Sonoff Motion Sensors False Motion Detection

⚙️ Custom Apps and Drivers Custom Apps
63

No - I have absolutely no illusions or expectations of these devices working outside - nor do I need them to as the Eufy motion sensors do this accurately.

All of my specific and detailed testing has been inside the house in highly controlled environments to ensure I can observe and log the specific errors and check their reliability.

I do not believe I can have 10 Sonoff motion sensors that are false-triggering more than any other random batch owned by others.

My best guess is most people use these for non-critical purposes such as light switching and therefore are accommodative or don't really see the variety of errors these things can throw up?

Every one of my 10 Sonoff motion sensors will false trigger at some point within 48-hours, will occasionally latch 'permanently' in a 'motion detected' state making a T/F state check useless, etc.

And these devices are all in 'protective custody' for accurate false-positive testing purposes - not out in 'the wild' where conditions are not controlled.

Hence I have given up and I will do it another way.

64

Selling them? I have had great luck with them.

65

I've got one triggering camera snapshots so every error definitely shows up. However other than the sun issue mentioned before I don't think I've actually had a false trigger in the recent (6 month) past. The rest are used for lighting triggers so I can't speak to those but seem to work fine.

I've had a hue sensor do this occasionally on a very hot day.

It does seem rather strange although I have no better insight unfortunately. If you got them all at the same time from the same seller it's not beyond the realm of possibility that the whole batch had issues.

66

The first thing I do when I get these Sonoff motion sensors is use a kitchen knife and slightly bend the battery contacts so that there's a better connection to the battery.

I used to have a lot of false positives when nobody was in my apartment. I even thought it might have been birds or traffic triggering them. It wasn't. It was a slightly loose batter connection.

Since then, any issues I've had have been related to the hub or placement of the sensor rather than them dropping off or giving false positives.

Try the batter connection trick and see if it helps (it's been mentioned elsewhere).

I recommend these sensors.

3 Likes
67

Same here. I have about a dozen of these sensors, and they are rock solid for me. But I did do exactly the same as you with the battery contacts :+1:

68

I am reasonably confident I now know what is causing the Sonoff ZB motion sensor instability/unreliability issue - it is high ambient temperature related.

The devices went from what appeared to be stable for a couple of weeks - to totally unstable and useless - and have now returned to a reliable/stable state again.

I have the devices in a couple of unused upstairs bedrooms for testing purposes, but one pair is in a ground floor level cupboard that we don't open.

The pair in the cupboard have been the only stable pair.

It is summer where I live and we can get daily temperatures of >40C (>105F)

The unused 1st floor rooms can get to 34C on a hot day - the ground floor cupboard is a fairly stable ~22C.

We have had a mild summer - but daily temperatures recently got to 40*C - and have returned to being mild again. The Sonoff motion sensors have similarly gone from stable in mild temperature conditions - to highly unstable once the weather got hot - and then stabilised and became reliable once the weather became mild again.

So...it appears these devices become highly unstable with false triggers, etc, somewhere above a 30*C ambient temperature.

I have now left one pair in a 1st floor room as a reference pair for when the weather heats up again - and relocated the other pair to lower, stable temperature conditions on the ground floor.

I am fairly confident what I am going to observe once we get 40*C days again.

If, through testing, I can convince myself I have a 100% level of sensor reliability at <30C (or whatever I test to be reliable) temperature range - I will use a Sonoff ZigBee temperature sensor in the individual rooms to test an ambient temperature of >29C (for example) - and at that ambient temperature level - disable/ignore the Sonoff motion sensors as a reliable alarm trigger source - until the ambient temperature is lowered to a level where I consider them technically reliable again.

These Sonoff motion sensor devices are clearly technically problematic in higher ambient temperatures - but MIGHT prove to be 100% stable and usable once this ambient temperature variable is totally understood - and can be disabled/ignored prior to a point where they are likely to false trigger.

69

Unfortunately even outdoor hue sensors are susceptible to false triggering/latching at over 30C (see my post above as well as here Hue Motion Sensor (Outdoor) False Trigger in the Cold? - #8 by rocketwiz). This will be a common issue across the board as I would not expect to see temperature compensation built into these sensors' circuits like they are in proper security sensors like from Bosch.

I would be interested to see if the Eufy sensors perform better under these circumstances.

70

Unfortunately this was not the case for me.
It was random ones throughout the house with the temperature no higher than 25C.
I tried tweaking the battery terminals, as others have mentioned, new batteries from different manufacturers nothing seems to make much difference.
I even put one in a box and left it alone in complete darkness. It would randomly report every couple of days.

71

Agreed - mine still throw up false-positives occasionally at lower temperatures.

I circumvent and can make these sensors 100% reliable (so far so good) by always operating them in pairs and testing for false positives (both in a 'motion detected' state simultaneously) using eWelink Scene Logic before triggering an event.

My observations show the heat problem causes a sensor in a pair to false trigger and then errantly remain in a 'Motion detected' state for an extraordinarily long time (potentially hours) - and is in that state when the 2nd in the pair false triggers.

My T/F eWelink Scene Logic test is then defeated - because the logic test is true - both sensors are simultaneously in a 'motion detected' state (albeit errantly), making them unreliable and useless under those conditions.

72

I got rid of all mine apart from a couple which I keep playing with. I don't know why I just don't give up on them. :man_shrugging:

73

So am I now - I have a 35*C day coming up in a few days. I have 3 Eufy sensors outside and will closely monitor them for false triggers on the day.

74

I have 4 operating pairs and a considerable amount of summer heat to go yet.

IF in lower ambient temperatures I can get a minimum of 45 days (pick a number?) 100% reliability from each pair they are probably trustworthy - ELSE they all go onto ebay.

1 Like
75

I have 3 Eufy motion sensors mounted undercover on the outside of my house. 2 of these were subjected to ambient temperatures of around 40*C where they are located in the last 2 days.

None of the 3 Eufy motion sensors false-triggered at any time on either hot day - and all did sense movement when I occasionally tested them in the heat.

It's a very different story for the Sonoff motion sensors.

I now have 11 Sonoff motion sensors and this is how well they are going (they operate in pairs in a vain attempt to control their abysmal reliability)...

  • EVERY LAST ONE of the 11 Sonoff motion sensors false trigger a minimum of once per day - but more likely tens of times per day.
  • ONE will no longer initialise and operate.
  • TWO randomly disconnect from the ZBBridge and need re-initialising to jerk them back into operation.
  • FOUR - including two of the immediate above mentioned - remain in their 'Motion detected state for hours on end a wait until the 2nd sensor in the pair false triggers and then provides a false-positive alarm when eWelink does a True/False check.
  • I now have just 3 pairs of working sensors remaining in test mode - and purely for none-other than for interest reasons to see if I eventually get down to none - or maybe one pair that remains stable.
  • There is no way on earth that I would trust these Sonoff motion sensors devices to trigger an important event - even if one pair remained theoretically reliable. A notification - MAYBE.
  • And as much as I would like to flog them on eBay - I could not do so with a clear conscience - they are unreliable, untrustworthy, JUNK.

Not all ZB devices are junk however...

  • I do now have 8 Sonoff ZB window/door switches, and they are proving reliable so far (only 3 days) - and the battery is much better as it is a common, cheap 2032.
  • I have 9 ZB temperature sensors - 2 are located in my in fridge and freezer for triggering SMS alerts - and they work well even though shielded by metal - and appear quite accurate.
  • I also have 2 ZB switches for panic buttons and they work reliably from the occasional testing I do.

So I am dumping the useless Sonoff motion sensors for the 100% reliable Eufy motion and door sensors - using specific app notification names as a trigger for IFTTT Pro to THEN activate eWelink Sonoff switches and Scenes that I have setup to be 100% reliable.

I go out of mobile service range occasionally, so I am probably going to need to setup a mirror image BU phone and leave it home to ensure notification events can still trigger. (Still thinking this one out carefully and will have to test)

It's a slow process making sure your alarm is 100% reliable and is not going to annoy the neighbours and call 'wolf' I have decided - but would have been considerably faster if I hadn't not wasted time and money on the @%!& Sonoff ZB motion sensors.

76

Bummer to hear your experiences with the Sonoff motion sensors. I will say, I have grown quite fond of the Aqara contact sensors (they call it a door and window sensor). I know this is a different type of sensor, but they offer motion sensors as well. They pair directly to Hubitat, and for the contact sensors there is custom drivers which work amazingly. I would assume their are drivers for the motion sensor too.

Most, if not all of Aqara's products claim a 2 year battery, and there is a youtuber who did a study comparing a bunch of contact sensors, and this one outlasted them all by far. I haven't had their contact sensors long enough to claim battery life. I have had a temp and humidity sensor of theirs up and running in my basement for almost 2 years now (end of April will be 2) and it is still on its stock battery.

I assume you wanted cheap sensors considering that is really the only reason to buy the sonoff ones. Their contact sensors are very well priced, at $18 USD (usually with a 10% off coupon on Amazon). Their motion sensors are about $20. They seem larger in size compared to the sonoff ones, but if that isn't a problem then they might be another option to consider, if available in your area.

Again, I haven't used their motion product, so I don't know personally if it is any good, but the contact sensors I have are solid. Hopefully someone else here has used them, and can chime in on how they perform. Though, if you have gotten by well using the Eufy sensors, then maybe that is the one to stick with

77

Thanks for the response. It wasn't so much cheap sensors - I have accumulated over 40 Sonoff 240V and 12V Wi-Fi switches over the past 4 years and they have been highly reliable, as has the eWelink app and its Scene logic.

I stupidly extrapolated that sound design and high reliability to their ZB motion sensor range.

Now that I know how to reliably trigger eWelink actions from differently worded Eufy notifications to Android using IFTTT, I will stick with the motion sensors I know work with 100% reliability.

IF the Sonoff ZB window/door sensors prove 100% reliable over several weeks, I will then trigger further eWelink events off them - and at ~$AUD9 each in bulk purchases they are cheaper than the ~$AUD40-50 Eufy ones to put all over the house, otherwise I will fall back to the Eufy window/door sensors.

78

I ordered a 2-pack of the SNZB-03 sensors in May 2021. The one I set up immediately has been reliable to a fault. The other one, I set up 2 days ago triggers motion at an almost 79.5 minute frequency. The first night, I thought it was the ceiling fan, which I shut off. Now, it's sitting in a corrugated cardboard box, away from any heat sources, and still triggering regularly and reliably in the 79-80 minute range, in a locked room no one has been in for like 8 hours... I could sensationalize it and say that's a 50% error rate, albeit with a sample size of 2.

I may try the battery tweak, but since it's to control the lights in my home office, if I have to go back to speaking to the Echo, it's not a huge loss, just annoying.

79

Strange little buggers. I have over 50 of them. Some of them were just not working right at first but just a few days in they seemed to settle down, I have not had to replace a single one.

80

I have now bought these from various places and one by one they seem to develop the same problem. After a number of months (I am going to say 6 on average) they start cycling active / inactive randomly. They will do this a number of times in a row and then eventually stabilize to inactive. Let's say every 1/2 hours they start for 4/5/6 times in a row. This makes them pretty much unusable. Have thrown 3 away thus far!!

81

Try changing the battery on one of them, Sonoffs go through batteries pretty quickly compared to Xiaomi.

82

There is the possibility that cheap capacitors were used, and failed, creating the problem. I fix people's stuff all the time by replacing every cap in, for example, the power supply section of a flat-screen monitor or LED/OLED TV, and this has a very very high sucess rate. One can buy sets of nearly every capacitor value known to man in both tantalum and electrolytic versions, so if one is willing to get creative with leads and solder axials into radial circuit board footprints and visa-versa, one can do this cheaply for an entire neighborhood. I used to pull out a scope and hunt for the bad cap, but I have found it easy and quicker to just to replace them all blindly.

Now, how this could happen with surface-mount devices is beyond me, but solid state chips and SMD resistors don't have a lot of failure modes once they get past the "infant mortality" period.