Aqara fp1 stuck on detected occupancy?

So…Does the solution suggested by crlogic really helped? I have the same issue with FP1 placed in bathroom. It’s faced in opposing direction from the door, being on the eye level. But there’s a laundry hanging to the left, right before vent, so it is possible there’s a really small movement there. But then again, shouldn’t FP1 being able to distinguish person?

Haven’t seen them advertise this ability. What makes you think that it should? Seems to be a common misconception. Perhaps the FP2 will…

I don’t remember actually. I think I may hear it in ‘Game changer’ video but it could be said as a suggestion for the delay in both detecting and clearing…or maybe that was the video with comparison of FP1 and DIY? I remember there was conclusion where DIY had some false activations when FP1 didn’t.

But then again, if it doesn’t do this, then why it needs so much time to set itself to detected? The only thing that comes to my mind are situations when someone would walk through but didn’t actually stay in the room. That would mean it don’t occupy this place…so the solution here would be to have two entities: one for occupancy (currently present) and one for detecting motion. In a way that second one would fire right when there’s a movement and right after it’s gone.

Very good question

I don’t have an FP1… …thought I saw on Reddit someone saying that the FP1 has two sensors reported just like you suggested. It doesn’t?

Nothing through ZHA at least. Z2M says that there should be additional sensor for approach event. But it’s not there…or I can’t find it at least because logbook shows this event firing (few examples):

MotionSensor05 Leave Left - Sensor event was fired
13:31:39 - 17 minutes ago
MotionSensor05 Approach - Sensor event was fired
13:31:22 - 17 minutes ago
MotionSensor05 Away - Sensor event was fired
13:31:19 - 17 minutes ago
MotionSensor05 Enter Right - Sensor event was fired
13:29:56 - 20 minutes ago

But to be clear - having this wouldn’t help here because Leave won’t fire the same way detected doesn’t change.

Small update about getting detected/clear states and about events:
I just read the manual through the Translator app (it’s Chinese of course). And Aqara states that sensor needs about 6s to detect presence (whether person stays or not) and it determines within 30s to be sure that person is no longer there. Positioning of the device is important. They even tell differences between wall mounting and celling/stand mounting. There are also some tips regarding detecting user when sleeping.
As for the events (enter, approach, away, leave), they theoretically should fire faster then occupancy changes:

  • enter should be PIR-like, which means as soon as sensors sees something,
  • approach/away is self-explanatory with additional info that the distance settins relates to 1m, 2m and 3m distance of detection,
  • leave still requires up to 30s to know if someone truly leaved detection area, which means it’s still useless for fast detection. It also fails to fire sometimes, not sure why though - the guess here is detection distance setting.

I couldn’t find exact info of how motion sensitivity works.

One thing that is really missing from ZHA implementation (and Z2M as well) is a way to build the map of the room for sensor to know some basic elements that are inside to help identify everything faster and more accurate. The example shows bathroom floor with toilet bowl icon in place. Manual states that there are more icons that can be set. I think the whole map building is related to the missing regions implementation, where you can define squares of detection.

1 Like

My wife had a Kleenex on her night stand that was fluttering in the breeze created by the ac and it was triggering the sensor. They are very sensitive.

The FP1 does need to be paired with a motion sensor for most uses. I find the device works great. Restricting its view with metal allowed me to exclude potential problems. Since a motion detector is needed anyway, the FP1 only needs to see the places where movement stops (couch, desk chair, bed).

I find it responds from the change of away to home surprisingly quickly.

I use Z2M

I write a litte automation for debug and testing purposes. I noticed that the presence_event changes when I press the reset_nopresence_status button.
So I wrote this automation that triggers the reset_nopresence_status every few seconds via mqtt.

It is not a final solution :smiley: But I really want to know If it give you (z2m users) the same results…?

1 Like

I just picked one up through Aliexpress and I’m having similar problems. Integrated through ZHA, it will detect presence within a few seconds but never clears. I tried placing it face down on the floor like some suggested but it still won’t clear. Does this indicate a defective piece of hardware?

Do you have a metal bread baking pan? Place that over it while in a concrete floor. Should really turn off after that.

I put it on a brick and covered it with a metal pot and it still won’t turn off

Yeah that doesn’t sound right. Is there a factory reset option?

No clue. The instructions that came with it are all in Chinese. I’ve tried un-pairing it and re-pairing it in ZHA multiple times and nothing has changed.

Just got mine today and had the same problem. Turns out I was using a cheap third party power adapter. Switching back to the original adapter or an Anker charger solved the problem.

1 Like

I wish the solution was that simply for me. I’m using the included adapter already, after reading your post I even tried a Belkin charger and it didn’t make any difference. It’s still stuck on “detected” all the time.

I think I may be on to something, I believe I have a manufacturing defect with my FP1. I ended up breaking down and getting an Aqara M2 hub to pair my FP1 with, it acted better but it was still getting stuck on detected all the time proving it was not an issue with ZHA or Home Assistant. On a hunch I took it apart and pulled the circuit board out of the shell. Lo and behold, it worked perfectly with just the circuit board exposed and aimed at my room. It was detecting motion and occupancy within a few seconds of entering the room and within about 45 seconds to 2 minutes of leaving the room it reset back to nothing detected. I theorize that perhaps the plastic shell might be milled out too thick and is somehow reflecting the microwave beam back to the sensor and inducing false readings. Unpaired it from the Aqara hub and re-paired it with ZHA and all seems well. As it seems the hardware is working fine I might try to just fabricate and 3d print a new enclosure.

3 Likes

Just came across your post, how is it holding up until now? With bare pcb

Still working fine, occasionally it takes slightly longer to reset back nothing detected, but on average it takes about a minute to reset on its own. I haven’t 3d printed a new enclosure for it yet so I don’t know if having something in front of the sensor will cause the issues to come back. Funny thing is that I bought a second one of these FP1s and it worked out of the box with absolutely no problems, so I’m quite convinced there is a manufacturing defect with the original one.

I also had the impression that it never cleared it’s state at initial testing when I had it just lying on the table.
Once I set it up where I wanted it, it worked flawlessly. I’m impressed.

I now have it plugged into the USB socket on my zigbee-controlled power strip, and have it working together with an Aqara P1 motion sensor. When the FP1 detects that I have left the room, HA turns off the USB socket on the power strip. When the P1 motion sensor then detects motion, the USB socket is powered on again, so the FP1 can monitor my presence.