Zigbee 2-1 Switch (On/Off & Dimmer) | Project New Horizon (Blue Series)

Bringing back up the topic of zigbee binding support:

It looks like it’s been confirmed here that this dimmer will support direct zigbee binding, but I’m curious if it could be confirmed that it will support binding to a zigbee group and not just individual devices.

The use case is for the switch to directly control a large zigbee group of lights (e.g. hue bulbs paired to zigbee2mqtt and added to a group) without having to blast many zigbee commands to each of those lights: that is, it would simply send the single command to say “if you’re part of group ‘ABC’, turn off” etc. Big advantages are, of course, reduced zigbee traffic, reduced latency, and more consistency in the lights being sync’d together i.e. turning on/off/dimming in sync.

Surprisingly (?), as I play around more with zigbee, it seems that a lot a switches simply don’t support binding to groups, but instead need to be bound to each individual bulb. Thought I’d ask the question here!

1 Like

@EricM_Inovelli - Any idea in groups vs individual?

That is a good question.
I am interested in hearing the outcome of this.

Out of curiosity
Is the goal to eliminate a hub?
And do you have experience with a hub to compare to?
(as in, will a hub accomplish this and then the switch is just an input device)
And how much traffic is needed for this to become a problem?

My lighting goal sort of boils down to two primary constraints:

  1. Have RGB smart bulbs in every socket (mostly for circadian lighting, but also the kids have a lot of fun with it).
  2. Don’t be dependent on a hub for basic on/off/dim from the wall switch: switches should ALWAYS “just work” (and look and feel) like you’d expect a hardwired switch on a dumb light to.

I definitely don’t want to eliminate a hub: I have an incredible amount of legitimately life improving automation occurring on it- lights only playing a small part. However, I don’t want a hub to be a dependency in the basics of my lighting: it should complement and add-on to the basics (via scene commands, lights being triggers for other automations, etc).

In my experience with hubs [being a dependency between a switch and a bulb], I have never gotten the lag down to a point that it’s natural and feels like a hardwired switch. Not a huge, huge issue, but if it can be solved via direct binding without downsides, why not? And, to that point, I consider bulbs turning on/off at slightly different times to be a downside (which is, to my understanding, more likely to happen with you have to bind each bulb individually vs binding a group). I certainly notice this sync issue on my Inovelli Red dimmers that are directly associated to multiple Inovelli RGB bulbs (zwave has no concept of associating to “groups” as far as I know(?) ).

Now, to your last question: I have no idea! I’m currently playing around with an Aurora dimmer bound to multiple individual hue lights and it’s pretty impressive in terms of controlling them in sync despite having to make individual calls to each of them (if I understand how all of this works correctly :slight_smile: ). So, perhaps, binding to groups vs individual bulbs is somewhat unimportant? Or, maybe it will become an issue once I move all my hue bulbs over to the network and traffic grows- especially for whole house/whole floor on/off commands.

1 Like

I have to say I admire the work you are doing to figure this out.

My eventual goal is similar to yours only with adjustable colour temperature bulbs and strips. Currently I only have 2 smart bulbs and the rest are ordinary led bulbs. I have slowed any progress down due to Matter. I decided I didn’t want to put more hardware in knowing about the impending release of a standard backed by so many major players. I have to say though I will be getting a blue series switch, because well I want to play, it meets all my hardware requirements with respect to design, and let’s face it, it’s beautiful.

Do you have any where you are documenting your journey? I would love to check in from time to time.

Yep- similar thinking here with regard to Matter + Blue series. I’m already sitting on enough hue for (mostly) my entire house so sort of stuck with zigbee there.

I think these posts are the most I’ve “documented” it :slight_smile: I’ll probably be sticking around these forums as I get my hands on the Blue series and inevitably am troubleshooting and sharing experiences!

Not quite accurate. Absent noise on the line, 50/60Hz AC is pretty close to a perfect sine wave. There are no pulses whatsoever. Dimmers that chop the waveform create sharp edges (like in a square wave) that can be hard for devices with fast response times to deal with. @kevink summed it up nicely: incandescent bulbs average whatever is coming in. Thermal time scales are MUCH longer than those for the diodes and transistors in modern devices. Modern LED flicker is a side effect of progress, I guess…
image

Quick update regarding the outstanding questions:

  1. On/Off vs Dimmer functionality – is this done via firmware or hardware
  2. Group vs Individual ZigBee Binding

Regarding #1 – We’re still working on this with the manufacturer. We do have a preliminary response, but I’m not sure I fully understand, so I’m asking for clarification before explaining it here. There is a difference inside the switch on how the switch handles a dumb switch vs an aux switch, which leads me to believe there are different ways the switch handles loads. Similar to the dimmer switch, when a dumb switch is used, you hear a, “click” which cuts power completely to the circuit. This is why you cannot use a dumb switch in a non-neutral setting as no power, “leaks through” to power the switches internal circuitry.

My gut is telling me that this new switch is basically our dimmer switch, as it now makes sense how they were able to solve the non-neutral setup in on/off mode. However, what I want to explore is if we can utilize that other relay (used during the dumb switch mode) to have a, “hard on/off” vs “simulated on/off”.

EDIT: For full transparency, I’m adding the manufacturer’s response. Someone way smarter than me can possibly interpret:

“It’s a mos tube instead of the relays because it’s a 2-in-1, the relay inside will only be used under a dumb switch setting. On/off and dimming are using the same hardware (mos), when on/off there will only be 0 and 100, when it is set to off (0), the electrical conduct of the p-n junction in the mos tube will be cut off (but it is differ from the concept of a physical cut off of a relay).”

Regarding #2 – This is something being tested as we speak. Right now we haven’t gotten groups to work, but honestly, this wasn’t on our radar as we’re new to ZigBee and didn’t know this was a thing, so thank you for bringing it up! Individual binding is working great, however.

2 Likes

“Not quite accurate” back at ya :stuck_out_tongue:

Being a sine wave does not mean its not pulsing. Its still a wave that goes up and down and crosses 0 and reverses direction. It absolutely is pulsing 60 times a second. Being a sine wave doesn’t change that fact. Its going from 0-170v-0-170v which is a pulse and has a width of 1/60 a second. It doesn’t have to be a square wave to be PWM. Curved waves pulse too.

Granted, full sine wave AC is smoother than the chopped wave from triacs and mosfets. And I agree the chopped wave can cause some issues (mostly noise) with different types of load. But it doesn’t mean its not pulsing. All AC power is pulsing and a relay does not change that.

1 Like

Thanks for these quick responses, @Eric_Inovelli and team!

Quick follow up question re: #2 binding: Are you aware of any limit on the number of bulbs that the individual binding will work with on a single dimmer. e.g. can I bind 10 bulbs to one dimmer or is there a limit like the 5 bulb zwave association limit with the red series dimmers?

This would really only be important if the group binding ends up not working out. I’ve noticed that some dimmers seem to have hard bind limits (e.g. 7 devices for the Lutron Aurora dimmer), but others seem to allow many more (e.g. GE Zigbee in wall dimmer). Of course, I’m assuming these limitations are with the dimmer hardware/firmware itself and not my controller or controller software (Sonoff Zigbee 3.0 and zigbee2mqtt in my case), but I’m new to “advanced” zigbee as well so learning as I go. I think I have a max of 6 bulbs I’d want to bind to a single dimmer, but would be great to get confirmation if possible!

This is a very old comment that no one ever replied to, but could you potentially use a faceplate screw as a heat sink for the thermal probe? You’d need to somehow thermally isolate the probe and screw from the metal heat sink tabs (maybe a plastic ring between the screw and housing?), and thermally couple the faceplate screw to the probe (maybe mount it such that you can actually screw directly into the probe?), but that could be a pretty elegant and invisible solution.

I love having extra temperature sensors all over the house, so this would be awesome for me, and maybe even awesome enough to get me to upgrade my existing switches.

I’m contributing nothing to this conversation but…

Wow. It’s amazing the level of depth & transparency @Eric_Inovelli has provided in this thread. I had been holding out for a restock of the Reds to complete a number of outstading projects in my house (10+ switches needed), but no longer…I’m now in the long line for the Blue Switches.

I could have picked up any ol’ zwave switch instead of waiting for the Reds to come in stock. I’m really happy I didnt.

Project Update: Received some clarity around the relay question (I’ll explain below) and we’ve closed any new firmware requests so that we can submit for ZigBee Certification (or CSA, I’m really not sure what it is). New ideas will be captured and put into a firmware upgrade in the future and the only thing we’re working on for the production run now is squashing any remaining bugs.

Targeting early May for a release – hopefully pre-orders can open in April once we have an official production scheduled. I’m going to be conservative with pre-orders this time!


Relay Question/Comments
I want to first thank @contactcr for bringing this up and for the rest of you who piled on and asked for clarity. This is exactly what the community is for and I want to share that because of you all bringing this up, we were able to make a last minute change as there was a misunderstanding (language barrier – no one’s fault – if anything, mine for not being super clear).

TLDR: We were able to add the relay function (as the hardware already supported it in 3-Way (Dumb Switch) mode) so that there’s more of a, “hard on/off” for single-pole settings. We were unable to put it in 3-Way (Aux) mode because of the way the smart switch detects signals from the traveler (and the aux and dumb mode detection is different). So, while the normal MOS + relay scenario we implemented works great and we haven’t had any complaints around flickering or safety concerns in on/off mode, at least there will be an override if someone experiences this.

Long Version:
NOTE: I am not an engineer so if I screw some technical terms up, I apologize – I understand the concepts of what’s happening, but I will likely mess up some of the vernacular.

When we first briefed the team on our requirements, we said that we wanted this switch to work in both a neutral and non-neutral setting (like our Red Series Dimmer) but also, if they could figure out how to get it to work in on/off mode, that would be a bonus.

Looking back, I can now see where the language barrier came into play. We told them that our current Red Series Dimmer had two relays in it as when it was put in 3-Way (Dumb) mode, you could hear a, “click” just like you hear in the Red Series On/Off. Keep in mind, none of us are engineers and we rely heavily on the expertise of our manufacturers (for better or worse – I know there are pro/cons for hiring internal vs outsourcing – not here to debate that lol). So, we figured they could just take our Dimmer switch and use the one relay (MOS) for the dimming functions and the other relay (the one that clicks) for the On/Off function.

When we received the switch, I remember thinking how there wasn’t a clicking for the on/off, but didn’t think much of it because the on/off worked great and even worked on my extremely cheap LED can lights (I live in a rental and the builder is notoriously cheap).

It wasn’t until @contactcr brought it up that I remembered not hearing the audible, “click” and I went into full panic mode lol.

Luckily, when testing the 3-Way (Dumb), I heard the, “click” and at least knew the relay was in there.

__

What we ultimately decided was that the current implementation of the switch will work just fine and has worked fine across 20+ beta testers and in internal testing (both here and at the manufacturer). But, in the rare case someone wanted the, “click” relay (or traditional switch relay), there should be a way to access it outside of 3-Way (Dumb) mode.

We were able to configure the firmware to give single-pole the ability to have the traditional relay. Unfortunately, this will not work in an auxiliary setting as that would require a major change to how things were designed (the way the smart switch detects the travelers in an aux vs dumb switch setting would ultimately confuse the switch).

So, we have the following options now:

  • Single Pole - On/Off (MOS) - Works in Neutral or Non-neutral
  • Single Pole - On/Off (MOS + Relay) - Only works in Neutral
  • Single Pole - Dimmer (MOS) - Works in Neutral or Non-neutral
  • Single Pole - Dimmer (MOS + Relay) - Works in Neutral only
  • 3-Way (Dumb) - On/Off (MOS + Relay) - Works in Neutral only
  • 3-Way (Dumb) - Dimmer (MOS + Relay) - Works in Neutral only
  • 3-Way (Aux) - On/Off (MOS) - Works in Neutral or Non-neutral
  • 3-Way (Aux) - Dimmer (MOS) - Works in Neutral or Non-neutral

Hope this helps clarify things and again, thank you so much for bringing this up @contactcr and for the rest of you guys for building on it. As mentioned, this is exactly why we have the community and the reason why I always say 1,000 project managers are always better than 1!

9 Likes

@Eric_Inovelli thanks for the detailed reply (and just in time it appears). Sounds like a really versatile switch in all respects!

Thanks, @Eric_Inovelli. My guess is that you will get Zigbee certification from the CSA. :slight_smile:

FYI for all, Matter has received another couple of months delay. Although pertinent to the Blue Switches future thread/matter support from Inovelli, 0.9 of the matter spec should become available to you for testing in June, which would allow several months of testing before it is a full go this fall.

Connectivity Standards Alliance Matter Update March 17, 2022

Matter smart home standard delayed until fall 2022

4 Likes

Just saw the PR on the ZHA_Quirks GH repo for the switch. Pretty cool to see a company out in front with that, all the manufacturer attributes right there too nice job!

4 Likes

Can anybody explain in layman’s terms the significance of one relay vs two. Been following this thread for a long time but it got super technical the last few post. :rofl:

One is a relay in the traditional sense (mechanical, ie why you hear a click). The other is a MOSFET, used in a similar fashion as in a SSR (solid state relay). The mechanical relays are better for highly inductive loads like fans/motors, whereas the MOSFET allows for on/off for mostly resistive loads and also allows for dimming through either leading or trailing edge control of the AC waveform.

4 Likes

Thanks for the explanation. So basically this switch is the best of both worlds. Can’t wait til it’s released. I need like a 100 of them!

5 Likes

I am going to have to keep the context of the sentence in mind. It wasn’t till you said this that I realized @Eric_Inovelli wasn’t talking about the Canadian Standards Association.

1 Like