I'm currently using Ecolink motion sensors for my driveway and entry lights. They are mounted well out of the weather, and seem to be working OK. But they are SLOW to respond! I've seen folks praise the Xiaomi sensors (and Zigbee in general) for speed, but they don't have the detection range I'd like.
The Hue sensor looks good in terms of range, outdoor use, etc. But how fast is it to respond to motion? I have no Hue devices and will be pairing directly to the hub. When going out the entry my lights don't come on now until I'm well out the door, and guests could certainly benefit from the entry lights much sooner than they come on now. How quick is the Hue?
@BrianX I have the hue motion sensor and it is very fast. It also has a light sensor which is very accurate. You can pair the hue motion sensor directly to the hub with the built-in driver. Also the light sensor is independent of motion which means it will give you a light level even if there is no motion. Some motion sensors will not send a light level unless motion is tripped first.
I haven't specifically measured these compared to other Zigbee sensors, but I've heard others report that they are fast (including the above--yay!). I have never seen a Zigbee motion sensor as slow as most Z-Wave sensors, so you're bound to see some improvement. I've also never measured the Ecolink specifically but have measured the similar Monoprice sensor (they're likely identical hardware with different firmware), and they were consistently the slowest--so slow that I took them out of my experiment eventually because they were so boring. I've seen people praise Ecolink elsewhere so I'm thinking maybe I got a dud (if these are really the same) or they've never used a good sensor, but I don't know. I'm guessing the latter, even though some ST users gave me flack for that viewpoint.
FYI, that is the only way these will work on Hubitat anyway, so definitely no problems there!
I tried Zooz and SmartThings and both were pretty fast but didn't have the range that I wanted and were not really "outdoor" sensors. I bought the Hue "outdoor" for $50. Definitely pricey but in my opinion, well worth the $$$. Great range with adjustable sensitivity and seems to be well made. The Hue brand motion sensors will by my 1st choice going forward. ...and with Black Friday, Cyber Monday and Prime Days coming up, I'll be keeping my eyes open for them.
I've got two Hue Motion sensors (paired to the Hubitat directly) and they are indeed fast in my experience as others above have been saying. One turns on 4 lights in the master bathroom and the other is in the laundry room that is the main entry to the house coming through the garage.
The problem with measuring that is that you need a zero point. The best I can easily do is compare the time one sensor responds versus the time the first sensor responds (to the same real-world event, like me walking into the room). Looking at a few trials I ran, the Monoprice responded anywhere from about 0.8 to 1.3 seconds slower than my fastest sensor, which was always an Iris v2 in this trial. In some cases, it mysteriously took 4-5 seconds longer (this is why I stopped seriously considering this sensor in future tests).
Not exactly what you're looking for, but the closest I have.
My z-wave plus motion sensors are < a second most of the time based on stop watch testing. Is it <500ms? Who knows? There is no standard quantitative measurement for motion response time. So fast/slow is all most people can anecdotally give you.
Yeah sorry I forgot most sensors don’t have the flashy thing like my Iris v2. With those, you can easily watch the flash and the logs in Hubitat and estimate within couple hundred ms. But yeah if they don’t flash, you have no way of knowing when motion was actually detected by the sensor. Good point
That's not really the problem: in my experience, the LED (many sensors have these, including the extremely slow Monoprice) and the reporting of motion happen around the same time. The slowness seems to come from how long the sensor itself take to do those from the time the actual motion starts.
I could be wrong but looking at the Iris V2 circuit board, it looks like the LED is in series with the out pin of the PIR sensor, meaning it flashes when the sensor sends its high pulse (motion detected). Not to say other motion sensors have to be designed the same way.
That could indeed be the case for other sensors, too; I'm not convinced the delay is inherent to the protocol (certainly the 4-5 second times on the Monoprice aren't). The lens or sensor could just be different enough that it won't detect motion until so many "zones" are activated or however they work under the hood, and perhaps all the Z-Wave manufacturers chose the ones that need more compared to what the Zigbee manufacturers chose. That part I don't know about--only my real-world experience that any Zigbee motion sensor is faster than nearly every Z-Wave sensor I've tried.
I just tested the HS-MS100+ which is z-wave and it took 214ms to turn on a hue light that is connected to the hue bridge. Not too bad for a z-wave motion sensor!
I also just tested the Zooz ZSE18 and it took 233ms to turn on the same light. So I am going to test the Hue sensor next.
The Hue motion sensor came in at 215ms. So all are pretty close. I was just curious.
What did you compare it to? Manual measurements with a stopwatch? You said "to turn on a light," but the time between motion registering as active and the light turning on isn't really the duration of interest, which is why I'm guessing you have used a stopwatch and tried to measure from something else (like you waving your hand at the sensor). Adding the lights into the equation also adds a second possibility for a different delay (the Hue LAN integration should be consistently fast but isn't really of interest, but assuming it's consistent, its effect can probably be ignored).
My tests involved a small app I wrote to log the event times (as registered in the Event object, not the log time which was often a couple hundred milliseconds later), but the only "zero point" I had to use was the time of the first sensor to respond. I can certainly imagine better ways to do some of this (some sort of controllable IR source that would trip all the sensors at the same, specified time, instead of me walking into a room with the sensors all pointed at me), but none are practical (I don't have one of those hypothetical devices) or easy (still have the measurement problem: do I let Hubitat turn on this imaginary device and time from there?) to do at home, where I did these measurements.
I certainly didn't go to the extent you have and I only did it to see what the time was from trip to light on. I just used the logs which is all I have to use and that is where the times are from. So like you said I was just tripping the sensor by walking by it. I realize it is a crude measurement and probably not very accurate. I was curious because of all the people having all these long delays.
Not at all, in my experience. Z-Wave 2.5 version of Ecolink motion sensor is the fastest z-wave motion detector I've have yet found, it has an incredible LOF and if set to test mode, resets in 15 seconds (which is outstanding). Experience varies based on how the hub is performing as well, so I'm not sure how to bench mark the actual message to the hub.
For what it's worth, I tested the Monoprice shortly after I got Hubitat (January of 2018) on the same hub I was using for everything. This summer, however, I did a bunch of Z-Wave and Zigbee sensor tests on a hub dedicated entirely to that purpose. I hope to format the results in a more easily share-able format soon, though everything I said about about design flaws still applies (e.g., no good way to measure time zero).
I've heard good things about the Ecolink (I know they have at least a "classic" and a Plus version and I'm never sure which people mean), so I'm glad to hear that! I've never used one, so again, my use of the Monoprice as a proxy was just an approximation based on the guess that identical hardware might perform similarly. Some day, I might get a real one just to try, but safe to say, the Monoprice firmware does not make me want to risk it.
In my experience with the Ecolink (fairly new, zwave plus) are just slow to detect motion. Once the led flashes, the lights come on fairly quickly, but it often takes several seconds of me walking past the sensor before it sees me. What I'm looking for is a sensor that sees motion faster.
