I’m starting to see a pattern in my devices’ LQI - it’s high and stable through the night, but when people wake up and start getting on their laptops all my devices’ reported LQI starts bouncing up and down.
Things I have tried:
Zigbee channel energy scan on my SMZB-06M, showed 25 as the lowest energy so I started there before pairing anything. 15 and 20 were pretty consistently higher.
WiFi Access Point, was set to “autoselect channel” which looks like it had picked 6. I revised that to stay set on channel 1. I saw the “energy move” on the Zigbee channel scan to concentrate more in lower Zigbee channels, but there’s still something hitting pretty hard on Channel 26 (but not 25).
Even after this, I’m still getting the bouncing LQIs… the devices are always responsive when I ask them to do things, but every once in a while my door sensors will unpair themselves (I have had 3 sensors installed for about a week now, two of them have dropped twice).
I had a “weak link” in the mesh, so I tried adding a SMZB-06 in Router mode to bridge the gap. It did improve reported LQIs at night quite a bit, but in the daytime we have three laptops running on WiFi, two of them on 5GHz which I would think shouldn’t do anything to Zigbee? Anyway, when they get fired up, my network starts getting jiggy.
Picking a channel for Zigbee is a useless exercise. The Zigbee protocol utilizes frequency-hopping, so whatever channel you select will change as soon as the device has difficulty sending data.
Thanks for that… I haven’t dug under the covers of Zigbee protocol yet, I assumed with all the channel optimization talk that the channel you select is somewhat important.
So, it’s to be expected that when people start streaming YouTube over WiFi, I’m going to take a lot of hits to LQI, but that seems to mostly be O.K. in terms of actual performance?
I need to take an inventory, but I have more than 75 WiFi devices, nine of them cameras, 20 or so Zigbee devices, a few Z-wave, a couple of 433MHz, and two Bluetooth thermometers. I really haven’t looked at my map or the LQI in a long time.
Not digging yet, but peeking under the covers of ZigBee protocol, I find that it is not wideband or spread spectrum, which makes sense with its low power goals.
Correct. Zigbee does not yet employ frequency-hopping. That was a misread of some specs.
Zigbee 3.0 does not explicitly mandate frequency-hopping, but Tuya, Microchip Technology and Nordic Semiconductor have incorporated dynamic channel switching to mitigate interference.
Yeah, I went with the SONOFF version of those, 12 scattered around, and I added a SLZB-06 as a router at a kind of wasp-waist point in my mesh.
For what it’s worth: I have a metal roof (roof is steep pitch: equilateral triangle, so 30 degrees off vertical), and where I initially installed my SLZB-06M ZHA controller the antenna was aligned vertically maybe 50cm from the metal roof. It has a good “view” of the room below, but… I had the opportunity so I relocated the antenna another meter away from the roof and oriented the antenna horizontally. Time will tell - I only did this a couple of hours ago, but some of my problem devices have showed better and more stable LQI in the past couple of hours than they ever have during the day when everybody is active on their wireless devices.
Zigbee does not yet employ frequency-hopping. That was a misread of some specs.
