What does it show between it and other devices on your network, etc?
1 Like
the bad switches the LQI would be much much lower mine don’t get above 100, except a neighbor report to the one in the wall in a double gang box with a good blue right next to it… Also if you have TI CC2652 coordinator the LQI shown may not be accurate as the different radio manufacturers all do different things with it.
3 Likes
Is there a go-to resource where different chip+antenna combos are compared? Would be nice to have a feel for how far off the numbers I see really are. My mesh used to run on Hubitat, which reported 250-255 for everything, even when I knew that must be wrong. Move the same set of devices over to a Sonoff Dongle-P with z2m behind it, and I get what seems like a more realistic representation: farthest, most obscured devices are in the 60s, while nearby routers are 130-160. Still, no idea if those numbers could be improved, since I don’t know the scalings used.
A bit to unpack here… The danger of using a hacked on switch… its a component on the zigbee radio. Call me naïve but I’m pretty sure we’re not playing with fire here, as long as the radio is all we’re messing with. As for my radioshack fat tip soldering iron, I retired it with my radio shack tone dialer sometime in the mid 90’s. 
As to whether it is worth trying. This is a big if, but if Inovelli doesn’t RMA them and I’m fairly confident, why wouldn’t I try? I saw some comments basically saying that if I was able to fix the switch then I shouldn’t be asking for replacements. My mind is blown by that sentiment - and I’m starting to feel like I’m chatting in an Apple fan club thread. A big part of home automation, at least for me, is the journey, not the final result. Most of the tasks accomplished in this hobby, while certainly not useless, take far more time to achieve than they’re truly worth. But we do them anyway… because we like to. The satisfaction of taking a few Blue switches that were headed for the trash bin and bringing them to life with my two hands… If you still think this is about saving 50 bucks I don’t know how to explain this any better.
1 Like
Do the affected switches work fine if placed right next to a good switch?
1 Like
The test switch and my coordinator are the only Zigbee devices I have at the moment; is there something else I can do to test or debug further?
If this is true that may be part of it for my case; I’m using this TI CC2652 coordinator. Are CC2652’s bad or just different?
2 Likes
Gotcha I wouldn’t be concerned, it’s hard to draw conclusions without more devices on your mesh, but that batch of switches is fine.
1 Like
Just different. Hey I made that device so stand behind it. The different radios just compute LQI differently and it’s hard to compare. when I was helping to try and figure out the issue I made sure to have an network of SiLabs/EFR32 only devices so that they were all the same family for the testing.
5 Likes
Fixing the resistor is whatever. You still have to take apart the switch to some extent which effects the whole switch. If you mess up and release the “magic smoke” inside the wall who knows…
I reassembled one and the paddle would get stuck on the air gap if you press up too hard just right. My point was just cause you can doesnt mean its worth it. Plenty of other time wasters in home automation to worry about as you say.
I use tubes coordinator, am mindful of wifi channels and still only 1 device out of 15+ is above 100 LQI but the location of it is more central and no devices in the same room. My network is still very reliable.
I think a lot of firmware issues are mixed in with these signal strength reports.
1 Like
In my experience it caused good switches to not connect.
From the image I saw the resistor looks to be just a selector for the RF output. It is incorrectly installed for the RF signal to go to the module pad and not to the RF connector.
I received an email about faulty switches and replacement. How do I go about that? I purchased 60 blue series switches. (6 x 10 packs) Out of the 60, 51 were faulty / in the batch mentioned in the email.
I sent in information on both forms including scanning each switch and including all the ones that matched the addresses in email.
Will we hear back from these or did they just go into the abyss? Will we receive email confirmation? And lastly, when will we receive confirmation of replacement switch order + delivery?
1 Like
When Eric gets time to review your submission. This is not automated.
Next batch of switches inst due until December.
3 Likes
Silly question but where is the form y’all talking about?
I thought Eric said not to submit it yet?
Is it this?
Well, I submitted now too. I never got an email. Hope that’s fine.
I’d like to strongly echo what some are saying: If there’s no harm, I’d really like to keep my defective switches. Maybe using them very close to my coordinator, try my luck with the 0201 resistor etc. If people want to RMA/send them back, fine, but do not enforce upon everyone just because some people shout for it.
I really appreciate the trust Inovelli puts towards us and I hope most (if not all) are honest enough and agree. As some people said, it is not trivial to fix and not worth the few bucks. Not even for a criminal. I would do this purely for tinkering, playing around and, if successful, being happy to have resurrected a dead switch and save it from a landfill.
1 Like
Neat! It came highly recommended and I’ve been enjoying it so far. PoE is super handy!
I wonder if there is any way to force a zigbee router to act as an end device? Presumably, this issue is caused by the “good” switches trying to route through the bad ones? Could be wrong here as I’d also expect the Zigbee protocol to consider the LQI across the entire route to the coordinator and mesh through the best paths.
Is there any way to use the defective switches as an aux switch?
1 Like
Not a complete aux but May be able to use it in a multi-way smart setup. If you can get the switch to bind anyways.
Ouch, 94 out of my 100 switches are bad batch. Going to be a rough set back. At least they are not in the wall yet.