is this normal?

[mcrossley]

When you check on ecowitt, it must of course be during one of the periods that MX is showing bad data - what are the current values (ignore the graph).

Also, a further thought, the error will actually cause the humidity to flat line for the duration, the temperatire will continue working (it is process first). The weird values for the temperature must be coming from Ecowitt.

It's really hard to tell from the graph when exactly the bad values occurred, can you zoom in on a section that is covered by the log file so I can see the minutes, then I can examine the data at that point.

[mcrossley]

Can you post the whole log file, I can see no instances of bad data for the CO2 in section above.

And which channel is the soil sensor?

[Nossie]

Hi Mark, thanks

the soil temperature sensor is on channel 1,

I've restarted and will run it for a few hours then post a log - I've taken a screenshot from within cmx, not so coincidently I think when the soil temp changes the CO2 average goes to 0, it's almost like they are fighting with each other?

I'll upload the data so you have a decent amount to review -

Didn't produce the results I expected - ignore just for kicks I'll disconnect the CO2 sensor and see if the soil temp stops thinking I live in Africa.

[Cortmalaw]

Some comments on your graphs from a different angle:

1. CO2 levels should not bounce around like that unless someone is breathing on the sensor. An outdoor CO2 sensor should vary very little over time, so much so that its hardly worth recording. Is that an indoor sensor?

2. There's little point in displaying PM10 and PM2.5 on separate graphs. That's because PM2.5 is a count of all the particles smaller than 2.5, and PM10 is a count of all particles smaller than 10 (which includes all those smaller than 2.5). To interpret, you need to be able to see how many particles are under 2.5 (ie mainly smoke, as well as very fine particles under 1 (PM1)), and how many are between 2.5 and 10 (ie dust and pollen). The only easy way to see that is to show PM10 and PM2.5 as two lines on the same plot. And that plot really should show PM1 (fine particles) separately as well - though I believe Ecowitt have strangely decided to omit that important measure!

See my plots at https://io.adafruit.com/Cortmalaw/dashboards/cortmalaw

[Nossie]

no it shouldn't

@Mark - here are the closer screenshots, I'll upload the logs too

[mcrossley]

Thanks, I'll look at this later today, meantime I have found and fixed an error in MX where it decodes the CO2/PM 24 average numbers.

You are one of the first users of the Ecowitt Cloud API - certainly with a significant number of extra sensors, so I'm afraid you are finding the initial implementation problems.

It seems it's your freezer sensor that is temperature only - what model of sensor is that?

[Nossie]

The freezer sensor is the (WN30AN) https://shop.ecowitt.com/products/wn30

That's fine Mark, as I said to Hans in another thread I don't mind being a test bunny


The soil temp sensor is this one - https://shop.ecowitt.com/products/wn34s (WN34BS)


Both are temp only, they don't do humidity or pressure.

[mcrossley]

Ian, I cannot see anything wrong with the soil temperature in either the Ecowitt data or MX, it may be a consequence of the CO2 error.

Would you be willing to try a test build for me (or share your Ecowitt API details so I can test)?

[Nossie]

I send you the api details by pm

[mcrossley]

Thanks, I spotted a silly error in the fix straight away with live data!

[Nossie]

I found the problem. Your temperature and humidity sensor #4 is only sending the temperature in the JSON (or at least the Ecowitt cloud API is). So MX is seeing it as malformed and throwing an exception when it tries to access the humidity values. It should really cope with this; I'll add it to the to-do list. (but, even if the sensor is bust, I think the Ecowitt API should still send the fields, but with dashed out values)

Code: Select all

        "temp_and_humidity_ch3": {
            "temperature": {
                "time": "1705042474",
                "unit": "℃",
                "value": "1.7"
            },
            "humidity": {
                "time": "1705042474",
                "unit": "%",
                "value": "98"
            }
        },
        "temp_and_humidity_ch4": {
            "temperature": {
                "time": "1705042474",
                "unit": "℃",
                "value": "-29.5"
            }
        },

Meantime, check your sensor, and if that is OK, is the hum data shown on the Ecowitt web site? If it is then it's a support call with Ecowitt.

I just realized I missed this entire post - how awkward haha

[mcrossley]

Good news is the graphs look as they should now after running for a few hours against your data - thanks.

[Nossie]

both of them? excellent news indeed

[Nossie]

@Mark

Hi - thanks for fixing the average CO2 issue - that does indeed seem to be resolved, however the "Garden Soil Temperature (30cm underground)" is still flicking between 3.2C and 46C, can you see that from your side?


cheerses

[mcrossley]

Hmm, it wasn't doing that for me, I'll check again in case something got reverted before release...