Stable Z-Wave

Trying to figure out if I expect too much of my setup?
I’m running openhab2 on Ubuntu using apt with “Testing releases” and zwave binding 2.4.0
I used to run the security development branch but have now switched.

I experience issues that response time on actions/events vary a LOT, and sometime doesn’t happen att all.
Does anyone have a running setup like mine (~25 nodes) where the events/response accuracy is around 99% success?

The reason I wonder is because I’m using this as home larm system but I can’t really rely on it when response time what it is.
Is this a consequence of my setup or is this the way it is?

I have my controller on floor 1 with nodes on ground floor and my cellar but there’s not so that the nodes in the cellar respond less accurate.
I have a Aeon Labs Z-stick gen 5, could this be a source of problem? Is there any better controller?

I’m fine with the answer that “this is normal” but if everyone else have super stable systems I want one too :slight_smile:
Thanks
Tommy

Generally, responses should be within ~1 second - normally a lot less than this. If this is not the case, then I would look at your network - try to establish a pattern and see if you can improve the mesh by adding more mains powered devices in the “right” place.

Can you explain what this means? Do you mean that 1 request in 100 times out?

All Z-Wave controllers are the same - the Aeon is fine.

To my understanding I have main powered devices in “strategic” positions… Is there a way to see find out that my network is suffering from this?
Maybe filter for certain log entries like re-sending, timeout or so?

More if someone had a system where every action to a node succeeded and the node always responded, always. Like really stable :slight_smile:

Great, thanks

Yes, I would suggest to try and work out what devices are not responding - or are timing out - the most. I guess you can filter the logs looking for the timeout messages, or you can look at the response times which are also logged.

Apologies, but I’m still not sure what this means. I guess it is relating to timeouts, but I’m not really sure - sorry.

:slight_smile: English is not my mother tongue.
Rephrasing the original sentence…
Does anyone have system with similar size like mine (~25 nodes) but more stable like 99% success with events/response?

1 Like

I think I’m up to around 40 nodes- latest milestone seems stable for me on zwave

@TordYevl

Can you give some details about how you give the commands to the actors when you test their reaction?

Cause at the weekend I added a new actor (Hank Smart Plug) to my network (Z-Wave controller in the basement, new actor in first floor), and it also was reacting slow. I was giving commands with Openhab app, on the second floor, connected through a WiFi repeater to my Wifi router (local connection). With this setup, reaction occured about half a second to second later.

But then, I switched of Wifi, set phone to LTE and connected the app via cloud (remote URL). And strange thing was: Reaction to the commands was immediately!

So it seems to me that latency in local Wifi has to be considered when actors are acting with a delay.

Also you may use the Z-Wave network viewer in Habmin, my nodes are all interconnected.

2018-11-19%2011_29_31-Mozilla%20Firefox

You may also install a mobile SSH client and your device to tail the logfile the same time you give commands, so you can monitor reaction in realtime. Good luck!

I am using openHAB 2.3.0 Release Build by the way.

Thanks everyone for your thoughts, good idea to check the network viewer.
Where can I find what the different colors mean?

I have close to 100 nodes, and my network is quite speedy. Others have even larger networks and also have fast/stable networks.

Hmm. Where is this logged? I know you show it in the log viewer, but I can’t find it in the log file.

Maybe I’m wrong then :blush: . I’ll take a look, but I thought there was a log entry - it might just be in the viewer though…

Thanks for checking. If it’s not in the log, it would be a great addition. :slight_smile: Would make it easy to periodically scan logs to find problem areas… OTOH, I could just grab the relevant piece of code from your log analyzer. :wink:

Hmm. Looks like the log viewer shows the response time for the ACK, but not for the actual response message.