I've been a lurker for several months and have tried dealing with these issues on my own, but I am about ready to give up and could use some help. I have contacted support a couple of times for assistance, but I basically got directed to the forums for assistance.
What I am seeing is devices will become unresponsive for 1-2 minutes on occasion when multiple devices are sent commands to turn on/off. I also see some devices not reliably turning on/off on their schedules. I have stopped using the built in scene control app because I have found that to be unreliable with the issues I have been experiencing. When the issues occurs and devices with a direct association on zwave can still control the associated devices - I only have two such devices (for whatever that is worth). I have no ghost devices and the only thing I can observe with the buit in zwave table or the great Hubitat Z-Wave Mesh Details app is that there are a lot of device reroutes going on. Some devices I see re-routing occurring about 50-60% of the time. I assume what causes the issue that when mutlple devices are sent commands when more than 1-2 need reroutes the whole system grinds to a halt.
What I have tried:
Bought a second C7 Hubitat and re-setup ALL zwave devices on the new hub and relegating the old hub to Zigbee only (which I only have a few vents).
Updated all firmware on the zen switches - removed and re-added to network.
identified a few devices with suspiciously high neighbor numbers and removed and re-added those. Seemed to improve occasionally.
Always updated to the latest Hubitat firmware in hopes that would improve the issue - no dice.
Replaced all non zwave plus devices with zwave plus. I now have 120 zwave plus devices that consist mostly of GE/Jasco, but also has a fair number of Zen 72/Zen77 switches.
Replaced all my wall plates (which were metal) with plastic to help the mesh. No improvement.
Have tried both built in drivers and user drivers for the Zen switches and the GE.
BTW - I use a button program for a zen32 scene controller in my master bedroom as a test case as it seems to be the easiest to produce the issue. I used to have issues with my evening alarm arming turn off all lights, but found an app on the forums (All Off) that seemed to address that for the most part. Works great.
Is it normal to have so many reroutes and its just an issue with such a large zwave network? I was previously using a Smartthings hub and didn't have issues there, but didn't really ever go looking behind the scenes at all the stats because I never had delays like this. I'm out of ideas at this point and could use some help. As is customary, I am attaching my zwave settings screen (what a PITA screen capturing that was ).
Lots of disagreement on this topic. I think the issue is more about the messages/s, burst # of messages, and # of hops everything has to take than it is just the # of devices. I have 105 devices right now, and a dozen or so more I need to add.
Completely stable / solid. BUT I also don't have any power metering devices, or other devices that send a ton of, or high frequency, data on the mesh - just lots of switches, dimmers, motion sensors.
That device is not well connected. There's something 'blocking' the RF. Be it a metal AC vent, mirror, distance... that device is simply not well connected. It's reduced its speed to 9600 to try to get communications through.
Going to the Device Info page for that device and forcing the contact sensor to change, you might expect. there to be a 1-2 second delay. It's likely the device is sending and resending at the lowest possible speed, That means it alone is consuming 10x packet frames of your higher speed devices... the radios are busy trying to hear this one device for 10x the time it would take to converse with a 100k speed device. And with retry's it's probable 30x frames that have to wait.
For people with small sets of devices, the chance that there's a lot of competition for the RF bandwidth is lower than for people with a high quantity. (Yea, I know.. that's a duh, but sometimes it needs to be said )
You have a lot of devices and that's OK, but you will need to take the time to care for the mesh. You might consider going through the device list and moving any of those slow ones back to your C5 and once the C-7 is nothing but sub-20ms RTTs and RSSIs that are positive numbers, see if the responsiveness is significant'y altered.
A couple of clarifying things - I almost exclusively have switches/dimmers, plus 3 locks, 2 contact sensors, 1 water sensor, and a water valve controller (Zen titan). I have one polling device, but turned its reporting to every 12 hours because I dont need it.
Also, I only have 4 that have dropped to 9.6kbps. There are 12 at 40 kbps, and the rest (104) are at 100 kbps. The sensor you keyed in on usually is at 40 kbps and has a good responsiveness. It must have had a re-route because the standard deviation on the average RTT is 1828.42 ms. Also, that is a new zooz contact sensor and I think it only reports every 12 hours by default (other than when it opens/closes - which it has been doing just once a day). I cant imaging that using a lot of bandwidth.
Also, my front porch light is at 100 kbps and has 6 route changes with a direct connection currently (its about 30 feet from the hub with 1 wall in between). There are several others that are at 100kbps with high route changes, which I see as part of the problem. I am not trying to be argumentative, just trying to understand what is going on.
The negative RSSI is what I find really confusing because it is surrounded by other nodes (powered switches) and a lot of times the paths seem to fall back to direct at negative RSSI - even after it connects through other nodes and has appositive RSSI. At least that is what I have observed anecdotally.
Oh - and that screen shot was actually 12 screen shots of the browser window glued together. Next time I'll break it apart in a few sections.
I find this whole Zwave vs Zigbee rather comical. I have both and have had issues with both. Truth be told once I got the mesh nailed down I have more issues with the zigbee devices acting up (or most recently not acting at all). Everyone's use case and experiences are different. Both protocols have the place. Their strengths and weaknesses. Anyone in this community will have an anecdotal reason for and against both. Good luck with where you are headed next!
Awesome question! I tried to look for that commonality, but haven found anything. They are in different parts of the house and dont seem to use the same repeaters (at least its hard to tell). I mentioned that I found a couple of nodes (switches) that had a suspiciously high neighbor count before. One of them was on the edge of the network, but had 101 neighbors. I removed it and repaired it and it my issues seemed to improve for a bit. When I repaired it it dropped to about 45 neighbors. I thought I was on to something there and started to look at regions of the house and compare devices to see if some had suspiciously high neighbor counts. My thought was that they were telling other nodes they had a good connection and messing up the network. I thought I squashed them all, but I am still having problems. Its also difficult to tell what is abnormal because I assume there is just a natural variance in neighbor counts becasue switches face different directions and the RF won't be identical everywhere.
Not looking to get rid of my zwave nodes. Im actually really like them and LOVE the Zen switches. I am just trying to figure out how to fix. I spent a lot of time on it and Im out of ideas on how to isolate the issue. Certainly. not saying zwave is an issue, I have had it for years and its worked great for me up until now.
I'm sure you've thought of this but as soon as I read this I thought "ok what changed?" Was there some event that was the "now"? Your neighbor put up lots of strange looking antennas? There's a black van parked regularly across the street? Introduction of a new device? Something else happening in the 900mhz band maybe?
Not that I am aware of. What I meant with "up until now" was that I switched to the Hubitat. There is a small part of me that wonders if its the 700 chipset. Prior to this I used Smartthings and before that the Alarm.com GC3's built in zwave without any issues (both with older zwave radios).
But that is conjecture and just wishful thinking that a firmware update will solve the issue..... I have had way too much experience with holding out hope that a new update for something...windows, android, iPhone, Mac, whatever, will fix all my problem only to find out there was something else at play.
When I work up enough energy (and have enough free time) I may just try to split the zwave between my 2 hubs, use hub mesh, and see what happens. They are both C7s so it would be interesting to see how it fairs. Just not looking forward to redoing all my rules.
I have three of the zen25 paperweights in my closet. I did everything in the many threads to quiet them down. Logs even showed they weren't reporting. But when I tried to add one back during Christmas it started to crash my mesh within a day. Took it back out, and things improved quickly. I really don't know what the problem is, but i know i had them for about 2 years on a C4 and i never had an issue, which leads me to believe there is more to it than just being chatty.