This one?
No it's the Sonoff device.
?
That's the one.
Just reviving this to see if anyone is still seeing this.
I kept a couple back and joined one of them again.
After around a week it started giving false motion.
It's placed in a spot where it's not possible for it to register motion or any light changes.
Just wondering if it is my friendly poltergeist or if anyone else is seeing it. 
I had one that was getting false motion and returned it and got another. Seemed to fixed the trick. I have 2 of these devices and they are very reliable. Most likely you just got a broken sensor, similar to the one I had and returned.
I had about 15 of these working for some time with no issues then some started playing silly beggars.
Sold them all on ebay apart from a couple that I held back.
As said above just trying again to see if all OK. Appaently not, for me anyway.
I've tried putting brand new batteries in and this doesn't resolve the problem.
No big deal as I don't use them anymore but thought I would ask for peoples current experience with them.
1 Like
Very interesting then. Maybe I had a pairing issue with the faulty one I had? Not sure exactly what the issue for me was. I figured it wasn't working properly, and replacing it made me think it was just a bad sensor. Given the price for these, I assumed maybe the QC isn't as good as some of the higher end motion sensors sold.
Weird they worked good, and then just didn't. Maybe some code was added to the handler that is causing the issue? That would be my only thought if they used to work, and now have issues.
1 Like
I bought 11 of the SONOFF SNZB-03 ZigBee Motion Sensors - and all but one false-triggers at some point in the day. I was pretty disappointed with them until I worked out a solution to my problem.
To fix the problem and make these motion sensors 100% reliable you have to use them in pairs along with eWelink Scene logic - but my testing says you can eliminate false-positive triggers. I have setup 3x 12V Sonoff boards and eWelink logic to latch the board's relay 'ON' when a pair of motion triggers a positive signal (I don't have to keep checking anything that way, just the state of the board's LEDs when I think about it) - and they sit there inert for days until I deliberately walk in front of a sensor pair.
If you ask Scene logic to trigger an event IF both motion sensors detect motion simultaneously, it will not work, because Scene logic accumulates the motions triggers over time, not simultaneously. For example, you might get the first sensor false-triggered 3 hours before the second sensor triggers (most probably falsely) - but even though there is a huge gap in time, Scene logic does not know the difference - and executes the 'THEN' event on that 2nd trigger 3 hours later regardless - very frustrating!
Fortunately these motion devices remain in a 'Motion detected' state for 60-90 seconds (each device varies) - so you can use that latched-on 'Motion detected' state to your advantage to make these sensors 100% reliable. (This delay also means they won't sense movement again until they change back to a 'Motion undetected' state - but this is not an issue as you should be reacting to the first detection as you will be confident you have a 100% positive trigger. You can easily retest for movement once the state reverts to a 'Motion undetected' state again.)
So - to eliminate false positives - this is what I did...
- FIRSTLY - program a True/False test Scene where in 'IF', Motion Sensor 1 and 2 are set to have ALL 'motion detected' conditions met (both must be in this state).
- My 'THEN' in this T/F test process is to turn 'ON' a noisy alarm siren using a 12V Sonoff relay board 'IF' true'. (Hence I am very keen to get 100% reliable positive triggers)
- What this T/F test is doing is checking that the state of BOTH motion sensors are at this specific moment in time in a 'Motion detected' state (using that 'Motion detected' state holding for some time as mentioned earlier) - and IF true - it means a heat source change was positively detected.
- SECONDLY - you then set-up an initial motion detection test in your scene logic where in 'IF', Motion Sensor 1 and 2 are set to have ALL 'Motion detected' conditions met so it triggers on that 2nd event - BUT - we know the trigger could be false.
- To determine if the trigger is T/F - you 'THEN' add that T/F test Scene we set-up in 1, 2 & 3 above - and it instantly tests the state of both sensors and IF both are in a 'Motion detected' state, 'THEN' my set-up it will activate a siren via a Sonoff 12V device.
- So anyone walking into a room for example, would trigger a TRUE test and set my alarm off within a few seconds using the above Scenes.
- I have been extensively testing the above for a couple of weeks and have had no false positives to-date.
- I have now placed a pair of motion sensors outside to test if they function just as reliably in that environment - only 2 days - but had one positive trigger at ~3am this morning - both sensors were triggered at the same time - but I have a 100% reliable Eufy motion sensor to double-check and it did not trigger. (The Eufy motion sensor is useless outside of its own closed Eufy alarm system - but it can be adjusted for sensitivity and made 100% reliable in outside conditions even though they are not made for external use.)
- I don't have pets - but best guess is that a bat, rodent or similar triggered the event - and having no sensitivity setting - it likely makes these sensors useless for external use.
I hope that helps someone.
Its good you found a solution, but I wouldn't buy 2 just to make it work personally. For the price of two, I could just buy a different brand sensor that was more reliable for similar money. Don't get me wrong, if someone has a bunch of these already, and they aren't working properly, then this is a solution to get around that. Though, double the batteries sort of sucks, and it would probably be cheaper in the long run to just buy another sensor, than batteries to power 2.
I guess it has been a while since I last posted, so I can provide an update on my 2 sensors.
Number one is working amazingly. It triggers lights in my master, and probably sees 30+ motion events daily. I don't think I have even swapped the battery on it either.
The second one I moved into my garage, and use it to keep my garage lights on if I am in the garage for an extended amount of time. Honestly, I haven't checked in on it, since it has been cold I don't spend much time in the garage. May be working properly, might be dead.
Totally agree - I wouldn't buy 2 to do the job either (or probably one 
)- but I have them - and was determined to use what I had available to invent a solution.
For future reference, what are the better motion sensors (and any other sensors) that work with the Sonoff ZigBee bridge and eWelink please?
I ended up selling most of mine on ebay.
I kept 4 that I keep playing with and still get random motion detection for some strange reason. I just keep trying.
What did someone once say.
The first sign of madness is keep doing the same thing but expect different results. 
The best was I have found is put 2 in the same spot and use Zone Motion Controller set up to have both active to generate motion.
1 Like
I do the same except I space them a bit to provide better coverage. It has completely eliminated the cats from setting them off.
2 Likes
UPDATE - these Sonoff SNZB-03 Motion Sensors are UTTER CRAP 
After reliable operation, I am now observing/experiencing multiple different sensor failures...
-
Two of my motion sensors are holding their 'Motion detected' state in what appears to be permanently, so the next time the 2nd sensor false-triggers and the eWelink Scene tests for a true/false state - of course it is TRUE - but it ain't F>>N USELESS!
-
I have had 2 sensors stop communicating with the ZigBee Bridge and require re-initialising - again F>>N USELESS!
-
Two of the sensors out of 10 have now failed and will not initialise - a 20% failure rate is once again - F>>N USELESS!
I have now given up. Sonoff knows how to make very reliable switching gear - been using many for years.
SONOFF ARE COMPLETE DESIGN AND TESTING DROP-KICKS WHEN IT COMES TO THEIR ZIGBEE ALARM SYSTEM IMPLEMENTATION.
SAVE YOUR MONEY - AVOID the Sonoff SNZB-03 Motion Sensors - THEY ARE JUST USELESS PIECES OF ELECTRONIC AND PLASTIC JUNK DESTINED FOR THE TIP SOON AFTER PURCHASING!!
I have in excess of 100 motion detection devices maybe 20 different brands. I have had bad and good luck with several brands. I agree that motions act differently depending on the driver used. In the end I feel that my protection mesh of additional motion sensors ultimately creates a more secured area. Combining the motions also cut false motions to nothing. I try to use 3 for each room and they provide complete room coverage and my animals just do no set them off any longer.
My C5 HE hub seems to have no issues with the Sonoff motion sensors.
8 or so C-5 and 6 C-7's with no more sonoff issues than any other brand that I have. (12) One nice feature that they all do not have is present or not. If you change the battery after a not present warning I never seem to have any issues. Battery reporting as on most battery devices does not seem to work for me!!!!!
2 Likes
Setting off extremely high dB sirens near many other residences with 100% reliability is my design brief.
I cannot afford to have a single false trigger error if I want anyone to pay attention if my alarm ever goes off - let alone annoying the neighbours and breaching noise pollution laws.
The Eufy security alarm I use has motion sensors that are 100% rock-solid – they have NEVER false triggered - and I use them outside even though they are not made for that purpose - and is elegantly simple for my other half or other family members that I share the app with if we are out of mobile data range to monitor and control the system.
The problem with a Eufy alarm system is that it has a closed eco-system, a quiet siren with no breakout connector and no way to talk to IFTTT to trigger something useful like warning SMS’s and to trigger DIY ear piercing sirens. (I don’t want to open the Eufy control box and breakout a trigger as it will stuff the warranty).
I wanted to use Sonoff motion sensors and Sonoff ZB magnetic reed switches to trigger 3x high dB sirens. I don’t have the Sonoff ZB magnetic reed switches yet – but after testing these Sonoff motion sensors, I am very confident they will be unreliable, prone to failure, false-triggering, junk.
So, I have decided to use the highly reliable Eufy system elements I already have.
I am now building an electronic sound-sensing trigger system that will pick up the sound of the Eufy siren and latch a Sonoff wi-fi switch ON to get around the absence of a Eufy external trigger. I can then do nearly anything I want with eWelink Scene logic, highly reliable Sonoff switches and IFTTT.
The only useful Sonoff ZB elements are their temperature sensors and the 3-press switch which I will use for a panic button – SO FAR they have been 100% reliable.
I have totally lost my sense of humour with the Sonoff ZB motion sensors (and pre-emptively the Sonoff ZB magnetic reed switches) – I am just fighting too many electronic latching faults, device failures, connection failures and false trigger variables to ever trust them.
I have about 20 and they have been rock solid for me.
You are expecting cheap sensors to perform high reliability tasks in an outdoor setting (which they are not designed for) and then calling them junk when they can't deliver. The design of the dome makes them completely unsuitable for this use case.
For the record I have a few of these. My experience is that yes they are prone to (occasional) false triggering when used in an outdoor space (which they are not designed for) where for example there can be rapid changes in light/heat levels due to the movement of the sun and/or clouds.
This is mostly IMO due to their capture dome - the same comment applies to all similar style sensors regardless of brand. I might also add that OOB the sonoff's quite possibly will have partly depleted batteries, even thought they are not active (who knows how long they have been sitting around and also whether the included batteries are cheap rubbish - quite possibly). Changing the batteries even for a 'new' unit can make a difference.
My thoughts on increasing reliability are:
1 careful placement (this is IMO extremely important when attempting to use these outdoors as they can detect plants moving in the wind)
2 alternatively reducing capture area (I made a "snoot" for one of mine to reduce what it is "looking at")
3. in combination with (2) use more than one pointing at an offset and then only trigger a response if motion is detected in all of them.
