HS3 Z-wave was great for years, now SUPER flakey

[dmcentire]

Hi, I've been a very satisfied HS3 user for quite some time and have (had) a super high WAF! But now even my approval factor is quite low...

Z-wave events now take anywhere from 5-30 seconds to switch on & off lights. And now the lights often quick cycle on and off a couple to a few times in quick succession, often not ended up where it needs to be.

I did notice first thing in the morning (or after a long period of inactivity) responses are usually instant (like they used to be). I haven't had a lot of luck using the Z-Seer tool to find anything interesting, and I've tried optimizing some nodes I feel are important.

I've tried a new PC with a fresh install of Win 7 Pro and HS3, with a restore of the previous system and I still see the delays. I've rebooted many times and even added a restart plugin to restart HS3 each night to no avail. I also tried running the beta z-wave plugin v. 3.0.1.208 although it didn't seem to make any difference. Went back to my Intel NUC win 10 version since it didn't help.

I swapped the SmartStick+ with another spare new one, restored my setup on it and it still does the same thing. I played with the COM port settings regarding UARTs. I tried different USB ports, as well as removed the extension cable I was using to get the SmartStick+ up high in the house.

Nothing seems to help. Here are some facts that might jump out at someone more knowledgeable here:

1. I have a bunch (5-10) devices that are offline. Nothing I would call critical, just some plug-in appliance modules that are not in use. Should I delete them completely?

2. I have moved most of my lighting to Philips Hue, and love it. They're reliable, neat, and the JowiHue plugin lets me do some cool stuff. Plus they're on a separate system so they're reliable. Mentioning this because the HS3+JowiHue+Philips Hub all work great without any delays or weirdness.

3. Maybe there's a different Z-wave module (other than SmartStick+) that I should try out?

4. I was tempted to buy the HomeSeer remote z-wave Rpi thing but wasn't sure if that would help at all.

Short of starting over (delete everything, reset stuff, add back one at a time) which sounds like a nightmare (events, some modules are inside wall boxes, etc. I'd really like to diagnose this and figure it out.

One thing, I use MiniMotes everywhere and they often "blink funny" when the event isn't going to process immediately. On the HS3 log it shows the proper "Device: Family Room FR minimote-a Set to 60" type of log immediately when I press the button, but the action it's supposed to do doesn't happen. This shows up like it's "trying" to run the event:

Event Trigger "Family Room Toggle bookshelves mm 3"

But the toggled light doesn't turn on/off for a significant delay.

To me it really looks like some kind of "throttled buffer" that's fully congested and so full it might drop data to process. On the z-wave plugin page it shows this information currently for the SmartStick:

Counters Since Startup
Request Frames 14347
Response Frames 10901
App Frames Waiting 33

Is that any kind of clue?

Any thoughts on how to troubleshoot this would be greatly appreciated!


Zigbee (like the Philips) is getting more and more appealing to me although do they make appliance and dimmer plug in modules?

Thanks!


P.S. My HS3 and pc info:

HomeSeer Version: HS3 Standard Edition 3.0.0.368 (Windows)
HomeSeer: Is Registered
Operating System: Microsoft Windows 10 Pro - Work Station
OS Version: 10.0.16299 Service Pack: 0.0

[jmaddox]

I wonder if you have a zwave device that's going bad, those can create transmission delays.

Try testing each device in z-seer+ and see if you can identify a node thats on the fritz

[randy]

Jonathan's suggestion is good. I'll offer a couple more. If you optimized with the appliance modules in place they will be active in the mesh. If you remove them without fully optimizing the system, you may have some corrupted routes. You do not necessarily need to remove them altogether, but you should fully optimize after they are decommissioned to make sure they are not expected to participate in any routes.

The biggest issue I had with delays was due to excessive activity - mostly polling and reporting. I went through the Node Information page and was very surprised to find that 50 or more of my devices had been set for polling without my knowledge. There were some Aeotec energy reporting devices that had polling set on every child node. I manually went through and removed polling on a massive number of devices (root and child). I only left polling enabled on my older Jasco devices that do not support instant status. I went through all of my Aeotec energy reporting devices and made sure the reporting interval was not too frequent. I found that 5 of my smart strips had changed to 30 second reporting, despite the fact that I had set them for 2-5 minutes, depending on their use. It turns out that most of the Z-Wave plug-ins were changing polling and reporting on every rescan. This was only fixed late last year. I had removed all polling and set the reporting intervals properly, but each rescan was changing these values. I also disabled command reporting on all of my devices with a script, then only enable it when I need to troubleshoot operation. I was seeing delays of as much as 5-10 seconds and all delays ended after this process. I went through all of this last year and my system response has been almost instant since. The single biggest contributors were polling and reporting intervals on multi child devices.

[dmcentire]

Slightly interesting development

Thanks for those suggestions, I'll definitely give those ideas a solid workthrough to see if they help.

One thing that has improved it -- although I am not totally sure I know why -- is that I was also working on a different thread on this forum about someone having trouble with a Leviton in-wall dimmer and while I was trying to find a root device I stumbled upon one problem.

I have this device that's how I control the Leviton. But apparently there's a root device that it's missing. When I clicked on the link (it directs me to this URL: http://<hs3>/ZWaveConfig) and on that page there is a button that says "Reset Z-wave data". I thought that was for this device only (from what I could tell) so I clicked it. It only took a few seconds to confirm that it was done.

Then I looked at the logs and there were lines like this for EVERY z-wave device:

Apr-08 3:39:21 PM Z-Wave Creation of Z-Wave plug-in object from HomeSeer data was successful.

Apr-08 3:39:21 PM Z-Wave Warning HomeSeer device Guest room Table lamp indicates it is owned by this plug-in but no device exists in this plug-in to match it - attempting to create it from the HomeSeer device information.

Not sure honestly what all that means (sounds like it cleared associations with every z-wave device in the plug-in's cache and repolled it from the SmartStick+?) but it seems to have helped.

It's not completely fixed but devices do respond faster. I did notice some hiccups although things are a lot more responsive.

Maybe there's still a flaky device in the mix which I hope I can identify with Z-Seer.

[mikee123]

I can second the polling issues. I had the same problems with devices becoming unresponsive, events and commands taking ages to complete, sometimes they didnt. I changed all the polling (as described above), most of my energy reporting is now only every 20 minutes, with essential ones every 5 minutes. All unnecessary reporting I have stopped completely. I also had a few (xmas) decomissioned appliance modules. I have been told to either leave them in or exclude them. Since I have done all of that my system has become completely reliable and fast again.
One thing you can do to test if its the reporting causing issues. click on manage plugins, then click on the zwave PI, and disable polling there. This will turn off all polling which is not ideal. But if that solves your problems, you know where to start, and then slowly start enabling polling, crucial ones first. Thats what I did

[dmcentire]

Lots of missing &quot;root&quot; devices

I've been going through the devices looking for polling values, and many of them display this when trying to edit anything on the z-wave tab:

Error: Z-Wave device has empty ROOT object, it may have lost its reference to the actual Z-Wave device,
try resetting the Z-Wave database from Z-Wave Config

I have already done that reset thing before but it clearly didn't fix the ones I've been finding.

The devices generally work in my system but I can't see how this is a good thing. Does anyone know how to fix them? I suppose I could delete the device and re-add it into the system to create new instances, but then it'll also require lots of event updating.

Thanks!

[mikee123]

I had that issue occasionally. I cannot give you any good news... the only solution I found was exclude and include again. Which is a major pita. Redo all events, scripts etc. I had posted here before and did not really get any answers except for resetting the zwave which didnt help in my case.

[dmcentire]

Still delays but found one thing

So I have done extensive cleanup, poll time checking, optimizing, removing bad/missing nodes, removed devices no longer in use with the usual remove a node method, and have gone through every device to fix the bad root node (remove, reset, re-add).

Believe it or not there are still z-wave delays of up to 15-30 seconds for events to fire. Here's an example:

1. Push button on minimote to toggle a z-wave light.
2. HS3 log shows button push immediately
3. Right after that the log shows "toggle <light> event triggered"
4. Anywhere from 5 to 30 seconds later, the light toggles.

It's like there's a bottleneck in the data getting to the SmartStick+ and then out to the device (outbound commands).

So I tried something. I went to the Z-wave config, disabled the interface (click on the little green checkmark).

Then I disabled the z-wave plugin and waited for 10+ seconds.

Then renabled both items and it's responsive and fast like it used to be. Not sure how long it will last though. Maybe there's a way to disable and re-enable the SmartStick every evening or something to "clear it up."

I'll update with more findings as they happen if anyone is still interested.

[jmaddox]

I wonder if your smartstick is going out?

[dmcentire]

Already tried that...

The SmartStick+ was already swapped to see if that was it and the problem carried over to the new one.

Although it wasn't installed at the same time I started having issues I had installed a cell-phone repeater for our coverage in the house (was pretty bad before) but from what I can tell the frequencies are different and they shouldn't interfere with each other.

It's a four band model (LTE700, Cell800, PCS1900, AWS2100) but again I though z-wave was 908.42?

I wish I knew what was going on. In the meantime, I am going to proceed with moving into the Insteon world and if it's more reliable I'll start moving devices that way.

I already moved my lighting *mostly* to Philips Hue and love it; it's reliable and fun with the JowiHue plugin.

[dmcentire]

Still happening....

It's so weird that my z-wave setup has been working for years and now it's so intermittent.

I've slowly been moving off z-wave to Insteon and have migrated over 20 nodes off the z-wave system and it's still not working right.

Lights take from a few seconds to never to operate. I've moved off all the minimotes so those are not the issue (I'll probably sell them off).

Anyway, I find it really strange that all these issues have popped up after being so reliable. I am now recommending to folks I talk to about smart homes to be wary of z-wave as I am not the only one this has happened to.

Thanks for listening

[gnichols]

Could be RF interference.

Do you have any baby monitors or 900 MHz phones?

I'm having problems also. I've lost about 10 devices on my network. They just don't communicate.

https://forums.homeseer.com/showthread.php?t=191446

[dmcentire]

No 900 mhz devices

No monitors or landline cordless phones here so I don't think that's it. I have a feeling as I migrate away from z-wave devices I might find what was causing the whole issue but so far I have not.

I'm curious how the diagnostic tool works out for you. I ended up giving up on z-wave myself and have started migrating to Insteon for most of my devices.

I had already moved to all Philips Hue lighting for lamps and other bulbs and love it. It's just that my z-wave remotes (minimotes) started acting flaky and then I realized the whole setup starting falling apart.

Nothing had recently changed and there was no reason for it (been using z-wave since my Vera days a few years ago) but it got frustrating enough for me to switch to another setup.

With Insteon remotes and keypadlincs, and switchlincs controlling Hue lights I have eliminated the z-wave element and the lighting in my house works perfectly with HS3.

I still have a few z-wave relays, so we'll see how that all works out.

Thanks and good luck with your issues!

[bebaldin]

The one thing I have always found to be the cause of this kind of thing is some bad device flooding the network with traffic. Try turning on Log Send and Receive Device Information to the HomeSeer Log (this is in the Z-Wave Plug-In configuration page). You are going to get a TON of log entries in a very short time, but what you are looking for is some device that is broadcasting a huge amount of data compared to everything else on your network.

Whenever I have had a slow network issue, I have always been able to resolve it by doing this. Remember to turn this off as soon as you can, because it is recording everything that every device does.

When I used to use an Aeon SmartStick (before I bought the SEL Controller and Z-Net Interfaces), they had a utility that you could load up on a Windows computer and see what the entire network was doing for traffic. That was useful for the same kind of troubleshooting.