From all the experiences of the many different people with the Sonoff SNZB-03 motion sensor, we know it is really a crappy device falling off the Zigbee network regularly and totally unexpectedly. Even with a solid external power adapter you face those issues.
Most of the time you have to complete repair then the SNZB-03 sensor, which in practice is really crap.
Probably most of us want to completely dump those sensors in the trash…
But, has anyone any idea what the fundamental cause is of the regular ‘drop off the zigbee network’ behavior of these motion sensors? Is it instable firmware by Sonoff? Or are messages wrongly transmitted or received by those SNZB-03 sensors?
I never had issues with these devices falling of my network (conbee2/deconz)
My problem was that it has so ridiculously many false movement registrations, that the device is totally unusable.
I have testet about 5 sensors on my setup and the same sensors on a different setup with a sonoff!! usb dongle and ZHA and getting exactly the same issues.
Yeah, I think the main/most common problem with the SNZB-03 is false positives which makes them unusable.
Of course falling of the zigbee mesh is not less problematic but seems to happen not that much.
@Marc_Sway (or anyone else): does it make any difference regarding false positives when powered from a USB power adapter?
I’m curious too as to why this happens.
Power: These Sonoff SNZB-03 sensors run on 1 CR 2450 battery cell, 3 volts. This is already quite bad design since those motion sensors are known to be quite sensitive to generate many false triggers during the day and also at low temperatures (if you use those sensors outside), empty the battery quite quickly. Also, I noticed that when the battery voltage drops below 3V (for instance 2.95V) many of these SNZB-03 simple stop working. So, with these small (1 off) CR 2450 battery cell, easily within 1 week of working, the whole sensor stops property working due to an exhausted battery.
That is also the reason that any other professional PIR motion detector never runs on a single CR 2540 battery cell!
Thus the Sonoff SNZB-03 is badly designed from a power perspective.
For that reason I power those Sonoff motion sensors in and around my house now with wired power, via a 5V Apple USB power adapter and a stepdown converter to 3.3 V. This works much better, but I would even prefer to run on 4 VDC if I could to have some margin in respect to voltage loss over long wires:
With the wired powering from a stable power adapter it got a bit better, but what really reduced the number of false triggerings was the shielding of the PIR sensor itself!! That reduced in my case at least the number of false motion triggers with 50% or more.
Another note: since my SNZB-03’s are hard-wired powered, I can power off and on these devices quickly.
I noticed that once a SNZB-03 ‘dropped off the Zigbee network’, switching off and on the power again (I keep usually ca 30 secs in between), this helps to get the SNZB-03 ‘recognised’ again in the Zigbee network.
If the SNZB-03 is put in pairing mode and it pairs with Sonoff’s ZBBridge, is that pairing then happening via Zigbee or via perhaps Bluetooth?
For me it is unclear why Sonoff always is promoting that during pairing a sensor (like the SNZB-03’s) is very close to the coordinator (like a ZBBridge).
———-
I did some further experiments to understand and reduce the ‘drop of the Zigbee’ issues with the SNZB-03 sensors:
Since those sensors at my set-up are all powered hard-wired with a power adapter, I switch now every night the power off for 2 minutes for all these sensors.
Then after 2 min, all these sensors start repairing to the Zigbee network (I leave always on).
With this daily repairing actively of the motion sensors in the middle of the night. I noticed that totally random ‘dropping off the network’ events dramatically decreased.
My overall assessment of the SNZB-03 is that this device is a Not-for-purposed designed PIR motion sensor
It is cheap and small, but that comes with a price: unstable, quite unpredictable and bad performance with many false motion triggers, running quickly out of battery life and totally random falling of the Zigbee network events.
I had another 2 ‘drop of the Zigbee network’ events during last 24 hrs.
Each time I recycled (‘Off-On’) de power on the respective SNZB-03 sensor. And each time after that, this sensor came properly ‘online’ again on the Zigbee network within ca 5-10 sec and works fine again.
So, what is happening here? Why is the sensor dropping off the network and why does it quickly function fine again after reboot of the power?
Anyone any idea?
Next experiment I may do is soldering a slightly longer antenne wire to the SNZB-03.
I think there is enough information/proof in this thread (plus the one that Francis links to) that there seems to be a problem with these devices.
Normally I would be in favor of trying to get to the bottom of this kind of problem, not in this case.
The reason: I tried to get something done by contacting the manufacturer without any meaningful response.
In the past I did have another problem with a device and the manufacturer of it is active on this forum, so they addressed it immediately.
It’s their product, they make money out of it so maybe they are more active when the problem is out in the open.
I fully agree with you. In this case unfortunately Sonoff / ITEAD is not very responstive to (consistent) customer/product complaints regarding their SNZB-03 sensor.
Is this the typical Chinese manufacturer behavior to not really respond to customer product complaints or think: “the product is not perfect/high quality, but it is cheap, so customers just have to accept that the quality is too low / the device is Not (!)-fit-for-purpose”?
Thus to get this product really perform better there are 3 options:
Continue raising customer complaints at Sonoff/ITEAD, at large scale.
start a large ‘blame and shame’ campaign regarding the SNZB-03 via large (FD, Reddit, etc) social channels.
Or start a joint research/experiment project with a few smart techies via a GitHub project to try to identify the root cause and define a solution for the unsatisfying performance of the SNZB-03 sensor.
The other manufacturer I was talking about is also Chinese and there are also a lot of non-Chinese companies that have a bad reputation when it comes to after sales support.
My main reason for getting these SNZB-03’s was it’s size.
That they are amongst the cheapest is a plus but I would even have payed/ordered more if they were reliable.
We all know that a single, non-public, complaint is easier to ignore then a public one.
In this case they would be confronted with many cases about the same problem which could draw a lot more attention of anyone in that company and buzz on the net.
I would even just inform them in the first place to give them the opportunity to respond.
By posting the fact that basic functionality like not only responding when needed (false positives) and not falling from the network, you make it very simple.
The fact that this device seems also to be falling from the mesh regularly is also a PITA for those experiencing this, but that could start discussions about coordinators, integrations, smart home platforms, … which could make it more difficult to get any support.
I don’t have an account on any of those platforms (and want to keep it that way), otherwise I would have done it myself.
The 2 SNZB-03’s I have in use have never fallen of the network, more then a year in use. The problem I have that one of them gives false positives, that is why the third one I have stays in its box and does not get installed.
I believe the only way we get this SNZB-03 device improved, if we can find someone to open (and drive) a complaint campaign via a FB group, Reddit, or at a dedicated Sonoff/ITEAD user forum to really increase the attention at Sonoff/ITEAD, that they really must improve de product quality, since it otherwise is going to hurt really hard their own Sales.
Other option is that perhaps a few very technical people at this Home-Assistant forum would be interested and keen to work together in a GitHub project to really get to the bottom of the SNZB-03 issues and find a solution.
Perhaps someone at this forum capable and willing to drive this?
).