Controller is unable to transmit - only fix is to pull the plug!

I use Z-Wave JS UI with a Zooz ZST39 Z-Wave stick, flashed to the latest firmware.
I also have 15 Zooz ZEN76 800LR switches. I tried using Z-Wave JS UI to flash them all to the latest 3.30 firmwarE. I succeeded for 14 of them. For the last one, which is very distant, the flashing always fails. And it fails spectacularly : the controller goes into the “unable to transmit” state, and the whole Z-Wave network essentially goes down - at least, HAOS can no longer communicate, since I’m not using direct Z-Wave associations.
Restarting the add-on does not fix the issue. Neither does restarting HA.
The only fix is is to physically unplug the Z-Wave stick from the USB extension cord that’s going to my PC, and plug it back in. At that point, HA can start using Z-Wave again successfully.
This is 100% reproducible - I have never managed to flash that switch yet, despite many attempts. I can’t bring the device closer as it’s hardwired. I could bring the PC running HA closer.

I’m not sure where the root cause lies, but it’s a really difficult one to deal with. I’m not always physically present when I use my Z-Wave network - sometimes I’m remotely logged in via VPN. Physically pulling the plug is not really a viable workaround for this situation.

Has anyone else ever seen something like this before ?

1 Like

Did you update to SDK 7.19? Supposedly that version has a bug.

Yes :

FW: v1.20
SDK: v7.19.3

The next version of the driver (12.0.0) has a workaround (add-ons pending update) to reset the stick. If the theory about a firmware bug is correct, then it won’t be solved until Silicon Labs steps in. That’s my understanding of the situation.

You can try a “soft reset” from the ZUI advanced menu for a temporary workaround as well.

If there is anything I can do to help trace/fix this bug, I have a 100% reproducible test case, so I’m happy to help. At least narrow it down to firmware bug or otherwise would be helpful. Sure seems like firmware, though.

I don’t think any more data is needed at this point, but you can decide that for yourself:

Thanks - added myself to the two HA bug. I hope Silicon Labs can fix their firmware.

Are you sure you guys talk about the same issue here ? The issues on Github for SDK 7.19.x are about the unable to tx error happening during normal operation of the mesh. The issue @madbrain experiences seems to be only triggered when he tries to flash FW on a distant node, it does not seem to happen for him during normal operation. Or did I get the OP wrong ? These seem to be two different issues (but possibly linked to the same root cause).

I also saw this error once during normal operation of the mesh, when some nodes died, including possibly the same node I have been trying to flash unsuccessfully for a week.

Ah ok. Because if the mesh goes down for you only if you try to flash this particular node, then the obvious workaround would be to not flash it :slight_smile: At least until the FW bug (if it indeed is one) is resolved.

Well, yes, but I’m trying to flash it to get some bug fix in there. The other switches have been behaving better since the update, but not this one obviously. Dead nodes still happen though, which is disturbing in a network with 42 nodes (only 8 on battery, all others repeaters).

If I see this and the 7.19 firmware, at this point I just presume it’s the same root issue.

Yeah most likely.