Yesterday after leaving the house - I got messages on my phone that there was motion in a room while no one was home and the HSM was in armed mode.
I have a robust camera system in and out of the home and immediately turned on my remote camera to that room. Nothing there. I got at least 10 alerts over a 90 minute period until I got home and the system auto-disarmed.
When I got home, I reviewed and found that my Samsung Motion Sensor battery was at 42% - the room where the 'motion' was detected.
I swapped the battery - and the false reporting stopped. Whether this is a driver problem, an electronics problem or a battery problem is not the point - a device shouldn't kick false detections! If the battery is dead, die for craps sake! don't fake report, don't tell me the battery is 40% when it's obviously not (or is it? another point of distrust in my environment - battery reporting...) I swear - it's two steps forward, 1 step back! The HA industry has got to do a better job with devices - there is no way to try/test every known device, in every iteration in attempts to find bulbs that work or outlets that do Refresh right, or sensors that misreport batteries.
I think the HE list of supported devices should be made more robust to include ratings and comments that are helpful in determining the best options to meet a users needs.
There's your answer - the magic 42%. Mine never report anything lower than that, so although when they first report 42% they probably really are 42%, they carry on getting lower without telling you. At some point they start giving false signals which is probably their way of expressing distress. I won't be buying any more of this brand of sensor.
As you've discovered, battery reporting is notoriously unreliable. I have an Iris v2 motion sensor that has been reporting 25% for days but is still behaving fine; I have seen others die when the reported value is closer to 50%. My preference is to monitor device activity instead of battery level (to find devices that are no longer working, likely becauase of a dead battery), though I'm also experimenting with ways to monitor battery level, too as another possibility, just not the one I rely on the most--and of course activity would not have helped in your case. I think this issue has to deal a lot with the nature of the discharge curve of the lithium batteries most of these devices use (not linear, with a plateau in the middle of their useful life) and may also have something to do with how at least Zigbee handles reporting (people who know more than me have said that the specific level is "binned" into a tier that may not be the exact voltage reading, but I have no personal knowledge on that).
I've seen sensors of various brands (all Zigbee, though I'm not sure if that really means anything, as I don't use many Z-Wave motion sensors) that begin sending spurious motion events when their battery gets low. Looks like you've discovered that, too. 
For security purposes, that much you might at least be able to avoid by grouping sensors together--for example, I use mostly Zone Motion Controllers for HSM, with HSM needing two sensors in the same room or adjacent rooms in order to "trip" the intrusion alert. It helps with both this (rare) problem and (more commonly) things like the sun or a hot air register blowing on a curtain and tripping the PIR sensor. Chances are someone who breaks in is going to walk around and not just stand in one spot, so I figured that was a good tradeoff (plus I also have contact sensors).
On my monitored burglar alarm, I went with all contact sensors, along with a glass break and smoke detector. Not z-wave/zigbee. I found the motion sensors not that dependable. We had a state trooper prowling around our place when we weren't home because the suction cup holding up the wreath in the breezeway gave way. This was even between kitties. So, no more motions for alarm for me.
Like @bertabcd1234, I use a combination of inactivity and battery level monitors to determine actual battery status. Unfortunately, there doesn't seem to be a reliable way to monitor level.
Same here, though I have semi-proprietary Honeywell motion sensors throughout the house and they have worked flawlessly, even with kitties. I have everything integrated with HSM but I find the purpose-designed and built security components more "trustworthy" than home automation sensors.
This whole Home Automation (HA) market is not ready for prime time, and is certainly not ready for home security, fire detection, or anything else that is very important to you. You would think that by 2021: Z-Wave would be stable, reliable, and interchangeable, but it is not. That battery operated devices would have a setting to differentiate between alkaline and lithium batteries, and an option to "beep" when they are running down, but they don't. etc.
The devices, Hubs, and battery operated and mains powered devices, are just not robust enough. They are all hobbyist grade, and rely on consumers to beware and debug the devices. And this is not cost driven. My UL Listed wireless smoke alarms cost about the same as a Z-Wave switch.
My UL Listed alarm system, mostly hard-wired, but also wireless water and smoke sensors, has not faulted or failed in any way in the 25+ years since I installed it. Not one firmware or software update.
Certainly, you cannot compare a Home Automation system with an Alarm System... or can you?
This was my first thought. It is bizarre to me that people expect professional-level performance and yet buy consumer-grade components.
I would never rely on a security system that wasn't UL listed. The UL requirements take into account stability and other things. Zigbee, Z-wave, etc are not reliable enough to use for security. In my experience anyway. Batteries don't last nearly as long as in a Caddx or Honeywell wireless sensor. My garage door tilt sensors have had the same batteries in them for 8 years and are still working fine. I get that it's the only option sometimes when you don't wanna run a bunch of wires or if cost is a concern. What would be awesome is a GE Caddx (or Honeywell, or Cryptix) radio that can talk to the hubitat so you can use Caddx wireless sensors with HSM. You can already use the Connected stuff to do wired sensors and connect them into Hubitat. I don't think they do wireless though.
I have an Elk M1 Gold security panel, and that's what I use for security. It's wired to every door, every window, motions, floor flex sensors (they are called something else), and I have a few GE Caddx wireless sensors for glass breaks and motion. In 8 years of running this setup in this house (and 5 years at my previous house), I've never had a false positive. It just works.
Plus, the Elk is connected to my Hubitat, so I can use those security sensors for automation tasks and I can control the Elk with the Hubitat.
I do have a bunch of zigbee motion sensors, but those are only used for automating things like turning on and off lights, etc.
I don't think you can - home automation stuff should be there to augment those kinds of existing systems rather than replace them. Doing things like providing extra alerts, turning stuff off, maybe shutting of water mains etc etc.
Also I agree with your thoughts on batteries. I use usb powered sensors where I can and also hack battery powered ones with battery eliminator plugs where possible.
Right! I use HSM and Envisalink but my Ademco panel handles anything directly related to life/safety. Smoke detectors, CO, door and window contact sensors, sirens, monitoring, etc., are all done by the Ademco panel. HSM gives me a couple of Ring keypads for convenience, handles water sensors and main water cutoff, blinks the lights, and gives me a couple extra sirens/strobes throughout the house. But my zwave and zigbee meshes could melt down and my home security / fire system would continue to operate.
Which kind of begs the question. I have read that Zigbee has found a home in hospitals for wireless monitoring. Either there's a much more robust version of ZigBee, or they use it for non-critical monitoring, or I hope to hell I don't get sick.
Zigbee use in hospitals is probably not based on the ZHA 1.2 or even Zigbee 3.0 profiles (someone please correct me if I’m wrong), so that may have something to do with it. Probably even more important, the mesh is much more carefully planned and maintained.
I agree with those that have pointed out that multi protocol home automation hubs aren’t intended for use as a home security system.
this is exactly why I sacked-off the battery-operated sensors.
Since moving to konnected/nodencu/12v-pirs...
... perfection. not a shred of issue. I constantly wonder why this approach isn't more popular. I found the battery sensors absolutely rubbish.
My understanding from general research is that it is zigbee 3 (probably the same for NASA's Perseverance rover). However I imagine that the system components, from coordinator/router to sensors all come from the same vendor and the hardware/firmware/software for each component are designed to work within a single environment.
It's probably ok as they have insurance 
.
Yes in HA. Dedicated wireless alarm sensors also use 433Mhz - they have much better range than zigbee or zwave. Motion sensors from Bosch for example are extremely reliable and have better discrimination than any HA motion sensor but these can come with a hefty price tag.
Within HE, none of my sensors will give a true signal as to when the battery is close to giving way. Instead I would do what alarm vendors generally recommend, that you change batteries at regular intervals instead of waiting for the sensor to die. I've been keeping track of battery changes over the last 1.5 years in order to get a better feel for this.
I agree, I would not integrate Hubitat into my alarm system, except to report status. I also believe that was never the intention.
If you look at a wireless alarm device you will:
- pay a lot more that $40 or so for a "hobby/convenience" motion sensor.
- The sensors are monitored by the Alarm panel. There is two way communications and an awareness of the frequency being Jammed.
Many folks here have recommended against using Hubitat as a home security system. Having said that, I use Hubitat to shutoff my main water supply in a water is detected where it should not be.
As for ready for Prime time, I guess I'm more of a tinkerer but my Hubitat hub has been very stable and the Lights come on and go off at the right times, which is the reason I started with X-10 (long ago).
Now if you want to know what not ready for prime time, look at the old X-10 system.
An interesting devolution of the point I was making - while I did say 'False Alarms' that didn't necessarily mean Alarms in the traditional sense. Bulbs that are on that say off... for example. Regardless -
The Professional Grade vs Commercial grade - I don't buy that one. Because something is priced 4x higher than something else has nothing at all to do with 'grade' yet thats the earmark.
The UL listing idea has some merit - a central valuation that we can trust. Okay - maybe that works - but. The HE isn't UL listed. Does that mean my HE is crap? (be careful how your answer or Bravenel will appear).
My final point on the original post was the real idea behind it - This group - this compendium of knowledge that spans technology from front to back and meets and greets here - in the forums of HE - TAP IT. We're buying all the devices and testing them. We're creating feedback and test sets. we're solving the problems and answering the questions with blood, sweat, experience and love. The 'supported' list of HE is a weak start point but it could evolve to something more useful, just as a good app evolves. With feedback and adherence to solid verifiable info. I submit it would be effective to expand the supported devices list to become the standard of device information, function, performance and characterization. I'm tired of not trusting any device in my HE network.
While theoretically a good idea--and one I think I've seen proposed before--I don't think it's quite that simple: for one, different people often report different experiences with the same devices. This likely depends on your specific network characteristics. While a bad example because they aren't "officially" compatible, Xiaomi/Aqara devices are a good example of this: perfectly fine for some, prone to dropping off for others. The best knowledge we have on this is that it depends on your repeaters (due to the non-standard behavior of these devices), which will be different for everyone. Zigbee smart bulbs are another example: perfectly fine for some, problematic for the network as a whole for others. This depends on the brand/model but also what other devices you have on your network (only other bulbs is supposedly fine). This also makes things difficult to summarize into a table like the compatible devices list.
I could also see Hubitat themselves hesitant to start or even contribute to a list like this, lest they strain relationships with any manufacturers. However, I'm not saying your idea for community contributions is a bad idea. If you want to pioneer it, go for it! Maybe start a wiki post (if you can't make your own a wiki, I think a moderator and change it for you), invite the community to contribute, and see how it goes.
Because the cost/effort of hardwiring all sensors is substantial if you don't have an existing system.
I would love to hardwire everything and have done cable drops for my networking, but the sheer number of sensors in my home would break me. 
Edit: This was worded very poorly. I did do network drops, but wires for sensors would just be too much.
The price difference is there, but you can get hidden caddx door sensors from Ion or Interlogix for $25. Motions are more like $90, and glass breaks are expensive also.
But they are reliable, and have huge range.
Also, someone mentioned that they are 2 way sensors. Not true. The only ones that I know of that are 2 way are the Elk sensors that talk their proprietary protocol. I don't use those.
