Ok I used to have a rock solid C7 network.
My C7 had the antenna mod just for the Z-Wave antenna.
My C7 was installed in our MBR closet where all my other HA equipment is and a few other things.
I used a long ethernet patch cable to move it onto a shelf on the other side of the closet to get it away from possible EMI issues. As a result out of 38 devices there was only 1 maybe 2 that were not direct.
All my switches/outlets are Z-wave Plus.
After the upgrade to C8 I noticed that my mesh had a lot more devices with routings. There were NO zombie nodes! On the order of 25 were now direct with the rest being routed. My closet has a lot of wire shelving so I decided to move the hub outside of the closet to the nearby family room. The improved the direct count but it has now settled back to the 25 direct count.
Now the bad part. I have various HA routines. One of them is to turn on night lights shortly before sunset and another is to turn off a lot of lights at 11:30PM.
The 11:30PM routine is erratic in that it just randomly misses some of the lights in the list.
The log shows the routine firing and lists all the devices to be acted upon but some of the lights just fail to turn off. They are not showing in the log list as Turned Off.
My wife will then do a manual voice assistant that triggers Hubitat to basically turn off all the lights again and it will work fine.
Last night the light that failed to turn off at 11:30PM was a non routed direct light switch.
The hub in new location would be about 8 feet away from the switch on the same wall.
Other nights it would be one of the other lights. It really is random and some nights everything works correctly.
I have the same issue that exists in both C7 and C8. It only affects turning off lots of devices at once (in my case 17). Randomly 1 or 2 devices do not turn off maybe a few times a week.
I got tired of screwing with it and got an app here by Bruce Ravenel called "All Off". It lets you meter (delay) turn off spacing between each device and also supports any number of retires. In my case I set metering to 200ms with 3 retries. The problem went away for me.
Perhaps. In general, the zwave experience with the C-8 seems to be pretty good.
I had three devices that were connected directly to the C-7, which connected to the C-8 through multiple hops. Sometimes, but not always, reports sent by these devices were not being received by my C-8. I just excluded and re-paired these devices which resolved that issue entirely.
You might consider posting your entire C-8 zwave table (settings -> zwave settings). There are multiple zwave experts here who can make suggestions to the residual issues with your zwave mesh.
I had the same type of problem not too long ago. I did some zwave device repairs thought it fixed it nope showed back up. All the devices that were in the rule were direct route so kinda baffled me. The fix for me was either pulling the air-gap on the dimmers and unplugging the Aeotec MS6. Turned out to be the MS6. Just because devices say direct route doesn't eliminate them from causing problems zwave problems. Pull those air-gaps power cycle those devices helps way more then it can hurt.
It's just a good place to start if you know you have a solid mesh.
In my experience, absolutely. Though not nearly as widespread as the zigbee problems there are a number of reports of z-wave issues with the C8. Understandably they don't seem to be getting a lot of attention in the releases yet but hopefully as the zigbee problems get under control that will change.
I've had a myriad of z-wave stability problems since migrating to the C8. My C7 mesh was rock solid. Since the migration I've had devices that were working before now become unreliable, and worse, my entire z-wave mesh periodically simply stops responding.
I've been trying to troubleshoot it for a couple months now with mixed success. At first I had 2 ring keypads that used to be reliable become completely unstable, dropping off the network at random times or looking like they were connected but simply not responding. Then it was motion sensors that would go through batteries in a day or two. Finally got those squared away. Then (and this is still an issue) all of my z-wave devices would simply stop sending events, The motion sensor light flashes, but there is nothing in the event logs and the corresponding rules wouldon;tdnt fire. Super frustrating.
Some folks have noticed that using a PoE splitter, particularly with Unifi switches, can cause z-wave issues on the C8. No one is quite sure why but there does seem to be a trend.
In my case I keep thinking I have found the issue only to be trying to find the bathroom in the dark at 2am the next morning. I'm still looking though! In the mean time I've moved some critical devices back to a C7.
And ironically just after I composed this post... z-wave became unresponsive again.
As previously reported, I also have Z-wave issues since upgrading from a C7. I have noticed that the closer the device is to the hub, the worse it gets. I e.g. have three devices within 9ft of the hub that are continuously (between 10 and 20 times a day) changing routes. While most devices are connected directly to the hub, these (close by) devices often connect via 3 or more devices at the furthest part of the house to eventually get to the hub 3 ft away.
Sorry, can't remember - have you tried individual repairs on those close-by devices that are using the odd looking routing?
If you haven't yet, you might want to do the shut-down dance (Settings>Shut down, pull hub plug at wall when red LED appears, wait a minute, plug in again) and then to the repairs, and see if things improve. If you've been there, done that, apologies for repeating...
Interesting about the POE adapter! I used one as well. I will try the C8 included charger and see how that changes the game. One thing at a time as they say.
I'm seeing very similar Z-Wave failures. My C7 Z-Wave network was rock solid everything was working perfectly with very few redirects. I upgrade to C8 and it's a mess. I let it soak for 3+ weeks hoping the mesh would figure it's self out, but no luck. Random devices failing, huge delays in automations (i.e. automation to sense a door open and turn on a light would randomly turn the light on 5+ minutes after the door opened).
I removed all 50+ Z-Wave devices and did a radio reset. After the radio reset the network is still completely unusable / unstable. Even with no devices joined after the reset, it was impossible to get a general exclude to work.
I have a map of where all the devices are in the house and used that map to build my C7 network that was completely solid. I've put the C8 in the EXACT same position the C7 was and have attempted to rejoin some devices and have been able to get a few (after general exclude using Zstick 5). Yet even those devices are sketchy at best.
I've been letting the few devices that have been rejoined attempt to rebuild the mesh and even attempted to force it at one point with a repair. No luck. Many devices have numerous route changes.
Not sure where to go with this and would appreciate any ideas suggestions. Maybe my hardware is bad?
Thanks for the response. Yes, I've tried that many times. Things are substantially better now than 3 weeks ago but still imperfect. I get the often complaint about delays. Sometimes the devices (I'm talking Z-wave) would respond instantly. Other times, there's a delay lasting from a few seconds to many seconds, creating the impression that the Z-wave network is not responding. The only noticeable issue is the constantly changing routes of close by devices. The routes of devices further away from the hub stay the same even after 2 or 3 days.
How are your device RTT average numbers in your Z-Wave table, for those devices where you are experiencing intermittent delays? Any in particular stand out w/larger (worse) values?
RTT Numbers are all over the board, I'm seeing from 48ms up to 410ms. No visible rhyme or reason why some are direct connected to the hub (with 300+ms) others are through repeaters with lower (100+ms).
The Z-Wave routing process is supposed to find the best routing for each device, but I think we've always felt that to human eyes it sometimes looks quite odd. Luckily in many cases odd routing doesn't result in performance issues - e.g., I've had devices routing through four others (including wacky cross-house jumps) and still work w/out issues.
It can take time for devices/routing to settle, sometimes more days than expected to work it out, but there are obviously cases where no matter how long you wait the end result can still be sub-par.
There have been a small number of issues w/the antennas. Have you tried swapping them to see if that makes any difference?
@brad5 and I are a couple of the people cursed from POE splitters with the C8. C7 was rock solid, and even using the same POE splitter with the C8 (MicroUSB to USBC adapter), I have issues. I have tried 3 different POE splitters, as well as source power from a Unifi POE switch and a 30W POE injector and nothing seems to work like the AC brick does. It's very weird, if I use the provided power brick, everything is perfectly fine. If I switch to POE power, things fail. Though, others have been using POE with C8 and their setups have been totally fine. I believe @brad5 is still seeing issues, even with the provided power brick, so I don't believe everything is POE related.
Thanks for the info. I have switched to the brick and we shall see what happens. I also moved it back to the original C7 location as well. The power specs on the brick must be special because I grabbed one of my iPhone bricks without thinking and the C8 would not even show power. Why o Why do they have to be special.
Well, that did it for me as well. Swapped the POE splitter (after reading this) for the original power supply, and BAM every single device is working and my RTTs are all less than 10ms. My eyes aren't what they used to be but it looked like the power supply had output 5.7V listed on it, instead of 5.0. That would make sense, the hub would function, but potentially not have enough power for the radio(s).
