For the past year+ I tried many different mmWave sensors. By the spec many (if not all) looked very attractive devices but practically they are too far from being reliable. Therefore they are really hard to use in automations and quickly bringing (otherwise very high) WAF down when I tried to use them. To begin with, many different Zigbee mmWave sensors happens to be either impossible to link directly to the C8 (Pro) hubs (C7 seems to be much better for the direct pairing) and/or extremely chatty. So, this way all they are absolutely NO GO. Finally my choices are: Meross MS600 (Matter), Linptech (Zigbee) and potentially Aqara FP2 (as of today Home Kit). Common problem for all these mmWave sensors - is really hard to find a good location/orientation in order to achieve acceptable sensitivity with minimal unwanted interference.
Initially Agara FP2 looked very attractive, so I bought 2 of them. Unfortunately even with the latest firmware I still cannot use them in automations. Original firmware produced too many ghosts resulting in absolutely useless automations. Latest firmware happens to be very slow for the initial response and still occasionally produces ghosts. So both are still in a drawer but I hope, eventually they will become useful.
Linptech and MS600 are very sensitive to the spinning fans, air flow from AC, etc. and they have a backside unwanted sensitivity. So, finding a good spot for installation/mounting is extremely tricky. And very community famous Linptech approximately once a month becomes completely unresponsive. Only recycling power brings them back on line.
Bottom Line.
At least in my case,
Meross MS600 so far is most reliable mmWave/PIR combo sensor.
Linptech is also good enough except for the annoying monthly disconnections.
Aqara FP2 is still useless but I am hopping eventually it should become OK.
Am I doing something wrong (specifically with FP2)?
To add another product to your "to-try" list, you can look forward to Inovelli's attempt to integrate mmwave into their light switches. This project of theirs has had its ups and downs, and what comes out of it (and when) who knows. But at the moment there's some positivity built around them finding some promise in a new sensor and they've targeted a February release....at the moment tho just something for you to ponder and keep an eye on. Obviously this solution would limit the places where you could deploy it, but perhaps it might work for some use cases and home layouts. The latest on the project can be seen here in their community forum.
EDIT: Or, wait, did you and I already discuss this in another thread? Sorry to clog up your thread with it if so!
Well, with all my lengthy experience with many different mmWave sensors I already ruled out this Inovelli Switch with integrated mmWave Sensor. The reason is - it will be absolutely impossible to adjust location and/or orientation for the sensor because the location is limited to only whatever electric box. It is very unlikely this will be a good spot for the built-in mmWave sensor.
I'm an original backer of the Inovelli mmW switch, but I'm now fearful it's going to be crushed under the weight of its own feature-creep...
The pie-in-the-sky stuff being requested (and considered!) in its Inovelli community thread is getting pretty outlandish, and if they attempt to implement too many bells-&-whistles (which Inovelli seems all too willing to do lately in order to be a differentiator), then it's gonna be an additional 12+ months for the firmware to stabilize (I'm looking at you, Blue 2-1 switch).
When I saw this first time, I became very excited and almost pre-ordered few. But very next minute taking in account all my not very pleasant experience with mmWave Sensors my internal voice told me: "this is not going to work" (for the reasons I mentioned above).
Backside sensing: Solved (where it was an issue w/two of my five sensors) w/a little bit of thin metal behind the sensors. In one case right behind it, in the other case on the other side of the wall from the Linptech.
Lockups: Solved w/Ewelink plugs rebooting the sensor on a weekly schedule. Not ideal, but works, and once in place is set it and forget it.
Fans: We haven't been running the fan in the office, which is the only room w/fans where the Linptech are installed. I did do some testing a while back and found that the Linptech are (at least at the settings that I was using) sensitive to fans. I have some ideas about managing it by moving the Linptech to the bottom side of my wife's desk attached to another thin piece of metal to inhibit the upward spread of its signals. TBD...
Worst case I might have to replace the one in the office w/something else if I can't manage fan issues there. But the Linptech has been SO GOOD at sensing my wife when she's frozen still for extended periods, that I hope I can keep using it. Overall happy w/the Linptech and sticking w/them for now.
This is exactly what I am doing with my mmWave Senors. But this very well known problem must be solved by sensor's manufacture(s), but not by (smart) end users.
Again, this is non sense and not a solution for the problem but just a band aid.
Now you have to have a switch for the sensor.
If I am not mistaken, only FP2 has a setting for the "interference zone". This suppose to mitigate a least interference with a fans. This is very nice feature but FP2 even with latest firmware is still useless. Still produces too many ghosts and now became very slow (most likely because of attempt to fix ghosts but this was not successful). How you suppose to use this sensor in automations?
I'm just responding to your question how do I like the sensors, and saying that even with the accommodations that I've had to make I do like them, and I'm now able to use them successfully in all of my automations. Caveat that I haven't tried to deal with the fan issues yet. Aside from the lockups, it does appear that the state of the art at this point is still a limited across the board, Even on some of the more expensive options.
Definitely a your mileage may vary situation and of course preferred that they worked better out of the box.
I do appreciate your feedback. Thank you for shearing yours experience. Basically I am trying the same things for "fixing" what must not be a problem(s) in a first place. Semi-good news - I am not alone facing all these mmWave Sensors child illness. Bad news - all these problems must be addressed by manufactures.
I've said it before, and I'll say it again... It's telling that we haven't seen any big legit names (Hue, Leviton, Lutron, etc) do mmW yet.
Until a serious company takes these seriously, we're left dealing with inevitable early-gen & oh-so-typically-inconsistent-AliExpress-quality-issues jankiness.
And I don't count Inovelli as a serious company in this space.
Never used any Aqara stuff, so no opinion either way. But I'd lump them in along with the rest of the AliEx names -- I don't consider them to be a higher-quality company than any of those others.
It's not my intent to just bash the AliEx stuff -- mmW or otherwise, you just gotta set your expectations accordingly and be ready to deal with some jankiness (that wouldn't tolerated by company like Hue or Lutron).
I have a Linptech mmW and I too have learned to live with its need to be routinely power-cycled. If that's the price of admission right now to have a reasonably reliable mmW solution, I'm OK with that.
Sure, I look forward to more polished mmW solution someday from a bigger name -- but for now -- dealing with some jank is just part of the deal.
I'm using several of the Aqara FP1Es paired with their M3 Hub and linked to Hubitat via matter. Have had excellent success with them so far. Excellent motion detection speed and the AI Spatial Learning seems to work quite well at mitigating false positives.
I've never had good luck with any Aqara products in the first place but I did give the FP2 a shot based on the early reviews. Admittedly it's better than any other Aqara product I've ever tried but it's slow to detect in the first place and has a tendency to show ghosts, as others have pointed out. I'm planning on just going back to a standard PIR when I have some time.
I've more recently dialed in my custom lighting app and haven't had a need to chase the mmW devices lately. For example, in my office I have two Iris v2 motion sensors and the lights no longer turn off unexpectedly. The more often the sensors get triggered the longer it takes for the lights to turn off. That has worked well for me.
For now the mmW devices just sit in a corner and look at me with contempt.
Before mmW sensors my sort of solution was/is to use multiple PIRs. Unfortunately even with 3 MS PIRs in a Bathroom occasionally timeout was around 15+ min. This is unacceptably huge delay for the turning Off lights. mmW Sensor (currently Linptech) solved this problem instantly. But unfortunately Linptech Sensors periodically becomes unresponsive. The only "fix" for this problem is to recycle a power. Sure, this can be automated but it is non sense (at least for myself).
As we speak I am testing Meross Matter MS600 mmW/PIR combo sensors. So far they are working just fine and hopefully will not be latching. But time will tell how good they are.
I am testing using @kkossev's suggestion to change settings on the Linptech to get the device to re-calibrate (similar to power cycle), but from the logs it doesn't look like the device actually re-calibrates when the setting changes. Going to let my test continue for now, maybe I'll be happily surprised, as this work-around would remove the need to power cycle. Still not ideal, but much better than needing the power cycle.