I was thinking of a temporary solution, tell me if it makes sense.
can we create a second zigbee network in ZHA, on a seperate channel, just to place these switches? that way they can function without killing the other mesh? or will this still cause problems?
FCS is Frame Check Sequence. It’s a checksum to verify the signal integrity.
I did some Googling and found that TI chipsets may report FCS in a way that Wireshark isn’t expecting.
Does following those instructions resolve the FCS issue?
I’m no Zigbee expert, but this feels like a dubious claim.
The bad batch essentially doesn’t have a good antenna, yeah? So signals both sent and received are going to be weak. This should be pretty similar to a device physically very far away from anything else on the mesh.
Communication to and from the distant (either physically or “effectively”) node may be patchy, but the network should find better routes for other devices on the network (assuming there are other routers on the network).
No, it just has a 0-ohm resistor in the wrong location so it isn’t “wired” properly and therefore isn’t getting proper power to the antenna.
It’s a hardware issue, but not the antenna itself.
I believe that’s more or less true, but some folks don’t have a robust ZB mesh to begin with, so these weak/bad Blues can cause more havoc in those cases.
If you have a really robust mesh, sure, you can give it a whirl – it’s not like it’s a safety issue pairing them or anything like that.
My point was simply that the radio in the switches just has a bad “connection” to the network. It’s not like it’s actively spewing garbage out, or corrupting frames being routed through it, or anything like that. The only problem with the switches is that they’re effectively “far away” from the rest of the network.
Before the issue of the bad switches was identified, I installed 3 switches into a 4 gang box. One of these switches is a good switch, and the other two fall into the bad range on the IEEE numbers. All three switches seem to be working properly. ZHA is reporting that all three have an LQI of 255.
I tested the remaining 9 switches (all considered to be bad) in a separate 1 gang box about 15 feet away from the nearest Zigbee device, and not a single one of them was found by ZHA.
So I am curious, are the two working “bad” switches causing problems on my ZigBee mesh (if so, how do I tell), or are they okay because they are within inches of a good switch?
i can tell you by experience - after installing 7 (7 bad out of 18 total installed) ‘bad’ switches, they essentially poisoned my zigbee network over a couple days. devices started using these bad switches as router endpoints, and they were not able to communicate upstream properly. example: a motion sensor right next to a bad switch was able to connect to the bad switch (confirmed by ZHA visualization), but the switch could not relay the information reliably to the coordinator (retry… retry… retryyyyyyyyyyyyyyyyyyyyy… maybe? maybe go through? mostly failing in the end).
i am by no means an expert in zigbee, learning as i go along - what i can tell you for sure is as soon as i removed these switches from the network, unplugged the coordinator for an hour, and let things rebuild over a day, these actions returned me to a reliable/responsive zigbee network for all devices. yes i did the ‘heal’ procedure before removing the switches, multiple times, while i was trying to figure out what was going on.
Man, that’s disappointing. The Zigbee route discovery protocol seems like it should prevent something like that.
https://www.digi.com/resources/documentation/Digidocs/90002002/Concepts/c_zb_AODV_mesh_routing.htm
You’d think route discovery requests wouldn’t successfully make it through a bad switch, or if they did, the path cost would be too high to ever be chosen.
Any feedback on how well the good switches are acting as routers? I am looking at installing these switches throughout my house and using them as the backbone of my zigbee network. Do the good switches make good routers? Are they playing nice with your other zigbee devices?
yeah i wonder too - maybe the LQI/RSSI values are being incorrectly reported by these switches?
hey i know these networks can be affected by so much, maybe there were other factors in my environment making it worse… but things have been rock solid ever since removing them.
they are awesome, no issues with the good batches and overall extremely happy with them. highly recommend
There is no update at this time. If you have submitted your data for the defective switches to the form Eric created, then you should be receiving replacements. The replacements are slated to be here at the same time originally announced - within about 2 weeks at this time.
We do apologize that we have not been able to send out a conformation email yet to everyone who has submitted to the form, however in between everything we have also had to move into a new headquarters space so we have been extremely busy with the move and are getting settled in this week. I would say after this holiday week things will be mostly back to normal and we will provide better full time updates. We are a team of four and two people will be out all week for the holiday so it leaves us even more short-staffed than normal!
Long story short…as long a you have filled out the form you will get a confirmation (before the replacements are here) and your replacements will be shipped as soon as we have them. Thanks for the understanding and patience!
Can we add amended IEEE numbers? I did not include some of the switches that I thought were working properly, even though they were in the IEEE number range. I assumed they were okay because ZHA reported an LQI of 255. However, looking at the network visualization, I can see that they are actually causing problems.
Thanks for the update. I assume ZWP will be handling distribution just like last time?
@Courtney_Inovelli @Eric_Inovelli @EricM_Inovelli First off, the switches are great (except for the bad batches glitch) and absolutely appreciate the transparency and handling over the forums , putting trust in the community to let us keep defective switches, ship replacements etc. I am very understanding this is a small company and in any moment I prefer a company who is transparent, puts customer requests first to any big company. I think this can truly serve as a good example to other companies (hello, RainMachine …)
But, please do send at least an email out to your customers. Not only to people who have filled out the form but to everyone who has ordered switches!
It is mind-blowing to me that this still has never happened! You cannot expect everyone to proactively be in this forum and monitor all these deep threads. I myself have wasted hours of back and forth until I found the thread and many other people in this thread are the proof that I am not alone. Everyone who ordered switches deserve to know that there was a bad batch by which he might be affected and you are trying to fix the issues to the best of your abilities. It is, in my opinion, unacceptable to use the forum as single information channel and expect people to find their way here on their own.
You would just need to paste one of your early forum posts into an email and send it out to everyone who ordered … possibly with a reference that more info can be found / is discussed in the forum right now until further notice.
Or just
Dear Customer,
It has come to our attention that some batches of the 2-1 Blue Switches are defective. At this point we are doing everything in our power to fix the problem and we are planning to replace the units.
If you think your switches are affected, please fill out the form (LINK).
At this point we are very busy of handling this situation and provide updates in the community forum (LINK).
Once we have resolved the issue, we will send out another email with next steps.
We truly apologize (bla bla bla)
Thank you (bla bla)
This does not make much effort but would have gone a long way.
There was an email sent back on 11/7, it links to the form (also in this thread) to submit, the relevant IEEE addresses, background info, etc. If you never got that, can check with support, but if you’re here also then it wouldn’t be anything new.
I have never received such an email.
Which worries me more: Who else did not receive any email? And why not? Which other communication have I possible missed?
Thanks for the update. Assuming it will also be shipped in a way where Canadian customers won’t be dinged for taxes a second time on these. I believe there are a few ways to do this such as marking the shipment as warranty replacement or similar.
And here’s a wishful thought: maybe this time we ship Canadian orders first? 
I received that email.