Topic: MB Pro shows messed up sensor data for Ecowitt WH45 5-in-1 air quality **solved*

[Gyvate]

Hi Boris,
I've sent a separate mail, but I also think that the issue is worth being documented here.

the WH45 sensor or sensors (5 "pieces"), are displayed very peculiarly in the live data.
I think, there is still some streamlining necessary.

e.g. (see 1st picture) 338.0 looks like CO2 concentration and appears at Leak ...

Air #2 is probably the PM25 sensor of the WH45 - would fit ...

and then there is a lot of Soil#2x and Sol#3x data - actually too much.

Maybe the API (due to lack of comparison, the WH45 is brand new and I have one of the first models) "slipped" a bit here.

Brian Hamilton with his Weather Display ran in earlier times also in a similar issue, when there were new, previously unused sensors. (PM2.5, SM: WH41/43, WH51).

At Ecowitt the 5 sensors run with the prefix CO2-.


With the Min/Max display it is somehow even more confusing - there are missing entries (CO2) or 2 different PM2.5 sensors (WH41 and WH45)
appear with the same values (here 24 µg/m³), which is not correct.


You can find the (my) current correct values under:
 https://www.ecowitt.net/home/index?id=34418

Not that this would help you, but weewx runs in a similar (even worse) issue - it crashes. Deactivating the sensor and restarting the GW1000 is the only remedy so far.
Meteobridge at least doesn't crash  - just shows additional rubbish so far.

If you need more info or want to connect to my test system, let me know. My email should be in your contact list.

[galfert]

I'm experiencing the same issues with my WH45. Mine arrived on Dec 24, and my wife wrapped it up and stuck it under the tree! I just plugged it in briefly a couple days ago only for a moment and things didn't go well. I was really hoping to be able to work with the software developers to have this ironed out by now. They all have the recent API documentation (that includes WH45) but having real data is always better than just following the documentation. I had to unplug my WH45 until this gets sorted out. I wish I had an extra GW1000 so that I could keep a production system running while a test GW1000 was up to provide live data for the software developers.

Gyvate,
You have more than one GW1000. So maybe that would be beneficial. Thank you for reaching out the various developers.

[Gyvate]

I'm experiencing the same issues with my WH45. Mine arrived on Dec 24, and my wife wrapped it up and stuck it under the tree! I just plugged it in briefly a couple days ago only for a moment and things didn't go well. I was really hoping to be able to work with the software developers to have this ironed out by now. They all have the recent API documentation (that includes WH45) but having real data is always better than just following the documentation. I had to unplug my WH45 until this gets sorted out. I wish I had an extra GW1000 so that I could keep a production system running while a test GW1000 was up to provide live data for the software developers.

Gyvate,
You have more than one GW1000. So maybe that would be beneficial. Thank you for reaching out the various developers.

yes, I do. I have kept it (the WH45) running, but deactivated it for now on the GW1000/WH2650 from which weewx is pulling data (as this makes weewx crash).

The GW1000s which MB pulls data from have it still active, and my HP2551-C too as no harm can be done there - there all is shown as predicted.

My MB databases will now be probably filled with weird data for the WH45 sensors (or what MB right now thinks they are), but that can be cleaned up afterwards. All other data is stored properly as far as I can see.
If Boris wants - and I don't think he has a live system to look at himself - he can use one of my MB Pro installations. That might be very helpful.

The weewx guys, once they realize what's going on (so far no reaction yet on my post to the user group - might have to contact Tom and Gary directly), will be in the same boat.
In fact, I noticed the issue there first.

I first thought it had to do with my switching my weewx database output onto an external disk (old IDE 160 GB disk with USB 2 interface; nicely repurposed) to take the write load from the SD card of my RPi4
(something you can't  do [at least not easily; maybe some fstab entries would do - until the next update] with MB - hence the enforced SLC microSD), but then realized it was the processing of the sensor data.
Looks like the program's interface description doesn't match the new API of the GW1000 fully (yet). Then I had a look into the MB live data and found the above documented mess - luckily MB didn't crash.

[mcrossley]

I'm experiencing the same issues with my WH45. Mine arrived on Dec 24, and my wife wrapped it up and stuck it under the tree! I just plugged it in briefly a couple days ago only for a moment and things didn't go well. I was really hoping to be able to work with the software developers to have this ironed out by now. They all have the recent API documentation (that includes WH45) but having real data is always better than just following the documentation. I had to unplug my WH45 until this gets sorted out. I wish I had an extra GW1000 so that I could keep a production system running while a test GW1000 was up to provide live data for the software developers.

Out of interest, did you try it with Cumulus MX as I added WH45 support a couple of months ago?

[Gyvate]

I'm experiencing the same issues with my WH45. Mine arrived on Dec 24, and my wife wrapped it up and stuck it under the tree! I just plugged it in briefly a couple days ago only for a moment and things didn't go well. I was really hoping to be able to work with the software developers to have this ironed out by now. They all have the recent API documentation (that includes WH45) but having real data is always better than just following the documentation. I had to unplug my WH45 until this gets sorted out. I wish I had an extra GW1000 so that I could keep a production system running while a test GW1000 was up to provide live data for the software developers.

Out of interest, did you try it with Cumulus MX as I added WH45 support a couple of months ago?

@mcrossley
My CumulusMX instance runs with the same GW1000/WH2650 as my weewx instance does ... as I deactivated the sensor (array) for this console (as weewx crashed),  CumulusMX will not show anything - and if it picked up somethings, I didn't see it under the extra sensors in the interface. Where exactly would I find it ?

I've now changed the console IP and restarted CumulusMX - CumulusMX will now receive the sensor data from another GW1000 for which the WH45 is activated - I'm still running build 3097 (and read that 3099 is available). Are your changes already implemented in build 3097 ?

Will see what it gives tomorrow (today my later morning), now too late ...

Don't see any sensor in the extra data log (or in the extra sensor interface) which I could relate to the WH45 sensors ...
Is this because of build 3097 ?
Or do I just look at the wrong place ?

[galfert]

Cumulus MX gained WH45 support on Sept 14, version 3.8.4 build 3094.
I'm running Cumulus MX 3.9.3 build 3098.

I looked at Cumulus MX and I don't see the data anywhere. Not showing up under extra sensors...but I also didn't restart Cumulus MX to see if that makes a difference. I do have extra sensors turned on and I do see my WH41.

Here is the raw GW1000 API data capture, I'm sure this helps:

FF FF 27 00 6D 01 00 F8 06 35 08 27 D6 09 27 E1 02 00 D2 07 5E 0A 00 79
0B 00 05 0C 00 05 15 00 00 00 00 16 00 02 17 00 2A 01 71 4D 00 C4 1A 00
E4 22 3B 4C 05 00 00 00 FF FF FF 06 F5 FF FF FF FF FF FF FF 19 00 33 0E
00 00 10 00 00 11 00 0D 12 00 00 01 0B 13 00 00 3A 62 0D 00 00 70 00 ED
3A 00 2B 00 11 00 1E 00 0D 03 7B 03 D2 06 02

Looks like the WH45 data is there starting with hex byte 70.

[Gyvate]

@mcrossley
OK - mine runs CumulusMX 3.9.2 build 3097 - but no fields that could correspond to the 5 WH45 sensors in the data log nor extra sensor interface.
Seems to be oblivious of them. The support must be/remain somehow under the hood and doesn't show in the user interface as far as I can see.
(what in my perception means that there is no field, table entry in the database schema for them - unless they are meant to be put under userTemp, but these are empty)

My other extra sensors (extraTemp - WH31 1-4), air quality (PM2.5 actual and 24hav, WH41 1-2) and Soil Moisture (WH51 1-5) are shown.
Lightning count, distance, last occurrence (WH57) doesn't seem to be shown in CumulusMX.

[mcrossley]

Thanks chaps, I'll take another look. MX doesn't currently display the CO₂ data on the admin interface, but only makes it available via a web tag to include in your own data files or web pages. I'll have to check what I did with the AQ reading, I have a vague memory of just mapping to AQ sensor 1

The fact that it doesn't crash means something is working to some degree!   

[Gyvate]

Error detected in a joint effort - Boris will provide the fix with next release (update).

[mcrossley]

If you switch on debug logging in Cumulus MX, you should see the WH45 data being decoded.

At present only the current and 24 hour CO₂ values are saved to be used with the web tags.

I have yet to decide how to handle the data from these devices - the extra log file is already too cumbersome. Maybe a separate logfile just for WH45?

Then what to display under extra sensors....

[galfert]

For Cumulus MX here is my feedback as to where to display the data. I hope this makes sense.

Because Ecowitt stations only have extra temperature and humidity sensors #1 - 8 and because the WH45 has temperature and humidity I think that these should be mapped to Extra Temperature / humidity #9

For the PM2.5 and PM10 these should be extra not yet listed Air quality #5 and #6.
Then the CO2 should be perhaps a different new category because it has different units of measurement than Particulate Matter.

Actually what I just said should also apply to the Meteobridge. I don't understand it being Extra Temperature #8 now. It should be #9 because there is no Extra #0. My first Extra shows up as #1 in Meteobridge (this is not main outdoor nor main indoor). So if my first Extra is #1 then the WH45 should be Extra #9.

[docbee]

Update mapping the WH45 temp/hum sensor to "th9" is available.