I’ve had a case open nearly a week and now turning to the community.
When I have polling enabled on ~7 GE Z-Wave switches, my hubs ability to control locks reliability is zero.
I’ve disabled any custom drivers, I’ve disabled custom apps, I’ve removed “chatty” devices from my mesh. I’ve repaired my z-wave mesh as many times as I can imagine and rebooted the hub. Nada.
Is polling that stressful on the hub, or is this hubs control over Z-Wave locks that poor? The secure join process went off without a hitch on my three locks (2 Yale T1L and 1 Kwikset 912) and my GD00Z garage door control.
I’ll go hours without issue and then bam, no lock status, can’t lock or unlock, nothing. I turn off polling and then my locks start doing any back queued commands.
I’m not particularly willing to go spend $280 on replacement GE Z-Wave plus dinners as this whole setup performed perfectly on ST. I only migrated to avoid the cloud stability issues of ST which pale in comparison to the stress this thing has given me in the last 4 weeks.
I know there will be a suggestion for a z-wave repeater but my house is 1500sq ft single story and I never had any of these issues under ST. 2 of the 3 locks are unobstructed line of site and the 3rd is obstructed by a single pane of window glass.
As stated, this stuff is rock solid when polling is disabled.
Is there anything else I can try before I just toss in the towel?
Well at least you know the recommendation to solve your issue, whether or not you use that suggestion is entirely up to you.
I have a 1800 sq ft house, had two of my 3 Schlage Locks (zwave) were constantly having the exact same communication issues you state, One of the two had 5 separate "typical" zwave repeating devices within 10 feet of the lock that the lock itself was only about 20-25 ft away from the hub to start with with only one wall separation. Fought this problem for my first two months on the platform, Added 4 Aeotec Range Extender 6's (one by the hub, and one by each lock) and haven't had a single communication issue sense. Using the Reliable Locks user app polling each lock every 30 minutes none of the 3 locks have missed a poll, when before the Aeotecs 2 of the 3 almost always missed the poll.
The zwave radio/antenna in the C-5 hubs is weaker than those used in Smartthings, Wink, Vera, etc. However this is only really noticeable when communicating with "Secure" devices (Locks/Garage Doors) << Also make sure your zwave settings are set to that and NOT "All Secure Zwave".
Yeah, I’d just as soon fire up the ST hub and just start using that again.
Has there been an official reply ever as to why the z-wave antenna is so much weaker than what’s been shipping in other consumer devices for years now? It just seems like such an oversight if that’s truly the case.
Nope, but one could assume that the form factor of these hubs is about half the size of all of those other brands, thereby space is much more limited. And it is only noticably weaker with ONLY secure devices (locks/garage doors) no other devices have reported these issues.
I think they are two separate issues. I have multiple Aeotec range extenders and no issues with my z-wave locks. So I agree with @waynespringer79's suggestion.
But if I poll a bunch of devices using the z-wave poller app, my hub slows to a crawl for all z-wave related events. Zigbee-related events at the same time have no issue, so I think there's either a queue for the z-wave radio that is held up waiting for responses from polled devices, or the radio is too busy polling. So I stopped doing it.
I have no technical basis for this comment (that I'm willing to share publicly), BUT... Anecdotally:
Hubitat seems to have a MUCH weaker zwave radio/signal versus Smart Things. When I moved from ST to Hubitat, I could see that immediately. NOTE: This was on the C4 w/dongle - I have not retested with the C5 hub.
Hubitat simply does not work well with zwave locks. It is not just a signal strength issue. Yes, others will now chime in and say it works fine for them - good for you. But I am sharing MY experience. Example - As a test before swapping my locks out for zigbee, I left a lock 5 feet from my development hub, with no other zwave devices on the hub - so no mesh issues or badly behaving devices to interfere - and the lock still only worked about 80-90% of the time. It worked probably 99+% of the time on ST... NOTE: This was on the C4 w/dongle - I have not retested with the C5 hub.
Anyway, good luck. Hope you get your issues resolved. I don't have any standard zwave (aka non zwave plus) devices in my home, so I can't comment on the impact of polling.
I will just add one comment. Since locks are battery powered, the Zwave radio is not always on and the lock will miss communications from the hub. Zwave Plus powered devices that support beaming, will capture the communication to or from the hub and forward it on. With constant polling even if you had beaming capable devices on your mesh, there might be enough traffic to spoil the lock to hub communication????
So far the replies aren’t promising. To summarize what everyone is saying
Zwave locks work poorly. Buy many extenders for $30 each or new locks to move to zigbee
polling doesn’t work either. Buy many new zw plus switches and dimmers.
Kind of a bummer, considering my ST hub from 2015 or whatever did both of these no problem. I get that the polling thing was a GE/Jasco thing but other platforms seem to handle it fine.
I agree that it is a bummer. And I agree that it worked 'good enough' in ST, and does not in Hubitat - with the same devices.
You could try to discuss with support, but I'm not sure others really made any progress going down that path in the past.
I chose to replace every non-zwave plus device with zwave plus alternatives (except locks, which I went zigbee as it was cheaper - just had to replace the comm card in my kwikset locks). But I would have even if I were still on ST - zwave plus is just that much better and more reliable (and faster once you have only zwave plus devices - speed kicks from 40kb to 100kb).
I have devices that just don't exist in the plus variety.
I have to decide what is more important, 100% local processing or 100% device reliability.
On the topic of the aeotec repeaters, I see some have gotten away with one, and some need 4. If I have beaming zw plus devices near each lock, would it be OK to assume that I may only need one repeater, placed nearest to the hub? I fail to see what a repeater next to each lock will do when they're sitting next to a zw plus beamer. The only explanation I have for the issues now is that the hub is in a void where it itself has no beaming/repeating devices immediately next to it.
If you are using the Z-Wave Poller built-in app, you might try disabling it and creating a simple triggered rule to poll the devices with a longer interval and see if your problems diminish (try 15 seconds or 20 in a rule instead of the apps' fixed 10). If the cause of your lock issues is indeed polling saturating your mesh there should be some correlation seen.
Have you noticed if the polling (as indicated by logging) is occurring as expected for the devices you have selected? I stopped using the Poller app in favor of a triggered rule; I had used it with 5 GE switches and initially it seem to function properly. I never experienced mesh issues but with some firmware update this Spring (don't know which, the software version of my hub is fluid) the app stopped working reliably.
It would frequently stop polling all selected devices; sometimes the app would appear to ignore which devices were selected or unselected for polling (sometimes continuing to poll unselected devices long after polling should have been disabled for them), sometimes (per the log) it would poll one or more of the selected devices while ignoring others... In general it just seemed flaky and didn't respect the settings that were displayed.
There weren't any log entries to show anything amiss and I never bothered reporting the problem since frankly I didn't feel like wasting my time or customer supports' as I expected any effort to fix it to be low priority given the frequent advice to 'buy newer switches'.
The 'app stopped polling for no apparent reason' issue was mentioned by at least one other user in a relatively recent thread but either this app doesn't get much use or it's not a widespread problem... At any rate since my triggered rule has been in place (with the built-in Poller app disabled) I haven't seen any issues with polling in my environment.
Sorry, just to even the score on your inquiry result: My Z-Wave lock works perfectly, without repeaters and without polling. So you can withdraw your summary of what "everybody" is saying. It's just because some locks work on an older, less evolved, version of z-wave don't work that good doesn't mean HE is bad with locks. There are plenty of people using locks here without any problem, but it micht be that many of those locks are Z-Wave Plus locks like mine (DanaLock V3).
Though I agree with you that it is a bummer that it doesn't work as expected in your case. And I would definitely report this to support@hubitat.com despite of JasonJoel says.
@bobbyD and I have been in contact, but we've just not made any progress.
As mentioned, I've taken the following steps to resolve the problem to no avail
excluded and removed my Aeon HEM for being too chatty
set my mimolite device to the built in driver (rather than the gate/garage control driver)
an infinite amount of z-wave repairs
reboots, systems updates, more reboots, etc.
Realistically I'm not sure what is next. The Yale locks don't seem to have gotten the same attention as the Schlage ones, though I am not sure it's necessary. I am following suggestions from above which is to poll using RM rather than the poller app. I am testing that now at 15 seconds intervals.
I think some enthusiasts read the suggestions and thought "Oh I bet that's my problem" and purchased a number that would fix it. In your case, ONE might be enough.
ST has a permanent poller running. It polls everything and obviously that's not needed as Hubitat polls nothing automatically. ST has the cloud to 'delay things' and therefore have moments when polling is just 'filling in' the cloud delays. Hubitat doesn't have those delays. Running the Zwave Poller app is intended for a small subset of devices. You've probably noticed that you can select only switches and dimmers and only switches and dimmers using the 'non-smart' drivers (generic zwave switch vs generic zwave smart switch, and dimmer.) Those are the devices that were impacted by the Patent related to 'instant status.'
The price of a single hub dictates how many people will try it out, I feel. I currently have 4 Hubitat Hubs, 3 of them connected into one 'system.' That means I paid almost $300 for "my hub" and it's integrated with 120+ ZWave devices (9 zigbee) which means 'My System" is $$thousands. Adding a $40 repeater doesn't seem too harsh if it truly fixes a specific problem.
I personally have zero ZWave repeater-only devices, I have more than enough in-wall outlets, switches and dimmers. My ZWave Yale lock works fine, BUT it did NOT work fine the day I switched it off ST to Hubitat. The next day when I had migrated more of the devices around it to Hubitat, I guess one of them was the 'beamer' it needed and it's been 100% ever since. It's showing a battery of 52% which I think means the battery is on it's last legs.. but it still worked this morning.
On the other hand, I have 4 Hampton Bay Fan Controllers and EACH one needs it's own Zigbee repeater.
Summary... I 100% agree that buying repeaters JUST to get the hub to work the first "week" of use is excessive... but the majority of the people that answer questions here are rather fully commited to this path. For them, a magic bullet isn't as disturbing.
