enabled - Another thing I noticed was that the dimmer channel (i.e. pressing the rocker up or down to trigger scene 2.2 or 1.2 respectively) would not report data unless the relay was enabled. I experimented with this on the 1 dimmer that was in a previously 3-way remote box. Since there was no fixture (thus no load) at this location. The dimmer was just acting as a scene controller. The problem with this is that when the relay is disabled you have to do a bit of “nasty” programming to detect the scene and then do an event timing looping to capture how much you have moved up or down. You then send that value to the bulb’s brightness channel and the switch dimmer channel. It is really not the best experience. It took me 2.5 days to get the programming right and then experimentally tweaking the event loop parameters to get something that was passable for the average person. In my opinion, this channel should always send that data so that it can be read directly and processed that way the dimmer function is smooth and performs the way people expect.
- Group Associations - At one point, before playing with the dimmer controls, I associated the ilumin bulb to the dimmer. That worked well but I as I recall, you needed the relay enabled. Another issue here is that when I had the group association on and the switch was off, there by disconnecting the bulb from the network. The switch also became unstable. Other scene events would process but they would do so unreliably
- “Air Gap” - This is another thing I’m not understanding in the context of deploying with the relay disabled. When you actuate this function the load is disabled and you will also see the switch go dead but the line and neutral will still show a voltage across them. What I was expecting was complete isolation- e.g. disconnect the line from everything so that in either relay mode, no current flows.
- Health Check - The Z-Wave signal test can not be performed unless the relay is enabled
Current solutions
- Wiring - because of the disable relay issues what I ended up doing is connecting the line and load both to the line terminal. Thankfully there are two ports at every terminal so that was no big deal. I just do not like not having a “real” switch in the box or having to go to the breaker panel for a physical disconnect.
- Event loop for dimmer scenes - as I mentioned I’m not crazy about this and based on my wiring now I could just leave the relay on and do things the easy way but this work was a labor of love and if there are possible firmware fixes coming, I don’t want to have to switch it again.
- Other scene controls - I’ve been making heavy use of the non-dimming scene controls. One switch has all 11 in use and the other two have only 2 or 3 openings. These devices easily solved the issue having more hardware based controls so that I nor anyone else in the house has to always pick up their mobile devices or use the computer to control the fixtures.
All in, I do not regret these purchases. If I had not already put in Lutron Sunnata dimmers in the places I didn’t want or need smart control, I would have purchased these anyway and just not have connected them to the network. I like the look and feel of these switches and at night they have the same vertical LED that the Lutron’s have (only difference is the color). I did however want to report my experiences to add them in the mix in hopes that a firmware upgrade will address my observations.
