Zigbee Device Connectivity Without Repeaters (During a Power Failure)

The primary reason I bought a Hubitat was to allow it to function even if internet connectivity was lost, as network connections can be flaky in both Ocho Rios and Los Cristianos, and a Samsung hub is mostly useless without internet.

But, prototyping here in the USA, it looks like I must avoid reliance on any zigbee repeater, as while I can battery back up the hub, router, alarm siren, and WaterCop, I cannot expect battery back up for things like Trafardi outlets or other "zigbee repeaters".

But there seems to be no battery-powered zigbee repeaters. One could simply turn off the water at a house when one is not in residence, but there is still a sprinkler system to run in the dry season in the tropics, and the cleaning service needs water, and needs a functioning bathroom to use. But the leak detectors... how to make them work without repeaters? What about external antennas to get better range?

The minimal core functionality of things like intrusion detection and fire/leak detection has to be able to persist, even if both power and internet are down. (Detecting the loss of each is an important virtual sensor.) A limited set of devices can easily be run off a deep-cycle marine battery that can be trickle-charged by a solar panel for even extended outages, but I don't understand how the sensors stay connected and useful, given the need for repeaters in so many configurations. Must I use the old-skool wired sensors, and run wires back to some Iris DWS800s, or other "dry contact" sensors within range of the hub to avoid this problem? Has anyone enhanced the dongle antenna on the Hubitat C4?

These are not large places, but keep reading about so many people solving their problems with a repeater or three.

Can you just turn off all your repeaters and see if the sensors can reach in a no power situation, its ment to be a mesh so direct connection might be a fall back option and if the place is small

Zigbee devices form a mesh network, and even in homes that aren’t very large, repeaters are often necessary because there’s all kinds of reasons why the signal may not propagate well from hub to end device.

Beefing up the hub’s transmit power probably won’t do anything, because the end device still needs to send a signal back to the hub for it to be of any use.

You should take a look at some of the threads created by @iharyadi. He has several home-made zigbee devices of his own creation, some have battery backups that, I think, still can repeat when electrical mains power is lost. That might help in your situation.

You are absolutely correct, which is why “real” home security systems tend to have wired sensors that all go back to an alarm panel with battery backup and cellular connection to send alarm signals to the central monitoring center if the phone lines are down.


That’s true, it’s conceivable one could have a functional zigbee network without relying on repeaters. But there wouldn’t be any good way to test that out until OP sets everything up, and it sounds like he isn’t at the location where he intends to do that.

1 Like

There is at least one supposed battery-capable repeater. The Centralite Nightlight 3420. It is a plug in device with battery backup in case of power failure and states that it keeps the repeater function while on battery.

I have NOT tried out this function though and I will be the first to admit it has not come through on some of it's other claims. Although I am trying to work with them to see if there is firmware to fix those flaws.

which is why “real” home security systems tend to have wired sensors that all go back to an alarm panel with battery backup and cellular connection to send alarm signals to the central monitoring center if the phone lines are down.

Silly me for thinking that any of the home automation products might address this, this most basic reason one would mess about with "door sensors", "window sensors", and "smart locks". :crazy_face:


With a power situation once the repeaters are cut it will also take time for the devices to reroute themselves back to the hub, if they can.

One option may to invest in something like an xbee (theres a dedicated thread in this forum). I found it is a great tool for not only mapping the zigbee mesh to see how things are routing but it was also a very powerful repeater. It would grab 18 of my devices from around the house. Maybe just adding this in the mix with a batter backup will help keep things stable when there is a power outage.

I also recently added a Sylvania LED Light Strip to my network in my bedroom and found out that it was a very powerful repeater. Didn't even know this until one night when it got unplugged and a lot of my devices stopped reporting in. I then mapped out the mesh with my xbee and found that it was grabbing and repeating a whole bunch of devices around my house as well. Throwing a battery on this as well would also help in a power outage.

So if you can plan out the mesh with a couple key devices instead of a bunch of smaller repeaters, then you can probably invest in some battery backups to keep them on for cheaper.


Yup I hear you. But there’s a reason most (all?) home automation platforms put a bunch of fine print in that basically says “never rely on this in an emergency.” :rage:

Edit: while I’m sure opinions might differ re: what makes for a “real” home security system, I’d argue that if it’s not UL-certified for that purpose (or an equivalent independent safety testing lab), then it shouldn’t be relied upon in a police or fire/life safety emergency.

1 Like

Well no... none of the DIY home automation products are going to address this. In fact MOST have specific warnings or statements in their documentation stating exactly NOT to do this or rely on them for this.

Is this possible? Absolutely.
Is it cheap? No.

1 Like

This is fine when one is in the USA, but in many other places, one's best security is the gift of a shotgun to one's neighbor who lives there rear-round. There is no "ADT" service, so one must cobble things together as best one can.

As a parable, I have a Kaba 900 series mechanical pushbutton lock on a vinyl gate on a vinyl fence. One could ignore the robust lock, and simply kick out the vinyl slats on the fencing or gate, but the idea here is to keep kids from wandering into the pool, and a keyed lock would be annoying. The Kaba was the only "real lock" that was both weatherproof, not junk, and keyless. Insurance requires a locked gate for a pool, and this thing is self-locking, each and every time.

Understood. But then you kinda have to accept there are deficiencies that come with a solution that’s not intended to be held up to a gold standard.

I’m not trying to be argumentative, but how does the comparison to the mechanical push-button lock apply? It sounds like it meets the requirements your insurance company has deemed necessary for the purpose of insuring you against the risk of a child accidentally drowning in the pool.

Most homeowner’s insurance companies would not offer a discounted premium if a lock like this is installed. If there were a UL-certified, centrally monitored intrusion and fire alarm system, they probably would offer a discount on the premium.

But that’s because in that case, the homeowner has taken what the insurance company agrees are adequate steps to mitigate another risk they presumably insure against, stolen property and a house gutted by fire. If the system isn’t certified to meet the standards of an independent safety testing lab, then it’s not good enough (for the insurance company, that is).

1 Like

All - will you consider continuing this topic in the Lounge (or Lounge/Debate Chamber) because it has meandered far away from the original question (which fit in the Support category).

I've seen that Hubitat staff have had to step in and close other threads over the last couple days. Perhaps we can make a little less work for them by making better use of community forum categories.

I know @bobbyD does monitor the Support category as well, and i'm sure he has plenty on his hands without our adding to it.

For what its worth, here's the forum FAQ.

The section relevant to my concern is below:
Screen Shot 2020-05-10 at 1.41.22 PM

Thank you for your consideration!


Great idea, I have created a new thread here, assuming anyone still wants to discuss.

1 Like