Understanding the impact of zigbee repeaters

I am wondering if folks have ideas and advice on how to understand the effects of moving zigbee devices around on the reliability of the system, and how to do it properly.

Case in point, I bought a Cree zigbee “connected” light bulb (60W equivalent) to use as a repeater (this being one of the least expensive repeaters). I paired it while I was standing close to a Sylvania Smart plug, with the hub pretty far away (about 80 ft and 3 walls). Pairing went fine and the bulb was fully functional, I turned it on and off about 10 times with no problems.

Then I disconnected the Sylvania plug, and the bulb was no longer reachable. This was surprising because a nearby Smartthings Multipurpose Sensor could still reach the hub, even though it’s powered by a small battery. I waited only a few minutes before I gave up and inserted the Sylvania plug again, after which the bulb resumed working.

This raises a couple of questions. 1. Is the multipupose sensor using the plug as a repeater, or not? I moved the plug after I noticed it wasn’t reliable (I just sent a separate posting about this) but maybe it’'s not really needed? 2. Was the Cree bulb not working because it was out of range, or because it was still trying to communicate with the Sylvania plug? I could have checked by moving the bulb close to the hub, but even that would not have been conclusive—what if it started working again, but not because it was closer, but because enough time had elapsed and the network had reconfigured itself?

I am developing the opinion that it is idiotic to think that this technology is ready for plug-and-play. It’s probably usable, but only with serious diagnostic tools, which aren’t easily available. I have received the advice to go through a “zigbee heal” upon reconfiguration, but that’s not really practical (unplug the hub, remove the batteries, wait 15 minutes). What else can one do?

Zigbee heals can take hours, not minutes. Worst case scenario, reset the device and re-add it

1 Like

Its mentioned somewhere in the forums (and I’ve done it myself), the way to force a full zigbee heal was to unplug the hub and remove batteries, wait 20-30 minutes and then plug it back in again. The devices will start finding new paths once its back up.

I think by removing the plug you were doing this to any devices connected to that plug but probably didn’t wait long enough.

Why isn’t it “practical”? That’s exactly what a field technician would do in order to improve network reliability. It’s just that you might not see the full results until the next day, it can take a while for each device to rebuild its routing information.

As far as the bulb and the plug, unfortunately, as you noted, SmartThings doesn’t give us any mapping utilities. It’s never been clear to me why, as most of their competition does. But they don’t.

There could be many reasons why you saw the outcome that you did. The sensor might be using an entirely different device as a parent, keeping that path open. And that parent might not have any additional child slots available, meaning it would not be available to the bulb to use.

The one thing you do need to be aware of is that with mesh networks, you don’t get instant reconfigurations the way you do with something like a Wi-Fi network. You have to have some patience and give time for the individual devices to do their own routing algorithms. So when you are physically moving devices around, you lay them out the best you can, but, again, you may not see the full results until the next day.

Thanks so much for the replies! I am sorry if I sound negative, but here are the facts as I see them.

  1. If things worked out of the box, everything would be fine. But they don’t.

  2. When things don’t work, one can either give up, or figure out why. There are two ways of doing this: tools and experiments.

  3. There are no tools. OK, that’s not true. I have seen the XStick:


This is likely to be helpful (zigbee only though, but there’s probably other stuff for z-wave) but it is going to add quite a bit to the learning curve.

  1. A 30-minute delay when making one change and experimenting with its effect is just not going to work for me, never mind the multiple-hour wait for the network reconfiguration. There are way too many variables (distance, state of device, state of hub, etc.)

I am sure that all these problems can be overcome with a lot of time and patience, but I also have a job. So it’s a bit disappointing.

If you lay out a good network backbone and stick to certified devices, you shouldn’t have to do that much troubleshooting.

Have you seen the wireless range FAQ?

Start with post 11, read that, then go back up to the top of the thread and read through.

If you follow those principles, you shouldn’t have to be fiddling with individual devices unless something really strange is happening. :sunglasses:

I forgot to thank you for this reply! Very useful, thanks.

1 Like