Z-Wave Network Repair / Routes

Good point on the potential issues if you remove even more Z wave repeaters. Getting the balance right can be tricky.

One small point: The iris plugs were intended by the manufacturer to be both a Z wave repeater and a zigbee repeater: that’s why they have a Z Wave radio in them. :wink:

That said, over the years a number of different community members reported different kinds of issues with the Z wave aspect, and there were apparently multiple versions of the firmware put out, some better than others, and no way to update that at this point if you happened to get a bad one. So I’m not surprised if anyone finds a specific device o& that model to be more reliable for Zigbee. It’s just not the intent of the design.

Agreed on the original design intent but I don’t believe the ones I own have any zwave markings on them. Let’s compare it to the Bluetooth radio in the ST hub :wink:

EDIT: I just checked one in my office and it actually has a Zwave Plus sticker on it. I need to check the others as this might be a more recent one…

1 Like

Not necessarily true though. I’m down to 6 Zwave devices from well over 100, and those remaining devices are scattered at all corners of the house and haven’t moved. My super old Aeon HEM is under the house and is the furthest zwave device from the hub, and it’s connected directly to the hub. It’s literally under the house on one side of the building and the hub is in a room tucked behind a dresser mirror on the exact opposite side of the house. It’s also configured to report instantaneously, so it’s a very chatty device in my Zwave mesh and works great.

image
image

All my other zwave devices are zwave+, and they’re GE fan controllers and one water meter. Given my small number of devices, my repairs take 2 to 3 minutes w/no errors.

I am convinced that the more zwave devices in the mesh, the harder and harder it is for ST to not have repair/mesh issues. I have no idea why or what causes this, but for me I believe I just had too many chatty devices, as well as some old GE devices finally starting to fail. I’ve mentioned this before elsewhere, but I also noticed significant zwave mesh improvements once I got down below 30 devices.

I’ve noticed a greater range than expected on many of my devices as well. So as long as @HalD 's devices are all in acceptable range, then removing won’t hurt. Some devices are able to report back the signal strength so he could use that feature to check whether any particular device might need a repeater.

I concur that 30 to 40 zwave is the limit where you start seeing issues. Last time I had to exclude everything and rebuild, I started cursing the world at around 40.

@HalD - There is a good chance you do NOT need the repeater in the Iris plug… and while not everyone may have had (or known to have) a bad experience with the zwave portion of it, I am pretty sure they only caused me trouble. Included as Zigbee only, they are great so I truly believe a long period with the zwave repeater excluded would be a good test to do. If you really need a repeater, then I would advise an in wall mounted device instead (this way it can never be removed).

2 Likes

LOL, yep I remember those days…

2 Likes

I’m pretty sure my Z-wave mesh is fine.

All I own are Z-wave Plus smart switches and dimmers. I have mostly Zooz but also a couple of Levitons and a couple of GEs. All of them use the stock DTHs and show as local. I only use Smart Lighting and new app Scenes, no WebCore.

All but three of my devices are in one room. The heart of the house is a single open room that incorporates kitchen, dining room, wet bar, and living room. The front entrance, three hallways, and three sliding glass doors open off this room. With ceiling can lights, hanging fixtures, counter lights, lighted cabinets and most of the exterior lights, there are over 30 Decora switches around the perimeter of this one room.

All but four of my Z-wave Plus devices are in this room. One of the four not in this room has to route thru this room to reach the hub.

When I do a Z-wave repair, everything in this room plus the one that routes thru it throws errors, usually “cannot update neighbors” or something like that. There’s a long, rambling topic I started about this last winter.

All my devices work. I just get random 2 to 15 second delays, usually when triggering multiple devices via a scene. Different devices in the same multi-gang box can be switch instantly or have delays.

I do not have any Iris plugs or other Iris devices. At one time I had some Aeotec Z-wave Plus repeaters. I removed those several months ago as they were never repeating anything.

I have multiple devices in my kids’ bedroom that I use for their night routine. As you can imagine, they must be super reliable to avoid causing tantrums. At first the entire control was via zwave and it was an inconsistent nightmare. Stuff would have random delays or simply not work. It felt like everything running perfectly and fast was the exception.

I got so fed up I removed the $90 zooz zwave power strip and replaced it with a bunch of Smartthings ZIGBEE plug in outlets. Night and day… it has been flawless ever since.

The issues were likely not due to the power strip, but rather to the traffic on my zwave network. On several occasions it turned out that chatty devices were wrecking havoc. Lately I was pointed to Zooz 4-in-1 sensors sending a message every 30 seconds, which is too much (per ST).

The point I am trying to make is that you might not have routing issues, but maybe routing delays caused by traffic. I would scour all the devices you have and tone down / turn off any reporting you see. If you must keep the reporting, try to spread it out in frequency.

I had devices reporting Voltage ?!?!?!? frequently. While I can see it helping to find issues with the power grid (brown outs/surges), or faulty connections (bad connection could show as a drop in voltage), I think it is silly to have these devices report 120V over and over. Another one that can be toned down in most cases is Energy reporting. The only reporting that I tend to find more useful if a bit more frequent is Wattage or Amperage. Cutting down on the reporting of all able products can speed up reaction time on your network if traffic is your issue.

Sorry, I confused you with @jlv. He is the one that should try excluding the zwave repeater in the Iris outlet to see if his issues go away… or maybe I am just 100% confused between your issues and his issues now :wink:

Just to see what happened, I did a Z-wave just now.

I have 23 Z-wave devices, not 25. Forgot I had replaced 2 with GE Zigbee.

Opened an IDE tab to the hub events page and another to the hub utilities page. Triggered the repair. Then refreshed the events page every couple of minutes.

Repair ran about 7 minutes. Between 5 minutes and the end I got 11 errors for 10 devices in the big room. One device reported “could not delete old route” and “could not assign new route”, the others only reported “could not assign new route”.

Only 10 devices giving errors is a great improvement from the last time I tried a repair.

I’ll give it another try after the upcoming firmware update.

How do I find chatty devices?

My only energy-reporting switch is a single GE Zigbee. None of my Z-wave report energy. I’ve only got three sensors, all are Zigbee.

I believe the only spot in the IDE where you can see most/all of the traffic is the Events List but othert han looking for specific messages, I don’t find it very useful as is. It would be nice if the IDE could flag chatty devices on it own… @Kianoosh_Karami hint hint :wink:

The only practical method I can think of is to look at the Live Logging, and review each devices. If you see lots of messages per minute for any device, I would consider it a possible culprit. I would expect some variance in messages depending on the activity in the home, time of day, etc.

You can also go see the additional settings for each device to see if anything is on by default which may cause traffic. The default for one of my devices was to report a value every 30 seconds which I was told is too much (at least in my context).

The plot thickens… are you sure you don’t have a misconfigured automation that keeps firing? Anything that would trigger more than once a minute is likely too much. Might work on low traffic networks but is bound to cause trouble as network and activity grow.

My process is similar… I open an Events List tab, then on another tab I run the zwave repair utility and click on the bolded link that appears in place of the utility as that opens a filtered events list. Do NOT refresh this list as it clears it out %#$%#@^%@$#%$# !! It’s ok, as you know, to refresh the list on the other tab. The filtered list will save you from having to hit refresh every few minutes unless you need to kill quarantine time :wink:

Energy reporting devices are the most common offenders of being chatty. IIRC the z-wave spec only recently added that you can’t report more often than every 30 seconds. Zigbee seems to handle frequent messages much better.

3 Likes

I opened live logging and watched it for 10 minutes while don’t morning chores.

Nothing other than cloud-to-cloud stuff. Honeywell TCC thermostats, Harmony Hub, and Alexa.

I have very few automations.

  • SmartThings Zigbee motion sensor in the garage set up in Smart Lighting to turn on Leviton switches for garage interior and exterior lights. This is flawless and instantaneous.
  • Smart Lighting automation at sunset that turns on 5 switches for front exterior accent lighting. These switches are in the big room. I usually hear them click on. Sometimes very close together and sometimes a couple lag behind.
  • Another Smart Lighting automation that turns those exterior lights off at 10pm if we’re home or 1am if we’re away.
  • Not z-wave but I’ve got a Zigbee peanut plug and virtual switch that use power allowance and Smart Lighting to flip the plug on for 10 minutes each hour.
  • Another Zigbee peanut plug runs the aquarium light, turns on at 9am, off at 9pm

I mostly use Scenes for the lighting in the big room and trigger those from the phone, tablet, or via Alexa routines.

There’s no indication in the events log or live log of anything firing repeatedly and nothing set up that should repeat.

I have a theory which I’ve never tested, so it may be worthless, but based on some previous discussions with @tgauchat (RIP) , I believe that while the Z wave protocol itself is fine, the addition of the cloud aspect in a SmartThings implementation means sometimes the system trips over itself in high traffic situations. So that the exact same zwave devices might work just fine with, say, a Homeseer primary controller.

OTOH, as I’ve also said in the past, I personally just don’t think z wave As a protocol is a good match for real time energy reporting devices.

So…I believe completely that Z wave devices on the smartthings platform Might have some fragility in some setups. I just don’t want people to blame the zwave protocol for what might in fact be aspects of the multiprotocol cloud-based architecture that smartthings offers. :sunglasses:

4 Likes

@JDRoberts - I think you are onto something… This is also why I purchased a Hubitat hub but I have not done much with it yet. I have to say that seeing it fire off the few things set up on it while ST was down, was ‘thrilling’ :stuck_out_tongue: How sad…

Speaking of @tgauchat - I have not seen him around in a while. His profile says he last posted in December. I hope the RIP in your post was just to say he abandoned ST forums rather than the traditional meaning.

EDIT: I found the announcement :frowning: Very sad indeed. He will be missed!

1 Like

Poltergeist?

1 Like

Seems the most likely explanation at this point…

:rofl::rofl::rofl:

1 Like

Wasn’t it said that when you see the “cannot assign a new route” error, it was generally a bad routing device or a ghost device?

I’m not seeing any ghosts in the IDE. I’ve had some in the past and removed them successfully.

How can I find a bad routing device?

Maybe the new ability to see the route might help… If you can pinpoint the devices that lag, or have issues, see what intermediate nodes they rely on to see if there is a common one. If there is, you could try to exclude it, run a network repair, and see if things work better. Then you can include it back and monitor. If the issue happens again, exclude it again, repair, and see how it goes… at that point you might consider ditching it.