Remote z-wave module


I have home-assistant running on my server. The server is the basement, out of reach of the z-wave mesh.
Now I do have a raspberry pi with a z-wave module, one of those gpio hats.
Can I let home-assistant running on the server interface to z-wave on the raspberry pi? How?

Are you running mostly zwave sensors or switches?

The state stream option is an easy feature to send all the states back to another HA instance. When you need to control a second instance and get the states it is doable but a bit more work.

Do you have some pointers for the second option? I’ve been googling for it but have not found anything.

Check out this thread: RPi as Z-Wave/ZigBee-over-IP server for Hass

I just implemented this last week and it’s working pretty well so far. Our server sits in a closet upstairs, but the RPi with Z-Wave stick is down in the center of the first floor. I have noticed a slight delay, but haven’t confirmed that it’s due to this or just the new Dome Pro Door sensors I added.

Edit: just re-read that you’re using a GPIO hat - not sure how that might play into this. Sorry!

Another option is to use a range extender. I use this and have a similar situation to you. (server in the basement, all of the sensors are on the first floor)

If you install HA on the Pi instance and setup MQTT statestream you can then send all the states back to the main instance and setup sensor/switches/etc. as needed using the statestream data as input.

To change the state of a device on the remote instance, from the main instance I have found this works best.

A powered switch or the range extender is a lot less headache


A wired switch, outlet or even a Z-wave light bulb is a better alternative to a repeater in that you end up with a usable device. For the most part, as you continue to add additional, well placed Z-wave devices, your network gets more and more reliable. I used a light bulb to bridge the gap between my house and my detached garage and even further remote shed. The light bulb was the least expensive Z-wave plus device I could find.

Well, another problem is that the server is a pc so even with repeaters I cannot get it to work that way.

I don’t understand your reply. Many people use HA running on a server. I have HA running as a virtual machine on a small server in my basement. I have a couple of Z-Wave devices in the basement plus I have the z-wave stick on a USB extension cable to get better access to the rest of the house. As long as you have wired devices, they can pass the communication, device to device until it reaches the stick. You can have up to 4 hops between the furthest device and the stick. More devices make for a stable network. The more well-placed devices you have the more alternate communication pathways exist.

Well the situation is that my z-wave receiver is on a raspberry pi; it uses the gpio-pins so I cannot stick it on/in the (regular pc hardware-) server.

Why not set up the raspberry pi as the HA server and forget about the thing in the basement. Place the Pi in a better location in the house. Why complicate things??

Because those SD cards keep failing. And also every RPI I put this z-wave adapter on crashes after a couple of days.
This server is already running and also much more stable.
Moments ago I decided to spend some money on an usb z-wave stick.

A Zwave stick was my next recommendation. Good luck.

I tried something different. I found that the z-wave device only talks serial. So using some socat trickery I connected the razberry to my server.

On the raspberry:

/usr/bin/socat -d TCP-LISTEN:23000,fork,reuseaddr /dev/ttyAMA0,raw,b115200,echo=0

on the server:

/usr/bin/socat -d -d pty,raw,echo=0,link=/home/hass/zwaveport tcp-connect: is the ip address of the raspberry pi.

Then configure it

usb_path: /home/hass/zwaveport

Sofar it SEEMS to work: it sees the controller and a couple of devices.

That may work but it will always be a kludge. Wouldn’t selling it to someone who wants to use it for its intended purpose and getting a supported device make some sense?