#SILabs Recommended Firmware
1 messages · Page 1 of 1 (latest)
what controller
and what's a 202 error
Looks like Controller_MessageDropped.
7.19 and 7.20 are bad firmware
https://devices.zwave-js.io/?jumpTo=0x0000:0x0004:0x0004:0.0 is my controller and seems to recommend 7.19.2 which is why I asked
DB Browser – The official device configuration reference! If you wish to find out if your device is supported, please search for it here.
anything less than 7.21.4 in the 7.21 line is bad
and I looked through that doc, but my network has been in place for years with no changes lately so while it could be new interference, it just seems odd that it'd just randomly start in the last couple weeks?
If that's the case, why would upgrading the firmware be a good response?
Well because the thing that has changed is many HA updates, zwave js updates. Seemed reasonable to think perhaps it's not liking my many year old firmware.
7.17.2 is probably one of the most stable. I've been using 7.18.8 on my older Zooz. I wouldn't go with anything besides 7.18.8, or try the newest 7.21.4.
Anything in-between is likely to introduce controller hang problems. 7.21.4 may or may not fix it.
gotcha. Ok seems like I should try to do some other investigation first
Also, if you don't upgrade in the proper order you can brick the controller
hmm where can I find info on the proper order?
So 7.18.8 first then 7.21.4
I saw this the other day. https://community.silabs.com/s/question/0D5Vm00000JAwbsKAD/zwave-v7172?language=en_US
At the bottom is a matrix of upgrades they tested
fun 🙂
according to that, you'd be safe 7.17.2 to 7.21.3 (presumably .4)
I also had issues flashing BRD4206A, neutered range of controller while BRD4201A works fine. But BRD4201A doesn't support LR.
One interesting thing I noticed is almost every report I find online says Z-Wave error 202: Failed to send the message after 3 attempts mine always says 1 attempt. I don't know if some retry setting got changed?
Could have been due to flashing one of those bad ones
sounds like maybe before I do all this it's an opportunity to buy a spare 800 controller in case things go wrong
Do you see controller jammed or unresponsive messages?
I’ll have to check. I assume that would be in zwave logs, not HA?
usually in zwave logs, unless you turned on the integration debug logs, then both
controller status entity would probably show it in history
I see some occassional jammed messages in the controller history. I'd have to dig in further to see if it aligns with these errors
doesn't look like it
not seeing anything interesting in the zwave logs either. I had the issue occur at 11PM ET yesterday. So I pulled the 9/20 logs. The word error isn't found. But the HA log has 2024-09-20 23:00:02.005 ERROR (MainThread) [homeassistant.components.zwave_js] Unable to set value 118-37-0-targetValue: zwave_error: Z-Wave error 202 - Failed to send the command after 1 attempts (ZW0202)
There would be something in the driver log
hmm not sure where to find that. I went to zwave js ui, then the logs store/logs/z-ui_2024-09-20.log.gz guessing that's not the right thing?
ok. Seems to happen every night when one of my "turn off lights" automations runs so I'll see if it happens today
If you see any of the jammed/unresponsive messages, the firmware update to 7.21.4 might help. Would be interesting to know if so. Only the 800 firmware has been confirmed to help so far.
alrighty I'll keep an eye. As is often the case, of course now that I have debug logging on when I manually trigger the automation multiple times it's working flawlessly 🙂
yup I'm getting some jammed messages
Is it just because this automation is trying to turn off about 10 zwave devices? Maybe I need to add some delays?
probably related to that. you could try multicast. it doesn't reduce the traffic (adds one extra command) but might have a different result.
not familiar with that, got a doc I can read up on?
cool, not going to lie I didn't know zwave supported that
requires same security classes
otherwise it splits it. you may still end up with some errors, but something to try.
the actual HA device I'm changing is a helper/group. So it sounds like I could just specify that entity and HA will handle the fact that it's multiple nodes?
I live in the middle of no where, I have encryption turned off on all nodes
probably not the best decision but I don't have zwave locks or anything
this sounds promising, thanks for the suggestions! I'll give this a try before dealing with firmware
S2 encryption doesn't have much overhead, but it does ensure messages aren't corrupted.
Yeah, when I get an 800 controller and have to repair everything anyway I'll probably do it
to move from 700->800? I hadn't heard that was supported yet
Been supported for several months at least
yes, unless you fail the node and do a replace
gotcha, that still requires going to each device though right? I've got stuff in my basement and attic which is why I've been lazy and avoided it
Tried multicast, failed with Unable to set value via multicast I have the driver log but nothing seems interesting other than a transmit status: Fail on the multicast command. So doesn't seem like this helped unfortunately
Not surprised
I'm not surprised this didn't fix this, but I'm surprised that sending 7 commands on a mesh of ~110 devices seems to be overloading it
Depends on if those nodes are doing anything. The more nodes, the worse it could be. Probably won't ever work if you're being affected by the firmware bug.
The nodes shouldn't be doing anything, they're receptacles that don't support meter/reporting and they're off so I wouldn't know what they'd even be able to be doing?
yeah firmware will be the next step. To your original point about being skeptical of firmware being the cause though - I've had this all working for years. I'm not sure what broke in the last month or so
They probably would have been seen as dead before. Maybe you didn't notice?
Z-Wave JS detects unresponsive controllers now instead of declaring nodes dead.
That makes sense then. I did sometimes see dead nodes that weren't really dead
this is a Christmas/Halloween lighting automation so it hasn't been run since last January. Maybe that was before the change
last question I'll bug you with, I'm going to get an extra controller just in case I do anything wrong and brick it. Any recommendations of an 800 controller? Figure it'd be a good time to ugprade
Zooz or Homeseer 800
Thanks for all the help!
Hmm the zooz website says this: NOTE: Do NOT perform NVM backup from 500 series or 700 series Z-Wave sticks for this device, this feature is currently not supported and will brick the product
That's for PC Controller software, or other software. Z-Wave JS supports it.
ugh, how has silabs still not sorted out the 700 series firmware? 😵
I guess at this point it's not going to happen and it's also time for me to pick up an 800 stick.
They are still releasing updates. 🤷♂️
soon ™️
A few notes on the upgrades I just went through in case it proves helpful to anyone else in the future:
- I went from v7.17.2 --> v7.18.8 --> v7.21.5
- It's been awhile since I've done this, so I found this link helpful as a reminder of what firmware file to use and where to retrieve them: https://github.com/kpine/zwave-js-server-docker/wiki/700-series-Controller-Firmware-Updates-(Linux)#silicon-labs-uzb-7-slusb001a
- I am not seeing any range issue using the
BRD4206Afirmware. YMMV? - I only did a few rounds of testing on each of the new versions, but in general I didn't observe any reduction in the frequency/severity of "controller is jammed" (granted, this was a torture test of turning off everything)
Did you apply the new firmware files via Z-Wave JS UI?
If you're plagued by jams, also check out https://zwave-js.github.io/zwave-js/#/troubleshooting/first-steps
you might just have too much traffic going on
Z-Wave driver written entirely in JavaScript/TypeScript