Recommendation for Z-Wave gateway?

ok bro :smiley:

2 Likes

What makes you think it will give a different outcome? Remember Zigbee can be more prone to Wi-Fi and other unlicensed device interference due to the RF it uses.

I am actually shying away from lots of work for doing this and tried to find some out-of-the-box device/solution - which doesn’t seem to exist. I didn’t know Hubitat though, so thanks for coming up with that :slight_smile: The number of manufacturers and ecosystems is rapidly growing over the last years, i’m always happy to get to know more of them.

Home Assistant can also be set up much like OH or it has its own device available too. Unlike Hubitat you are not required to buy their device.

My zigbee mesh (with Philips Controller) works for years without any issues. I think i had to unplug the controller one or two times since i got it, i could hook up any other zigbee device i wanted and the range seems decent for the building. Maybe i am just lucky with that, i dont know. I only have my own installation as reference.
The reasons for trying z-wave were curiosity and a few devices that i could not find in the zigbee world. Also i had the feeling z-wave might be a bit better, as you mention it yourself. My setup has too many issues though, and if i cant replace the usb-stick in the server with something independent from the server that doesnt require constant attention i probably wont stick to z-wave.

Hi Florian, I was also looking for a ZigBee controller that can communicate directly like with MQTT. It looks to me that the manufacturers don’t like this idea. They want to tie you to their ecosystem. So I tried several sticks that were not reliable enough. I also bought a “Sonoff ZigBee Bridge” and reflashed it with Tasmota. This was very easy and there are enough instructions on the internet. This device could be placed anywhere and it was very reliable. But similar to the sticks on the RPi, someone has to take care of new devices and update the firmware, which can be installed via OTA after the initial flashing. The current firmware does not support all the devices I own and so I decided to go back to the RPi, which now runs a TI developer board (CC26XR1) on USB as a bridge to MQTT (OpenHAB - MQTT - ZigBee → just the way I like it). This is just my experience and I hope it helps more than it confuses. :wink:

Zigbee chips with a serial output are available. Using a network interface & MQTT would involve more hardware & cost for a small potential market.

Although manufacturers would like to lock you into their system they must support standard Zigbee in order to claim they support Zigbee.

A few years back with openhab i had some zwave reliability issues similar to this. I created some watchdog type of stuff that watch to see if I was no longer getting zwave network frames and would restart the whole process. Not ideal but it worked. After awhile I eventually switched over to using my aeon zstick plugged in to a raspberry pi, running a docker instance of zwave2mqtt. I haven’t looked back honestly.

Not sure if it was how docker was interfacing to the serial, or how the author wrote the app inside the container that the moment the serial connection flaked out it crashed that app causing the docker to immediately restart or what but the zwave2mqtt instance has made my zwave experience LIGHT years better.

I was able to completely decouple my openhab install from the zwave hardware all together by doing that.

I’ve had a similar experience. My Philips Hue system works flawlessly. I’ve noticed in a couple of rooms where I have both Zigbee and ZWave devices that occasionally the ZWave devices turn on a few seconds after the Zigbee devices. It only happens very rarely though, maybe once or twice a week? I’ve not measured it so I can’t say exactly.

That is good but did you also move the position of your controller or any other aspect of your network?

Also so people are aware note this warning docs: unmaintained disclaimer · OpenZWave/Zwave2Mqtt@717206f · GitHub

1 Like

It does not sound like a big issue to you but if it is get a zniffer. You find all sorts of things like faulty modules or module configured to spam the network with reports that you are not aware of.

Being able to place your controller in a good location has a lot of advantages so I understand the question asked in the original post but delays like this, if they are rare, are probably not due to location but thy could potentially be if the routes used are very long. Again zniffer will show if the rooms that have delays are using many repeaters.

That’s good advice. I’ve read about zniffers and even bought another USB stick to use as one I’ve just never got around to doing it. :slight_smile:

1 Like

I did not know that but thought zwavejs2mqtt was interesting. Using a product like that means updates are not tied to the OH build structure.

That is one issue with the current zwave binding IMO. That dependency was needed for it to be an official binding though.

I am having the same challenge, as OH is running in a QNAP VM in the basement and my z-Wave devices are in the attic (where I do not want to move my QNAP for heat reasons).
Would it be possible to use a “USB network extender over RJ45 Ethernet, Driver free” to delocalise the stick in the attic and be connected through a direct Ethernet link to the Qnap? This would need no additional configuration. Or is this not possible?

All I know about QNAP is there are no end to problems openHAB users have trying to get OH to run on them (and they are pretty weak from a security perspective, never put a QNAP directly on the internet).

Until now, I have not run into any issues at all with openHAB 3 on a QNAP VM. Even my ETS recognises its dongle in the Windows VM (which is officially not supported, if I correctly remember). My openHAB 1.8 is running on a 14 year old iMac, so I definitely need a need to migrate. Could you tell me which OH & QNAP issues you are aware of? Until now, I only lived advantages, like easy snapshots before trying out new bindings, or duplicating the openHAB VM for experimental reasons.
Of course a NAS should not be exposed directly to the Internet. That is clear to me. Neither does my openHAB installation.

I just see post after post of people having problems installing OH on QNAP or having to go through complicated procedures to expose USB devices to openHAB running on QNAP. When OH 3 came out we couldn’t get confirmation from anyone that OH 3 even runs on QNAP so the installation instructions were removed from the official docs (that may be changing, there was some recent activity on the GitHub repo on this topic). Over all, given the number of users who claim to run OH on QNAP, they seem to have the most installation and configuration problems posted to the forum than other installation platforms and methods. But that’s just my impression.

1 Like

I use a Vera Edge hub to control Z-wave lights, plugs, thermostat, motion sensors, and a door lock via REST commands to drive functionality to/from OpenHab. There used to be a 1.x binding specifically for this on OH2 (MiOS) but it was not ported to OH3, instead I implemented a small lua code trigger for the Vera and a simple OH rule that generates HTTP commands back. The vera also serves as a backup HA system although I rarely use its interface. For what I control on my z-wave network its bulletproof.

You need to initially configure all Z-Wave devices through the Vera system though. Another possibility might be to have Vera work through MQTT. That is what the protocol was designed for.

I have set up a remote OH instance now with a pi3 (plus the z-wave stick) and already ported (and included in main instance) two z-wave devices successfully. I will report back how this develops in a longer run.