I’m no electronics expert so forgive the potentially silly question, but how come the scope is showing 2.405 GHz/2405 MHz if the frequency is supposed to be 2450 MHz?
1 Like
The batch shown in those pics tests good, but have they found switches where the frequency is significantly off? Or is that still waiting for confirmation?
FWIW, my switches are 94:34:69, 38:5b:44, and 04:0d:84. I’m not sure if there’s any correlation between the address and whether they work for me or not; the 04:0d:84 ones work fine, but they’re also the closest to the coordinator. The one I couldn’t get to pair at all unless I moved the coordinator closer is a 94:34:69 (and that one is still offline most of the time). The two that are borderline are both 38:5b:44. But I have some 94:34:69s that work fine (but they’re not too far away).
Well the, “good” news is these are the ones in question and are on the naughty list. So, I’m not surprised they’re giving you issues.
Thanks for confirming!
Yes these were the ones they tested to confirm were good. They are now looking for the ones with the IEEE in question (the ones you posted above). We hope they look at those tonight or at the latest tomorrow night. I know they were trying to find those ones as all they had on hand were the good ones.
5 Likes
In case it helps, here are my two switches after moving the coordinator closer:
0x943469fffe089a61 – Has never had pairing or connection issues. Shows online, with an LQI of 100+.
0x943469fffe089d93 – Has had connection issues. Currently, with the closer coordinator, it will successfully pair/pass the interview, but the LQI sits at 0. (Surprisingly it has shown “online” status inside the Z2M GUI for the last 36 hours and counting, but I’ve had cases before where it will eventually drop to “offline” at some point).
Edit: After watching further, the a61 switch above is now reporting LQI of 3 (but still shows ‘online’ status). So I guess maybe I’m not out of the woods on that one being fully ok? Just wanted to provide latest updates on my IEEE numbered switches, in case it helps with further troubleshooting.
1 Like
I had a single switch in my box of 10 that wasn’t either 94:34:69:FF:FE… or 38:5B:44:FF:FE…, and of course it was one of the three I hadn’t installed yet. I swapped it for one that was giving me trouble and it has a much stronger signal to the coordinator and has even taken over routing to the bulb it’s meant to control. Looking much better for when the rest are up to speed.
3 Likes
I was finally able to find one that would pair in the location that was giving me trouble, so currently I have a pile of 3 switches out of the impacted range. I’ve also noticed another switch which is paired, but misbehaving with constant timeouts noted in the zigbee2mqtt logs meaning that adaptive lighting stops working and I’m unable to set any config parameters.
Looking forward to a fix soon so I don’t have to swap that one as well. I still have 1 10 pack remaining that I’m not going to touch until this is resolved.
@Eric_Inovelli I am disheartened to find that 19/20 of the switches I received match against the IEEE numbers you mentioned
38:5b:44:ff and 94:34:69:ff
I’ve pulled out the one switch that doesn’t match the rest and will try to swap that one into a far away place. I’m anxious to hear any news on what will happen to the rest of our switches? I’ve got 19 potential duds here, 3 installed already (2 that paired but dc’d and 1 that never pairs) and waiting to find out out fate before I bother installing another.
Understood, we are working as fast as we can on the issue.
8 Likes
I too have some routing and signal strength issues; installed 7/40 out of the switches I ordered so far, and am definitely seeing things like switches installed in the same gangbox not being able to see each other at all. Also, zigbee2mqtt (repo image docker.io/koenkk/zigbee2mqtt@sha256:f9ac7e87c4962fc3abd4c3520f03f5c84e21df69df481fb82716761709c4667c) is showing an LQI of 255 for every switch in the devices view (is that the self-reported value?), but in map view, these lqi values are as low as 21.
Here’s my list of IEEE addresses; LQI values are from z2m’s map view as the device list says it’s all 255s:
-
0x385b44fffee8ee53- offline. Should see one switch 20cm away and a sonoff outlet about 1m away -
0x943469fffe05c966- LQI 29 to a sonoff outlet 4m away; an exception to the rule in that it’s connected to its neighbor in the gangbox at LQI 116/110. -
0x385b44fffee8ed8d- the neighbor switch; LQI 4/76 to a thermostat about 4m away, on the floor below. -
0x385b44fffeee16c7- LQI 12/98 to that thermostat about 5m away, and LQI 38 to a sonoff outlet about 1.5m away. -
0x385b44fffeee1cb2- LQI 47/110 to that thermostat about 4m away, 29 to that sonoff outlet also 1.5m away. -
0x943469fffe05cb8f- LQI 109/137 to that thermostat about 1m away (seems like this is the one good switch I managed to put in, but it doesn’t mesh with anything else, not even the sonoff outlet 5m away). -
0x385b44fffee8e989- LQI 5/89 to that thermostat, no other connections to either switches nor that sonoff outlet that’s 4m away.
I have about 33 more switches to try, but it takes a long time and trips to a breaker box that is far away to switch them out and try them, so I’ll stand by and wait to see how diagnosis goes.
One thing that I would love to know - is there a way to reboot the switch without turning off power and without factory-resetting it? I notice that they pick up zigbee mesh candidates when power-cycled, so if there’s an easy way to do that, it would help me bridge the period in which we don’t know how to treat this problem yet.
4 Likes
You can power cycle them by pulling the air gap on the bottom left (little white tab) all the way out (should see it go dark) and then pushing it back in.
Those are all in the main blocks of ieee space being looked at unfortunately (tip if you’re looking for which to install, the little sticker in the box has IEEE listed as EUI if you want to check to see if you have others as well). I have a few of each and while I was able to get one to pair, it definitely has low LQI despite being very close to the coordinator.
3 Likes
Oh man, thanks to you and stu1811 for that air gap hint. I had already evicted that knowledge from my brain!
As for the sticker, I’d love to know if anyone has destructured the QR code on the front of the device yet & knows if it contains the IEEE address; that would make it much easier to inventory large numbers of them 
Edit: There are plenty of QR code-scanning apps that will decode the front-facing sticker just fine & the format is easy to decipher too: Fields are colon-separated, starts with a Z: and the second field is the IEEE address. That’ll make it much easier to find which switch is which.
1 Like
Yeah, you gotta yank that air-gap out pretty hard – initially, I was just getting it about halfway and thought all mine were defective.
Google Lens can read the QR code to get the IEEE address
2 Likes
Yeah I used a qr scanner on all mine last night, I recommend only bothering to install them right now if they aren’t from the impacted batch, otherwise you’re going to risk disappointment
There are a lot of listed IEEE addresses. Is there a consolidated list of which to look out for? I’ve got my switches coming today (fingers crossed) and would like to avoid opening the ones that may have issues.
2 Likes
These are the 2 main blocks that seem to be problematic -
94:34:69
38:5b:44
2 Likes
You have to open it to find out the ieee
1 Like
Update, I have one single switch with a non-affected IEEE address, 040d84fffe02b4ce; that is able to mesh with all devices next to it and gets pretty excellent connection quality indicators. It even gets a direct link to the coordinator, despite being the switch that’s the furthest removed from it.
I have 2 04:0d:84 switches out of 10 and these are the only ones with an LQI consistently over 50.
The others are in the known “issue” group 94:34:69 and 38:5b:44 and are usually below 50.
I also have a switch inside a double gang box along with a z-wave fan switch and it reports as 0 most of the time.