All Temp/Humidity sensors keep quitting

🛎️ Get Help Devices
1

I'm pretty much at the end of my rope here. My issue is that every temp/humidity sensor I've tried has a very short window where it's actually of use.

I started off with three Aqara sensors paired directly to the HE, which worked wonderfully for maybe a couple months using the Oh La Labs drivers (Zigbee - Xiaomi/Aqara Temperature & Humidity Sensor) until some of them just stopped staying paired. One would drop off, I'd pair it again, then it would work for a bit then do the same thing. I believe this issue started after a blip in the power which crashed the hub. After getting annoyed enough with this and accepting that maybe the issues were just the Aqara devices not playing well with the hub like so many other people have reported, I decided to order some Tuya devices.

These Tuyas were zigbee 3.0 devices. Figuring this would be more friendly because of that, I swapped over to these two new ones and used the "Tuya Temperature Humidity Illuminance LCD Display with a Clock" drivers. These also worked well for maybe a month or so before they just started cutting out after a couple days. While this was going on, I kept playing with the Aqara sensors and their placement, figuring that maybe the zigbee mesh wasn't great. The Aqara in my bedroom would sometimes stay connected, but would ultimately give out. Eventually the Tuya that I had in my window for outdoor temps would only stay paired for 12 hours max. Re-pairing would rapid fire readings at the hub, then eventually disconnect. So onto another choice.

Last week I bought Ecowitt devices after learning about integrating them with the hub. I got the GW1100 gateway and two WH31s and installed the gateway in my bedroom and one WH31 in my living room. I did the actual integration with my HE this morning using this https://github.com/sburke781/ecowitt#ecowitt-wifi-gateway. It seemed to work fine, the numbers seemed to match, etc. Now hours later I go to check to make sure things are accurate and it hasn't gotten any temperature readings in 2 hours. Nothing but pressure and humidity, but not as often as I'd specified.

I'm about ready to heave the whole thing out the window since I'm clearly missing something here. I apologize for how discombobulated this probably reads, but this has been months now and Ecowitt was my foray into not using cheap chinese stuff from aliexpress. I figured I just needed something that worked on it's own, but apparently I was wrong. Please, for the love of all that is holy, let me know what I'm doing wrong.

2

Do you have a backup of your hub prior to the crash. If so do a soft reset. You may have a corrupted database if this started after a power outage. Otherwise @bobbyD might be able to assist.

My Ecowitt (albeit the GW1000) has been working flawlessly for over a year (pretty much 24/7) with the exception of the anemometer which failed and was replaced.

Also more details of your zigbee mesh (devices connected) would be helpful.

3

I have backups, but this first time thing that seemed to wreck up the Aqaras was months ago at this point (probably April if I had to guess based on when I ordered the Tuyas) and I've updated and reset the hub numerous times since then. If I didn't do an actual restore to backup after that and just kept going, would that be just pushing a corruption further down the line? I didn't realize that they kept dropping off until maybe a week or so later after I'd already done a software upgrade on the hub.

In terms of the mesh, I've removed just about all of any questionable repeaters except for one that's off to the side, which is a Gen 1 Iris plug. The only other devices aside from that single Iris that would act as zigbee routers are Sonoff S31 lites and a GE dimmer (device details says it's a Jasco 45852) which was the first device I had on the hub and has been one of the most reliable things. My apartment isn't particularly large, so I don't think that it's an issue of needing more repeaters, especially considering the problematic devices have been almost directly between other zigbee repeaters and at times within 10 feet of the hub itself. I do know that the mesh won't always take the most direct route to the hub though, which is why I started removing my old Iris plugs.

As for the Ecowitt, I am pretty sure I have it set up properly per the directions in the page I linked. I blocked out the IP address since I'm not sure if that's enough for nefarious people to do things with, but it is directly the xxx.xxx.x.xxx format which is the hubitat on my router and the IP address in the parent's device driver is also the correct IP address for the ecowitt gateway. I also reserved the gateway's IP to prevent future router resets from wrecking up the connection, and it did connect seemingly without issue in the first place.

ecowitt wsview
ecowitt wsview1080×2340 141 KB

edited to clear up a mistake

4

no its wrong. .the ip address on the panel above should be your hub address not the ecowitts.. the ecowitts in turn goes in the device driver on the hub

as for the sensors above.. i predict its battery issues.. the batteries on those sensors dont last for crap 2-3 months at most.. problably starting to go.. also how often do you have the configured to report.. too often ie more that every 15 minutes kills those little cell batteries even faster.

1 Like
5

My experience with Xiaomi/Aqara sensors is quite the opposite. As I track battery changes on all my sensors, I can say that the minimum life is about 18 months (assuming you use non fake panasonic, energizer etc). Many have gone closer to 2 years.

However, when you set up the sensors up initially, who knows how long they've been on the shelf already as they don't come with battery pull tabs and so are powered on as soon as they leave the production line.

1 Like
6

not for ones in wine fridges etc. maybe just in the open. the ecowitt sensors with two aa lithium batteries last much longer

1 Like
7

I don't run any in fridges so can't comment. However it's not apples-apples to compare a tiny CR2032 to 2x aa batteries (235mAh vs 3500mAh).

1 Like
8

I mispoke on that - the IP address in the Ecowitt app is indeed the IP address for my hub, and the Ecowitt's IP address is in the box in the device driver in the parent device. I'll edit that to fix it.

I've tested the batteries with the same thought as @rocketwiz and they are still alive, not even close to dead. Like still reading 3.1 volts and above. And the Aqara driver doesn't have an option to change how often it reports as far as I can tell. I believe I remember someone saying that there was another driver that did have that option but it didn't actually work since the Aqaras just report in whenever they feel like it.

The Ecowitt gateway is USB powered, the WH31 sensor has two fully charged Eneloop batteries in there, which I know is a nominal 1.2 volts, but considering the gateway isn't reporting as it should I am inclined to not think it's simply a battery issue with that either.

9

Have you checked your wifi channels for conflicts with the zigbee channel? And also do a scan of your neighbourhood wifi (either your wifi router might be able to do that or I believe there are phone apps for that as well)

As you have nothing to lose (unless you have a ton of devices) perhaps another (albeit drastic) way is to delete all the sensors and start pairing them one by one if you are on a clear channel.

1 Like
10

My zigbee channel is 13, my wifi is on channel 4, along with the landlord's wifi and the wifi network that my router is an range extender for (I live in an inlaw style apartment and my landlord's router is on the far side of the house, broadcasting their own network and the "tenant network," which my router extends so I can actually have a decent signal and also assign static IPs for things. 100% not ideal, but this is so far the best thing I've hit on without asking them to snake cat wiring throughout the hose to give me a wired connection.)

As far as other competing wifi from outside of this property, everything else is on different channels.

When you're talking about deleting all the sensors, are you talking just the currently problematic sensors, or everything? I can do that without too much issue since I don't have a ton of devices, but I would like to avoid that if possible. Also I'm willing to write off the aqaras and tuyas as they've displeased me enough that I'm okay with throwing the one remaining aqara and both tuyas in a drawer if and only if the ecowitts will actually report temperature and humidity on a regular schedule as advertised to the hubitat. If they won't do that then I'm going to try to get my money back and may have a stroke from the stress of simply wanting to be able to remotely know how hot a couple rooms are.

11

Is the ecowitt reporting in timely with ecowitt.net
That would rule out problems with the ecowitt.

12

Slso turn on debugging on the ecowitt on the hub snd see if anything is coming in.

13

I didn't register the ecowitt with any cloud services since the hubitat would be my cloud accessible source for the information. Would that have anything to do with it? My impression was that it didn't need that.

I just put the ecowitt driver back into debug as well since it had gone into Info as I believe it is designed to do after the first 30 minutes.

14

Zigbee 13 should be outside channel 4, however do you have channel bonding turned on (this the 40/20 Mhz setting on your access point) which **might interfere.

Maybe a first step would be to post your routing info http://xxxxx/hub/zigbee/getChildAndRouteInfo

Also with your ecowitt does the app show the correct values? If you are getting stuff off the phone app there should be no reason why the driver app on HE shouldn't be pulling in those readings. I assume both HE and the ecowitt are on the same network?

Also measuring no load battery voltage may not be accurate anyway for lithiums. If you have a fresh battery try also changing it on one of the sensors.

2 Likes
15

Battery life on many zigbee end-devices is poorer when a zigbee mesh is poor. I have over 65 Aqara sensors that do not drop off the mesh they're on. I also have ~20 really good zigbee 3.0 repeaters on that mesh. The battery life of the zigbee end-devices in this mesh is 2+ years.

6 Likes
16

I can't find any setting that would allow me to change anything that sounds like that. I looked through every menu trying to find where it would let me change the wifi channel and got nothing there either, which seems strange. Routing info is below.

On the WH31 that I have set up, the values on the display match what WS View is showing me. However the Ecowitt app just has dashes instead of numbers. I set them up originally in Ecowitt earlier this week, then after a bunch of screwing around with the WS View app, got them to show up in there as well. Ecowitt and hubitat are both on the same 2.4g network, but it seems to have erratic reporting for things. Picture of the log for the GW1100 in my bedroom is as follows:

bedroom eco log
bedroom eco log1373×787 71.1 KB

The batteries in the ecowitt WH31 are eneloops, which are NiMH that I just bought a couple weeks ago and charged fully once I opened the package. I'm not sure what the CR2032s that had been in the Aqaras are - they're just marked "Maxell Asia product." No fresh batteries in house, but I do have some "brand new" batteries that are probably around 8 to 10 years old that I got when Radioshack went under, which I'm going to go out on a limb and say that that isn't what you'd suggest using. The Aqaras never had these batteries in them though - just what came in the pack when I bought them.

17

Routing info:

Parent child parameters
EzspGetParentChildParametersResponse [childCount=3, parentEui64=0000000000000000, parentNodeId=65535]

Child Data
child:[Living Room Tu, B830, type:EMBER_SLEEPY_END_DEVICE]
child:[Bedroom Aq, 9DB1, type:EMBER_SLEEPY_END_DEVICE]
child:[Bedroom Lamp, E5C8, type:EMBER_END_DEVICE]

Neighbor Table Entry
[Couch Lamp, 0242], LQI:90, age:3, inCost:7, outCost:0
[Litter Night Light, 3FFC], LQI:170, age:3, inCost:7, outCost:7
[Washing Machine, 7B73], LQI:237, age:7, inCost:5, outCost:0
[Owl Lamp, 8F19], LQI:253, age:5, inCost:3, outCost:1
[6:15 Wakeup, C4CD], LQI:253, age:7, inCost:3, outCost:0
[Bedroom Button, E366], LQI:255, age:7, inCost:1, outCost:0
[Floor Fan, F2BF], LQI:255, age:3, inCost:1, outCost:1

Route Table Entry
status:Active, age:64, routeRecordState:0, concentratorType:None, [Iris Motion, 30AD] via [Owl Lamp, 8F19]
status:Active, age:64, routeRecordState:0, concentratorType:None, [Owl Lamp, 8F19] via [Floor Fan, F2BF]
status:Active, age:64, routeRecordState:0, concentratorType:None, [6:15 Wakeup, C4CD] via [Floor Fan, F2BF]
status:Active, age:64, routeRecordState:0, concentratorType:None, [Litter Night Light, 3FFC] via [Floor Fan, F2BF]
status:Active, age:64, routeRecordState:0, concentratorType:None, [Floor Fan, F2BF] via [Floor Fan, F2BF]
status:Active, age:64, routeRecordState:0, concentratorType:None, [null, D44D] via [Floor Fan, F2BF]
status:Active, age:64, routeRecordState:0, concentratorType:None, [null, 3EE8] via [Floor Fan, F2BF]
status:Active, age:64, routeRecordState:0, concentratorType:None, [null, 5C39] via [Floor Fan, F2BF]
status:In Discovery, age:0, routeRecordState:2, concentratorType:High Ram, [null, 0000] via [Owl Lamp, 8F19]
status:Active, age:64, routeRecordState:0, concentratorType:None, [Couch Lamp, 0242] via [Owl Lamp, 8F19]
status:Active, age:64, routeRecordState:0, concentratorType:None, [null, B777] via [Owl Lamp, 8F19]
status:Active, age:64, routeRecordState:0, concentratorType:None, [null, C6EC] via [Floor Fan, F2BF]
status:Active, age:64, routeRecordState:0, concentratorType:None, [null, 0124] via [Owl Lamp, 8F19]
status:Unused
status:Active, age:64, routeRecordState:0, concentratorType:None, [Outside Tu, 9837] via [Floor Fan, F2BF]
status:Unused

Now I'm at a loss at what like 90% of this actually means, but Washing Machine and 615 Wakeup were two Gen 1 Iris plugs that I removed weeks ago. Does neighbor table entry indicate that something was trying to route through those? I didn't remove the devices from the hub so I could just plug them back in if needed without messing around with re-pairing, but does the zigbee mesh not automatically repair itself? They'd been plugged into outlets for part of the time that I was having all of the issues going back to April or whenever, but I can't give any specific timeframes for when they were plugged in vs unplugged.

In case it matters:
Couch lamp is the jasco 45852 dimmer
Litter Night Light is a sonoff 31 lite
Washing machine is a Gen 1 Iris outlet (unplugged, but not removed from hub)
Owl Lamp is the one remaining Gen 1 Iris outlet that is still in use
615 Wakeup is a Gen 1 Iris outlet (unplugged, but not removed from hub)
Bedroom Button is a Sonoff button that also never works
Floor fan is a sonoff s31 lite
Iris Motion is a Gen 1 Iris motion sensor
Outside Tu is a Tuya TS0201

18

What device drivers and settings are yiu using for the ecowitt in hubitat.

19

Looks like you have the wrong drivers. You shoulda have at least three devices one for the 1100 itself and one for each wh31 and wh32
The ecowitt does not report events like you show above.

Screenshot_20220819-235306
Screenshot_20220819-2353061080×2460 259 KB

1 Like
20

This and the Owl Lamp are the only good repeaters in your Zigbee mesh. All others show high numbers for either inCost or outCost, which means that they either do not hear well the hub or the hub does not hear them.

Good working Zigbee repeaters should be listed in the 'Neighbour Table Entry' with high LQI and low inCost/outCost like this:


Neighbor Table Entry
[TS0121, 1B8E], LQI:255, age:4, inCost:1, outCost:1
[Switch Sonoff LivingRoom , 445F], LQI:254, age:4, inCost:1, outCost:1
[TS011F Leiki, 7DC4], LQI:255, age:5, inCost:1, outCost:1
[TS011F, AB16], LQI:254, age:4, inCost:1, outCost:1
[TS0207 Tuya Zigbee Repeater, C32E], LQI:255, age:4, inCost:1, outCost:1
[Radar Tuya _TZE200_ztc6ggyl LivingRoom, EB4B], LQI:255, age:4, inCost:1, outCost:1
3 Likes