I should have clarified, I'm using XCTU to determine that those two GE outlets are routing no end devices. In XCTU they only have links to other routers and the co-ordinator. Which really puzzles me, because there are end devices only a few feet away.
Just this morning I had another instance where one of the two centralite outlets that do most of the routing dropped off the network again. The cost on the GetChild page was 1 from hub > outlet and 0 from outlet > hub. I power cycled the outlet and everything starts working again. I'm still stuck on why the hub says it has a good link to the outlet but the outlet says it has a bad link to the hub, Would an outlet be susceptible to certain types of interference that would not affect the hub?
Cost zero == Centralite link won't be used... Can you try re-locating the outlet? One can only speculate why RF link in that direction is worse. Also, it might be interesting to see what the Zigbee logging page (in Zigbee settings) shows for RSSI on that device. That figure will include noise which kind of muddies the picture, however.
Edit: actually the RSSI shown on that page won't shed any light on the signal strength seen by the Centralite as it is hub centric
I used my Xbee to do a Spectrum Analysis, the first one at the location of the HE hub and the second at the outlet that dropped off this morning. Both are based on 100 samples.
Bedroom (hub)
My HE Zigbee channel is 15, which corresponds to the space occupied by wifi channel 4. So is it correct to say that the higher dbm represents less interference on the channel? My test showed a -52dbm at the hub and a -48 dbm at the outlet for (wifi) channel 4, which I'd think would be ok.
Does the spectrum analysis give enough information to say which is the better channel to pick for HE?
That outlet is currently being routed through the Xbee according to GetChild:
status:Active, age:64, routeRecordState:0, concentratorType:None, [Living Room Purifier, 6D5C] via [Xbee, 57A7]
The outlet is serving a function there other than just routing (controlling a dumb purifier), so if I move it it creates another problem. The other centralite outlet that does most of the routing (7E9A from an earlier post) has the same symptoms - incost of 1 and outcost of 7. It is located on a lower floor but closer to the hub.
When I moved the Xbee near the Living Room Purifier outlet, the Xbee showed an incost of 1 and outcost of 1 back to the hub. At least on the surface, it looks whatever is affecting the centralite outlets doesn't bother the Xbee.
Maybe I should just put Xbees everywhere!
Was just editing my post when I saw yours; as the logging page shows the hub's view it doesn't really shed any light on what the Centralite is seeing (this occurred to me after i posted). So as expected from the low inCost, the -61 /-62 is fine (for the hub).
multiple centralite outlets, ~20 ft. apart, main routers in the zigbee mesh, all show poor link quality back to hub although the hub has very good link quality to the outlets
an Xbee placed in the same location as one outlet shows very good link quality back to the hub
outlets will periodically go offline but I cannot link that situation back to any specific event
I could try changing the HE zigbee channel but there isn't much evidence that the current channel (15) is causing problems (i.e. no smoking gun).
I can try moving the hue bridge to another channel to create more separation between it and the HE zigbee network
I may also try contacting centralite to see if they have any suggestions. I know they manufactured a lot of devices for samsung and lowe's/iris so I'd think these outlets, or similar, are being used throughout the hubitat community. Maybe there are other users who have experience with the 4200-C ????
I'm not sure this means anything. I may be mistaken, but I recall someone telling me their supplier or something in the chain had changed after the bankruptcy filing. They may not be the same devices we loved before.
They (4200-C) do look very different externallly than the 3200-G series Centralite plug (which externally looks just like the Iris V2 smart plug w/Z-Wave repeater). The datasheet for the 4200-C shows it to have a stronger transmitter than the 3 series plug (maybe they dropped the ball on the receiver side-- or perhaps it lends itself to being placed further from the hub, which itself isn't famous for having a strong radio).
Maybe. There could be component changes that are affecting reliability.
I'd be interested to see the map. I'm often very surprised by the paths my repeaters take.
[Edit] Sorry Tony, misread your post the first time.
But it is interesting that the transmitter strength is quoted at +18dbm on the 3200-G, vs. +20dbm for the 4200-C. If the receiver sensitivities are the same, it can be out of range from the hubs's transmitter, yet it may still be in range of the hub.
Absorption/reflection from objects and building materials sure does have odd effects. I was so sure that a Trådfri repeater I had just 8 feet from the hub on the other side of a wall would join the hub directly. Nope. it hooked onto the other Trådfri outlet about 30 linear feet away upstairs. It's the one upstairs, equally far from the hub that is first in line
