On a quest to find the "perfect" mmWave device

TRS-80 · December 26, 2024, 6:52pm

Why?

At some point I realized reliable presence detection seems to be the cornerstone for so many other things in home automation.

Progression

I suppose everyone starts out with some PIR motion sensors. Until the lights go out when you are eating dinner, or sitting on the commode.

Then, as many did, I also became excited about mmWave. We came a long way since the original FP1 and DIY sensors, and yet years later I still find myself searching for the “perfect” mmWave device.

Criteria

Of course “perfect” is subjective, so here are my criteria (in descending order of importance):

Does not require any vendor “app” nor “cloud” nor “hub” (not even for initial provisioning).
Actually works:
1. Reliable presence detection.
2. No false positive/negatives.
I prefer Zigbee, but would consider perhaps Wi-Fi, if the device(s) work well.
1. I don’t have any Z-Wave devices nor gateway, so nothing against it, just not something I want to get into unless absolutely necessary.
Should not spam the network with messages.
Affordable.

I would prefer one (or more) consumer-off-the-shelf devices for the majority of my general presence detection needs throughout the home. Later, I may DIY individual device(s) for special uses (ESPHome + LD2410 for example).

I have spent an unbelievable amount of time researching this topic, ordering and testing units, only to still be searching for “the one”. Below follow some of my findings thus far.

Devices I already eliminated (and why)

I have yet to update this with every device I have tried so far. But I will try and do so as time permits.

Aquara FP2 - Eliminated because of requirement #1 (even though I have read many, many reports of how well they work).
Tuya Zigbee devices - I have researched (and bought) many of the more recent (and cheaper) Tuya devices. Many of these based on recommendations of the website Smart Home Scene (who seem to share my affinity for these inexpensive devices, lol!). I am still testing some of them (read on) however so far I eliminated the following from consideration:
- MTG275-ZB-RL - A ceiling mounted sensor I had high hopes for, but one of the worst spamming offenders, sending 60-70 messages per minute(!). I even contacted the seller (Wenzhi Factory Store) to see if the current units have received any updated firmware, but near as I can tell, they have not. Until this is fixed, deploying any more than a few (a couple?) of these might bring down your entire Zigbee network. Not supported by alternative firmware project like ZY-M100 are (see below), although there is an issue tracking that.

Kind of ‘meh’ devices

These are not so bad as to be completely discounted (could be used in a pinch) but there might be better options available.

ZY-M100-24G - One of many inexpensive devices available off AliExpress (maybe $15 at time of writing), unfortunately they spam the Zigbee network pretty bad (see table below). By now we have some mitigations for the ZY-M100 devices, although I have not tried that and probably never will as time is at a premium for me (please comment howevever if you have).

Devices I probably won’t try

Aqara FP1 - Have been around long enough now that I think they are well supported, even zones work without their app in Z2M by now(!). And price have come down into $30 range (on AliExpress). Although many people report issues with these dropping off the network. However, the FP1E seems to be essentially the same thing, except updated software (and certifications, and language) to be sold in Western markets, so I am looking more at that one now instead. Maybe one day I get around to trying the OG, but I wouldn’t hold your breath!
ESP based devices:
- Apollo Devices - In general, I find their stuff a bit too pricey ($35-37 per sensor, and then the mount is another $4.99, plus shipping). And we are still talking about something which comes in a 3d printed case. Yes they include additional sensors (which I don’t care about), and are made in USA, but I really have a hard time paying that much money for what it is. So much so that I will probably never get around to buying one.
  - MTR-1 - Smart Home Scene often recommends this one as “the best mmWave sensor”, FWIW (maybe they are friends, lol). This is also a multi-target sensor (LD2450 based) which is a bit different to everything else we are discussing here. I have read the LD2450 are not as good at still presence detection (which is my primary use case).
  - MSR-2 - This is more comparable to the rest of what we are looking at here, as is it a traditional (LD2410B based) mmWave sensor.

Devices I am looking at now

Currently my focus is on these devices:

Aqara FP1E - These are probably the best mmWave sensors I have tested yet! They update state changes quickly and accurately. Even the distance measurement seems accurate enough to base some automations off of (in pretty much every other sensor I have tried, this is just wildly inaccurate). They do not spam the Zigbee network with irrelevant traffic. The only down side is they are a little pricey. OTOH, they are readily available on Amazon, with Prime shipping no less (in fact, they are cheaper with Prime ($39 instead of $49)). The only thing that keeps me from recommending them is that I only have one, and have only been running it for a couple days. But I plan to order some more and keep testing them longer term and then report back my results. There is another dedicated thread for the FP1E where many other people have also been reporting good results.
Tuya ZG-205Z (round) - This is a newer device I think. Shows up as ‘TY_24G_Sensor_V2’ in Z2M. I only have one so far but can confirm other reports of messages only about every 5 seconds (which is not the best, but certainly not the worst). There seem to be 2 versions of this (or perhaps I am confused about the model number): a round, flat one with the same odd human logo on the front (like the ZY-M100-24G has), and an oval one that looks like Tuya ZG-204ZM except without the PIR sensor. I am interested in the former (reviewed here at SHS). I’m not super impressed with this one, but I only have one and only been playing with it for a day or two, I’m hoping I can tweak the settings into something usable for certain cases.
ZG-204ZM - This one is interesting because it is battery powered (2x AAA!), combining a PIR and a mmWave in a clever way to save battery. I have 2 units installed but still not sure what to make of them. One seems to perform a lot better than the other (I get inconsistent results in both spam testing, as well as detection performance). I have yet to determine if this is due to differing location, adjustment, or just typical Tuya shoddy QC. More testing is needed.

Might purchase in future

ESP based devices:
- Screek F2 - I know a lot less about Screek, but I like what I see so far. They are ESP devices like Apollo, but seem to be a little cheaper. People seem very engineering focused and working with the community to advance the state of the art. They have a lot of models, I may have mentioned the wrong model, I have yet to delve into their offerings in depth.
- Athom PS01 - Another one suggested by commenters to this thread. Less than (IMO, overpriced) Apollo (Athom is $29.49 at this time (2025-02-17) plus $5 shipping) and you get a lot more (charger, USB cable, mount) all included. Which is getting close to FP1E price, but of course this is ESP based and also has a light sensor (which is closely related to presence monitoring in terms of lighting control).
Tuya TS0601 - Mentioned by a commenter below as being serviceable in secondary situations. This is becoming interesting to me due to the price of the FP1E, which I think I would prefer in more critical areas. But then I could see a need for one (or more) other models which are less expensive but work well enough.

Spamming testing methodology

I fire up MQTT Explorer and start a timer. After some time passes (10 to 30 minutes), I jot down how many messages each device sent during that time frame and put it in a table and do some math to arrive at messages sent per minute. I have repeated this a couple times, and plan on doing more.

Spamming test results

Table is sorted by Msg/min. If you want to sort by other columns, copy into Emacs Orgmode and knock yourself out.

|------------+----------------------+---------+----------+---------+---------+------|
| Date       | Model                | ID      | Msg sent | Minutes | Msg/Min | Note |
|------------+----------------------+---------+----------+---------+---------+------|
| 2024-12-25 | Tuya ZG-204ZM        | zig-059 |        0 |      10 |     0.0 |      |
| 2025-02-16 | Aqara FP1E           | zig-063 |        3 |      13 |     0.2 | 6.a. |
| 2024-12-26 | Tuya ZG-204ZM        | zig-059 |       73 |      30 |     2.4 |    3 |
| 2024-12-25 | ZY-M100-24G          | zig-043 |       40 |      10 |     4.0 |    1 |
| 2025-02-16 | Aqara FP1E           | zig-063 |       74 |      15 |     4.9 |      |
| 2025-02-16 | Tuya ZG-204ZM        | zig-059 |      219 |      28 |     7.8 |      |
| 2025-02-16 | Tuya ZG-204ZM        | zig-059 |      148 |      16 |     9.2 |      |
| 2025-02-16 | Aqara FP1E           | zig-063 |       32 |       3 |    10.7 | 6.b. |
| 2024-12-26 | ZY-M100-24G          | zig-043 |      352 |      30 |    11.7 |  1,3 |
| 2025-02-16 | Aqara FP1E           | zig-063 |      363 |      27 |    13.4 |      |
| 2025-02-16 | Tuya ZG-205Z (round) | zig-065 |      137 |      10 |    13.7 |    4 |
| 2025-02-16 | Tuya ZG-205Z (round) | zig-065 |      375 |      26 |    14.4 |    4 |
| 2025-02-16 | Tuya ZG-205Z (round) | zig-065 |      210 |      14 |    15.0 |    4 |
| 2025-02-16 | Tuya ZG-204ZM        | zig-058 |      663 |      29 |    22.9 |      |
| 2025-02-16 | Tuya ZG-204ZM        | zig-058 |      402 |      17 |    23.6 |      |
| 2024-12-25 | Tuya ZG-205Z (round) | forum   |      300 |      10 |    30.0 |    2 |
| 2024-12-26 | Tuya ZG-204ZM        | zig-058 |      929 |      30 |    31.0 |    3 |
| 2024-12-25 | Tuya ZG-204ZM        | zig-058 |      399 |      10 |    39.9 |      |
| 2025-02-16 | ZY-M100-24G          | zig-043 |      578 |      12 |    48.2 |    5 |
| 2024-12-26 | Tuya MTG075-ZB-RL    | zig-060 |     1803 |      30 |    60.1 |    3 |
| 2024-12-25 | Tuya MTG075-ZB-RL    | zig-060 |      715 |      10 |    71.5 |      |
|------------+----------------------+---------+----------+---------+---------+------|
#+TBLFM: $6=$4/$5;%.1f

Notes for above:
1. In previous tests, I seem to recall this sending many more messages (but not as many as zig-060). Then I tested a few more times and could not reproduce (in fact, some times this one does not show up at all).
2. Here I did not test myself, rather:
  1. This review stated “about every 2 seconds” so I fabricated equivalent numbers for the table.
  2. And this comment claims “approx. every 5 seconds”.
3. During this test, The Missus was in the living room and kitchen.
4. First time actually testing this sensor myself. As for previous forum comments, whoever said “about every 5 seconds” seems to be closer to my results than the person who said “every 2 seconds”.
5. This one was not showing up during earlier tests today. Eventually I figured out the USB power cable had worked loose at the device end while I was adjusting the angle earlier. Once I plugged it back in, it was back to its old spammy behavior.
6. FP1E:
  a. With no activity, hardly sends any traffic.
  b. With activity, only sends what changed (motion or not, etc.).

Other (related) discussions

Besides specific devices, perhaps some other points of discussion:

Is the “network spamming” issue overblown? Smart Home Scene says he thinks it is. I am not sure what to think. Although it is probably less of a problem if you have a good coordinator, lots of routers, and generally follow good recommendations setting up a strong Zigbee mesh. However even then I think it will catch up to you at some point. I plan on deploying a lot of these devices, certainly the messages will all add up at some point and possibly even bring down the network (there are actually reports of this already).

Please discuss!

I would love to hear about any devices anyone has had really good results with.

And I plan to update this thread periodically with my latest findings (although, unfortunately, this may be few and far between).

Change log

2025-02-17:
- Add spamming test results.
- Move Aqara FP1 to new ‘Devices I probably won’t try’ section.
- Update review of FP1E, now that I have had a chance to play with it (only for couple days).
- Move ZY-M100-24G to ‘meh’ category, and update review.
- Update ZG-205Z review with initial test results, based on only couple days testing.
- Update ZG-204ZM review, and move into ‘current focus’ section.
- Add ‘Tuya TS0601’ for possible future consideration.
- Move Apollo devices to ‘probably won’t try’ and flesh out reasons why.
- Now that I have more testing results, I am also starting to include pricing into the comparisons.
- Add link to MTG275-ZB-RL tracking support in ZY-M100 alternative firmware project.
- Correct note in table (1.a and 1.b. supposed to be 6.a. and 6.b.).
- Addd Athom PS01 into consideration.

rouxdoo · December 26, 2024, 8:43pm

I have 3 FP2’s for larger spaces (kitchen/dining area, living room, master bedroom suite) and I simply love them. The ability to set up zones and automate separately off of that vs overall occupancy allows for a lot of flexibility - for example, overall presence detected → ambient lights on, specific area presence detected → area task light on. It nicely allows for a sense that the rooms respond to you in the space as you move around. I am not bothered by the app, it allows for a great way to map zones graphically. Requires plug access.

I have several Tuya TS0601 (little wired zigbee pucks) that I use for small areas and pass-through detection - bathrooms, approach to room entry, etc. I have not noticed any amount of unusual zigbee traffic or problem. Moderately configurable with the right quirk installed. Requires plug access.

I have a few esp32 boards using dfrobot_0610 sensors in project boxes that are great for rarely used mid-sized rooms - super sensitive response with a lot of configurability in distance (not as good as the FP2, though). They live behind curtains to hide the ugly boxes. Require plug access.

I have a couple battery Eve (matter) occupancy sensors watching exterior entry doors to automate alarm control (disarm if door opened while occupied inside the door). Not very flexible but solid performance as long as you watch battery levels.

I no longer depend on the handful of PIR motion sensors that I have deployed - just too lazy to remove them.

R0ot · December 26, 2024, 11:57pm

When you say 3 FP2 for big rooms do you mean 3 per room? Are you varying ceiling/wall mounted?

rouxdoo · December 27, 2024, 12:51am

I have one FP2 in my kitchen with 3 detection zones: general occupancy, stove/prep island/sink and dining nook - general occupancy turns on ambient light, specific zones turn on (or brighten) task lighting for the respective zones and roll it back on lack of occupancy.

I have one in the living room with 3 detection zones: general occupancy, couch/loveseat occupied, fireplace area - general occupancy turns on ambient light, specific zones…you get it.

I have one in the master suite with 3 detection zones: general occupancy, rouxdoo-in-bed, mrs-rouxdoo-in-bed - various scenarios actionable for night time, naps, wakeup, dressing, etc.

TRS-80 · December 27, 2024, 3:41am

Yeah, at first I thought the “zones” thing was a bit of a gimmick. but even in my setup I could see some applications. And your examples are giving me some more ideas. Luckily this feature is also available in some of the devices I am considering (e.g., FP1).

rouxdoo · December 27, 2024, 3:53am

I can tell you, the zones on the FP2 are no gimmick. Use the Aqara app to define and map the zones in real time - you can use the graphical map to add furniture, add doors and limits/extents/borders to the sensor response. You can even add exclusion areas for things like ceiling fans or kitchen cache racks. You can (using the app) walk around and see yourself or others as dots on the map of your room and fine-tune the zones you define.

You asked for the “perfect” mmWave device - I’m telling you that this is as close as you can buy today.

R0ot · December 27, 2024, 9:22pm

Did you end up roof mounting ?

rouxdoo · December 27, 2024, 9:58pm

Corner mount in the living room - excellent coverage.

The bedroom it’s on the long wall in front of the bed which covers everything when someone is in the bedroom but I do have a Tuya puck on the side of a cabinet facing in to the short entry hall to early-fire first entry occupancy events.

The kitchen is funny shaped so at a cabinet top corner facing kinda down - this leaves a tiny blind spot under it and in front of the microwave. Come to think of it that corner could use a tuya puck with a short tuned distance to cancel the blind spot.

TRS-80 · December 28, 2024, 11:59am

You seem to have missed the most important requirement.

TRS-80 · December 28, 2024, 1:02pm

Getting ready to order an Aquara FP1E to test, since I can’t seem to order an FP1 from AliExpress right now (they are being obnoxious about my credit card).

My plan is to keep updating the OP with the most relevant information as research continues. I just added some more recent testing results.

RonnieLast · December 28, 2024, 2:31pm

Did you also look at the Athom unit with ESPHome support? I have one for a year now and it’s good.

The best discovery that I made is that they can function well from behind wood - I have mine in my kitchen but installed behind the pelmet of the presses, it works well and has little loss of detection, except the IR sensor obviously does not work behind wood. It may be an option for some.

NRA · December 28, 2024, 2:50pm

I would not consider the FP1, I purchased 4 of them with very high hopes, they are virtual useless for me. We have ceiling fans and forced air climate control, and the three rooms without ceiling fans (bathroom, toilet and laundry) all have ventilation fans, so the FP1 always show presence. They have been relegated to the garage. I
did purchase 4 FP1E and they were a huge improvement. But both models suffer the same issue: They work for a while and then some sort of event occurs (power outage or a Zigbee device goes off line) and they go off line and can not be reconnected for weeks or months, all 8 are currently not included in my HA Zigbee network, typically I need to power down the whole house to have any chance of including them, so I have no enthusiasm to mess with them.

But I will be tracking your research.

rouxdoo · December 28, 2024, 3:02pm

That is one for the reasons I like the FP2 units. I’m able to map out interference zones such as ceiling fans, kitchen cache rack, etc. If there is an element causing false positives you can, in real time, view the detection graphically and mark it for later so it does not trigger next time.

Superdunder · December 28, 2024, 3:25pm

What about the Everything Presence One sensor?

Im looking to get a mmWave sensor for my living room as i soon realized the P1 is only good for low occupancy rooms…

peterxian · December 28, 2024, 4:19pm

I really like the Athom PS01. I’ve even made a few customizations to the firmware which is just awesome and easy to do — took less than two minutes.

TRS-80 · December 28, 2024, 5:05pm

Disappointing to hear, but you may have saved me some time. Which I guess is what this thread is really about at the end of the day.

I began to research the FP1E this morning, many people seem to think it is basically an FP1 except updated certifications (or whatever) to be sold in US or other markets outside China (recall the original FP1 was (and still is?) only sold in China), and/or simply some software updates while being essentially the same hardware. Which sounds plausible enough to me. And would explain the similar behavior between the models.

I have heard many of these work through drywall, walls (some times undesirably), but this is the first I heard through wood, interesting.

In my mind at least, I group all these ESPHome/Wi-Fi devices similarly. Most of reports I have read about performance (presence detection) have been good, some people even swear by them. I was just hoping to get mmWave on Zigbee. But as I said in OP, if I can’t I will have to start looking at these. Which I am about to start doing, I am so frustrated with the Zigbee options.

Many of these also seemed like cramming too much unnecessary stuff all into one device. But maybe that’s because I already have separate temp/humidity and PIR motion sensors (Aqara) which work well for me. I guess if you were just starting out, all-in-one devices like that would be more compelling.

But then we are back to numerous multi-sensors all reporting over Wi-Fi. Which sounds like a bad idea IMHO.

peterxian · December 31, 2024, 4:36pm

Speaking for my Athom, I’ve found all three sensors to be incredibly useful, e.g.: IF motion is detected AND the room is unoccupied AND the room is dark THEN turn on the lights. Note this required modifying the firmware so that presence detection is on a slight delay after PIR — thanks to esphome it was very easy to do that, but few if any Zigbee sensors give you that degree of customization.

Not knowing exactly what you have concerns about, but I’ve found that even with 80 devices on my WiFi network, these things use hardly any traffic and congestion is not an issue, as long as you avoid 24/7 streaming cameras and upgrade your ISP router to something with WiFi 6.

TRS-80 · February 15, 2025, 3:21pm

You raise a couple very good points. Customization, in general, and a light level sensor, specifically. As that is something I would also like to incorporate at some point into my lighting logic (for overcast days, etc.).

I am also experiencing some issues with Zigbee devices in general (Aquara battery based sensors like temp/hum, motion, etc.) falling off the network that I have as yet been unable to diagnose. Which is very annoying to say the least (much less hearing what The Missus has to say when things don’t work! ).

Yes, traffic/congestion are my concern. Although I hear a lot more people lately saying it’s a non-issue. Just something I heard about, never experienced first hand myself. So, maybe the issue is overblown.

Which is why I have not ruled out ESPHome based sensors, in fact I have all the bits to make some DIY ones already. Maybe your comment (and Zigbee issues) were the nudge I needed to pull the trigger on one or more of these commercial ones to try out. As I may have mentioned already, I have never heard anything but good about them.

TRS-80 · February 16, 2025, 7:03am

I’ve just spent the last few hours doing some more testing, especially the Aqara FP1E and Tuya ZG-205Z (round, aka ‘TY_24G_Sensor_V2’ as it shows up in Z2M) which are the 2 newest I’ve received since my last posts back around the holidays (the last time I had a chance to play with this).

At this moment I am probably the most impressed with the FP1E so far. It only sends messages when something important changes (motion, presence) and it seems to be very accurate and quick to update. And this was right out of the box, with no tweaking whatsoever (pretty much every mmWave sensor I have played with so far seems to need some amount of tweaking, usually a fair amount in fact, but not this one). Even the distance seems accurate enough to base some automations off of (this reading is mostly a joke on pretty much every other sensor I have tried).

I guess you pay for that though, these are currently $49.99 on Amazon (I am in US). But $39.99 if you are a member of Prime. Kind of pricey, IMO. But they do work very well. At least this one I have, I am thinking of ordering a second one just to make sure this isn’t just a fluke. And because I have a couple spots where I think some accurate distance might be helpful.

The Tuya ZG-205Z (round), aka ‘TY_24G_Sensor_V2’ did not do anything at all out of the box. I had to turn the radar sensitivity all the way up to get it to register anything. In fairness, I might not have it aimed the best. So more (performance) testing is needed here.

One thing I wanted to check right away was how much these new sensors spam the Zigbee network. Now I have been doing some of this kind of testing already, I just did not publish my results yet. So here those are:

|------------+----------------------+---------+----------+---------+---------+------|
| Date       | Model                | ID      | Msg sent | Minutes | Msg/Min | Note |
|------------+----------------------+---------+----------+---------+---------+------|
| 2024-12-25 | Tuya ZG-204ZM        | zig-059 |        0 |      10 |     0.0 |      |
| 2025-02-16 | Aqara FP1E           | zig-063 |        3 |      13 |     0.2 | 1.a. |
| 2024-12-26 | Tuya ZG-204ZM        | zig-059 |       73 |      30 |     2.4 |    3 |
| 2024-12-25 | ZY-M100-24G          | zig-043 |       40 |      10 |     4.0 |    1 |
| 2025-02-16 | Aqara FP1E           | zig-063 |       74 |      15 |     4.9 |      |
| 2025-02-16 | Tuya ZG-204ZM        | zig-059 |      219 |      28 |     7.8 |      |
| 2025-02-16 | Tuya ZG-204ZM        | zig-059 |      148 |      16 |     9.2 |      |
| 2025-02-16 | Aqara FP1E           | zig-063 |       32 |       3 |    10.7 | 1.b. |
| 2024-12-26 | ZY-M100-24G          | zig-043 |      352 |      30 |    11.7 |  1,3 |
| 2025-02-16 | Aqara FP1E           | zig-063 |      363 |      27 |    13.4 |      |
| 2025-02-16 | Tuya ZG-205Z (round) | zig-065 |      137 |      10 |    13.7 |    4 |
| 2025-02-16 | Tuya ZG-205Z (round) | zig-065 |      375 |      26 |    14.4 |    4 |
| 2025-02-16 | Tuya ZG-205Z (round) | zig-065 |      210 |      14 |    15.0 |    4 |
| 2025-02-16 | Tuya ZG-204ZM        | zig-058 |      663 |      29 |    22.9 |      |
| 2025-02-16 | Tuya ZG-204ZM        | zig-058 |      402 |      17 |    23.6 |      |
| 2024-12-25 | Tuya ZG-205Z (round) | forum   |      300 |      10 |    30.0 |    2 |
| 2024-12-26 | Tuya ZG-204ZM        | zig-058 |      929 |      30 |    31.0 |    3 |
| 2024-12-25 | Tuya ZG-204ZM        | zig-058 |      399 |      10 |    39.9 |      |
| 2025-02-16 | ZY-M100-24G          | zig-043 |      578 |      12 |    48.2 |    5 |
| 2024-12-26 | Tuya MTG075-ZB-RL    | zig-060 |     1803 |      30 |    60.1 |    3 |
| 2024-12-25 | Tuya MTG075-ZB-RL    | zig-060 |      715 |      10 |    71.5 |      |
|------------+----------------------+---------+----------+---------+---------+------|
#+TBLFM: $6=$4/$5;%.1f

Notes for above:
1. In previous tests, I seem to recall this sending many more messages (but not as many as zig-060). Then I tested a few more times and could not reproduce (in fact, some times this one does not show up at all).
2. Here I did not test myself, rather:
  1. This review stated “about every 2 seconds” so I fabricated equivalent numbers for the table.
  2. And this comment claims “approx. every 5 seconds”.
3. During this test, The Missus was in the living room and kitchen.
4. First time actually testing this sensor myself. As for previous forum comments, whoever said “about every 5 seconds” seems to be closer to my results than the person who said “every 2 seconds”.
5. This one was not showing up during earlier tests today. Eventually I figured out the USB power cable had worked loose at the device end while I was adjusting the angle earlier. Once I plugged it back in, it was back to its old spammy behavior.
6. FP1E:
  a. With no activity, hardly sends any traffic.
  b. With activity, only sends what changed (motion or not, etc.).

TRS-80 · February 17, 2025, 5:16pm

I have been able to spend a couple more days on this. I updated the OP with all the recent findings. There were enough changes that I felt a changelog became appropriate. I plan to continue updating the OP in evergreen fashion.

At this time, Aqara FP1E seems to be standing head and shoulders above the rest. But it’s only been a couple days. I plan to order 1-2 more and test them for longer time to insure consistent results.

However those are kind of pricey, so I will probably continue to order and test some of these cheaper ones looking for an alternative sensor which is less expensive to deploy in larger numbers in perhaps less critical areas.

I would like to reiterate that I have limited time (N.B. it has been months since OP), so if you have any experience with any particular models meeting criteria (whether I have mentioned them yet or not), by all means please discuss!