Zigbee + Matter/Thread Smart Smoke Alarm & Carbon Monoxide Detector | Project Nebula

Innovation & Firmware Discussion Innovation Projects
21

One thing I’d really love to see is backward compatibility with existing hardwired interconnected smoke/CO alarms, especially common brands like Kidde and First Alert.

Ideally, homeowners could replace just one existing detector with a Project Nebula, while keeping all their other detectors.

If any existing detector goes off, Nebula should detect the alarm and report it to the smart home system. Likewise, if Nebula detects smoke or CO, it should trigger the other interconnected alarms.

This would make Nebula much more attractive as an upgrade. Instead of spending hundreds of dollars replacing every detector in the house, users could start with just one and gradually upgrade over time.

I understand that compatibility may depend on specific models and safety certifications, but even supporting a few popular hardwired systems would be a huge selling point.

Being able to add smart functionality to an existing interconnected alarm system without replacing everything would be fantastic!

5 Likes
22

I second everything Rohan has said.
I would be very interested in this product…

Adding to the mantra but I think it goes without saying:
To me this has to be a next-level of devices out of Inovelli. Being a life/safety device it has to be bulletproof as a standalone device, and then all the smarts/intelligence and smart-home aspects come as icing on the cake that are value added features.

But the smart-home features can never interfere or detract from the core safety functions.
-Jonathan

1 Like
23

I’ve used the Nest Protects pretty extensively, and now have some Kidde smart ones at my new place…they’re nowhere near as good, but they’re ok. Excited to see that you guys are taking a stab at this!

I’d start by identifying the necessary non-smart features for compliance, which will differ depending on where you go, but I think Nest was pretty close to being compliant throughout North America:

  1. The new construction market. These would need to be mains powered, and need to communicate over wire. They still need a backup battery, could be 10-yr sealed or not.
  2. You have the retrofit market. They’ll want/need battery-powered units. These should communicate with each other (ideally not Zigbee or Zwave), and also the wired units in some way.

Here in Ontario (Canada) we have a few other requirements for new construction. I think we have some of the most strict requirements:

  1. They need voice announcements (which I think is preferable anyway)
  2. They need a super bright strobe light (this requirement is a bit silly). The Nest Protects didn’t comply here, as the path light wasn’t considered bright enough.

From a smart perspective I think the needs are pretty straightforward:

  1. A way to silence. Nest was great here, they sent out a vocal warning first and you could silence before they started going off
  2. Battery status.
  3. Control of the light, if you do something like the path light. A lot of people love that feature.

You can look at the Ubiquiti smoke detector launch to see the vitriol they had in ignoring a lot of these basics. I’m actually really surprised we haven’t seen a Nest equivalent yet…they really are a solid product and it’s too bad Google bought them and killed them off.

24

This is a really cool product idea! My thoughts:

  • Fewer false alarms are the killer feature that would get me to buy. Actual fires are thankfully quite rare, but false alarms are frequent and tedious. In my house, we removed two of them because the gas oven would trigger them. A smart SD that intelligently adjusts sensitivity when the alarm is disabled, and/or can use the signals from multiple SDs, the thermostat, cameras, etc to generate a more accurate alarm would be amazing.
  • Presence sensing isn’t interesting to me. I want devices that do one thing very well, rather than several things only halfway. I won’t pay extra to have it on a smoke detector as the places I really want presence detection are not the same places that the SDs are installed.
  • I don’t want to interact with my SDs. I just need them to work if there is an actual fire. I need to be alerted if one breaks or is compromised in some way.
  • Be able to trigger the alarm from a security system. SDs are loud, and loud is good if you’re trying to scare away an intruder.
  • Disabling the alarm should be brain-dead easy. With my existing alarms, the noise is deafening and there are two buttons and you don’t know which one you’re supposed to press and then you have to hold it for 5 seconds and you have to reach the ceiling and oh wait, it’s a different alarm in a different room that is actually triggering the alarm so you have to press the buttons on that one. It would be amazing is you come up with something better. Maybe the button is an LED so you can see it better when you trying to press it. Also make the button concave so you can use a broomstick to push it!
  • A reasonably bright light that comes on if it’s dark and the alarm triggers. This would be helpful to get out of the house or find the problem and fix it. Possibly the SD could be configured as a night light for other times.
  • There should be little or no lights on when there is no alarm. Too many appliances and devices are glowing when you just want to ignore them.
  • Matter over Thread is the way. Thread Router (Full Thread Device - FTD) when running normally on mains power, but switch to a Sleepy End Device (SED) role on the battery backup.
1 Like
25

Hardwired feels like a must to me, whether for compliance or not - ensuring that all devices alert when one detects and confirms fire is important.

Presence sensing has no value to me on this one. I’ve already got that covered with the dimmers, and anywhere the dimmer can’t cover, I’ve already got something else - I guess fall detection for a ceiling mount, but not a concern for me.

I’d instead like it to additionally have top notch air quality measurements. Something like the SEN69C would cover just about everything, and having room-specific air quality measurements could be hugely beneficial, while reducing the number of devices I’d have to deploy per room.

26

I love this idea! I’ve been on the hunt for matter over thread detector and there just isn’t one worth mentioning. The first alert that supposedly is a nest replacement is junk that has toms of false alarms, and mandatory junk app.

To echo comments above, wired interconnect is mandatory for me. A ‘lifetime’ battery is a huge bonus. Also, no matter how good they are, sensors and batteries have a finite life span, so as someone above suggested, making just part of it a replaceable module that carries just the sensors and battery would be fantastic.

One thing I’d love to see is configurable CM levels, maybe just a few presets, I might like minimum sensitivity near kitchen, highest in bedrooms, etc. it would also be nice to have a live values to display on a dashboard and/or trend line of some kind.

Another nice product might be a standalone CM sensor in a wall outlet, or even just a plugin one. I’m thinking if cooktop burners are on, and cm level rises, an automation can kick the vent hood to a higher setting.

27

Thanks for all the feedback, everyone! Figured I’d jump back in and go through what we already have planned and what I’m taking back to the team.

It’s still early in development, so there’s a lot we need to work through with the manufacturer. Some of this will also come down to what we can get through testing and certification.

A good chunk of what you’ve brought up is already in the requirements:

  • Smoke/CO detection and the alarms need to work without a hub, internet, or cloud connection. If your smart home goes down, these still need to do their job.
  • The current plan is hardwired power with a sealed 10-year backup battery. Wired interconnect is part of the intended design too. That message came through loud and clear here!
  • These will be combination smoke and CO alarms, with spoken alerts so you know which condition you’re dealing with.
  • Room assignment and reporting which alarm triggered are planned. We also want room-specific voice announcements, but still need to work out how that information gets shared between alarms.
  • The lower face is intended to be one large Test/Silence button.
  • A feature similar to pathlight is planned, along with brightness, timeout, and light level settings. We want it to be useful at night without lighting up your room.
  • We’re looking at ways to aim the mmWave sensor after installation, plus settings for sensitivity, range, and timeout. I get the placement concerns. Having a smoke alarm in the right spot doesn’t automatically mean it’s a good spot for presence sensing.
  • Zigbee and Matter/Thread are the priorities for launch.
  • Local and remote self-tests, test results, battery/power status, and fault reporting are in the plan.

Here are some notes from this thread that I plan on bringing back to the team to review:

Remote silencing is a big one. Being able to use HA, an app, or a switch button would be really helpful for tall ceilings. We need to find out what we can get approved before promising anything there.

I’d like us to look at earlier battery warnings and whether there’s any flexibility around when maintenance alerts happen. Same goes for automatic self tests and exposing a countdown so you have time to warn everyone or get the dog outside.

Compatibility with existing wired alarms is another big thing. I can see why replacing everything at once would be a tough sell. We’ll need to figure out which combinations we could actually test and support.

Also replaceable sensor modules, maybe. I don’t know yet whether a replaceable module is practical, since laws require detectors to be replaced after 10 years.

For the extra sensors, temperature is already being considered as an option. Humidity, air quality, combustible gas, and the other suggestions aren’t currently in the requirements, but I’ll bring them.

Appreciate everyone taking the time to explain how you’d actually use these, including the things that would keep you from buying them. I can’t guarantee every suggestion will make it into the product. Cost, space, engineering, and certification will all play a part in what we can do, especially for the first version. But I’ll bring these ideas to the team and keep you posted on what looks doable and where we run into limits.

Thanks again!

1 Like
28

Maybe able to get around this with semantics. The unit is a smart mount for the detector, the detector, a separate internal component, can be replaced after 10 years. Just a thought.

1 Like