Topic: Ecowitt WH41 - Calculations to take the humidity out

[ajay]

I live in an area that frequently has valley fog most clear nights during the summer and early fall. I have been working on a formula (modified from Colorado State U.) that seems to work quite well with the WH41 with regards to taking water vapor out of the readings. Typically I see readings at or above 60ug/m3 on fogging mornings when local observations indicate 10ug/m3 or below. The equation brings the WH41 in-line with local, official reports.

At least my version of the WH41 uses the Plantower light scatter sensor. (Same manufacturer as the Purple Air sensors.) This type of sensor is notorious for reporting water vapor as well as other particulates such as smoke...the stuff you really want to measure. Purple Air does their own correction and includes other formulas from the EPA, etc to correct for humidity. Ecowitt doesn't do anything and there lies the rub.

Is there anyway we can get Ecowitt to apply some form of correction or allow the end-user the ability to apply a formula as a correction? The GW1000 has offsets that can be applied but that does no good. There needs to be an equation applied to the raw reading using relative humidity.

I have applied the correction using weewx and it produces reports that are in-line with official recordings in my area. Although this works, it really needs to be done from the GW1000. Ecowitt...are you listening?

BTW, I am still tweaking the equation a bit and once I am satisfied with the results (over at least a few more months) I will share the equation. It does require having a relative humidity sensor next to the WH41.

[davidefa]

Please share 'your correction' now  ( and the source of the formula )

P.S.
On epa.gov there is this document
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKEwjemOemlsTyAhVHwKQKHbLVCnYQFnoECAMQAQ&url=https%3A%2F%2Fwww3.epa.gov%2Fttnchie1%2Fconference%2Fei13%2Fmobile%2Flindhjem.pdf&usg=AOvVaw27ntHjjc5yTO8_V4dRRcXQ
eq. 14 calculate a NOx emissions correction based on humidity, are you deriving something similar?
Is this correction function of humidity or takes in consideration 'condensing humidity' ( just wondering )?

P.P.S.
What sensor is inside wh41?
In this post has been reported it is an honeywell
in this ( on ecowitt forum ) there are a few photo of the inside of a wh45 ( but it is not clear what sensor is used )

[sky_watcher]

What sensor is inside wh41?
In this post has been reported it is an honeywell
in this ( on ecowitt forum ) there are a few photo of the inside of a wh45 ( but it is not clear what sensor is used )

The WH41 has a Honeywell sensor. From the case, it looks like the WH45 is using a Plantower sensor. Don't know which model it uses, but I use the PMS5003.

[ajay]

My particular WH41 has the Plantower sensor. It's stamped right on the side of the sensor.

I could not get the EPA's formula to work in 99% humidity conditions, but other formulas from other studies do seem to work with some modification. It will take some time however because I do not know how temperature may play a role and it needs to be tested in all seasonal weather conditions.

As I write this we are at 99% humidity and the corrected values from the WH41 are close to an official EPA reporting station within 25 miles of here. Will the equation hold up? I have no idea, but it's worth a try.

[ajay]

With regards to the equation, I first saw a version of it from a paper from Colorado State but later discovered their equation was derived by a paper in 2020 entitled "Development and Application of a United States wide correction for PM2.5 data collected with the PurpleAir sensor  - Karoline K. Barkjohn, Brett Gantt, & Andrea L. Clements

Here's a link to their paper: https://amt.copernicus.org/preprints/amt-2020-413/amt-2020-413.pdf

The PurpleAir sensor uses two sensors to derive the measurement while the WH41 only has one. In theory, the Ecowitt WH41 may require individual equation tweaking based on variability of the one (and only) sensor. That's why it would make sense for Ecowitt to allow for a calculated correction based on the raw pm2.5 value and relative humidity. The GW1000 already allows for offsets to correct rainfall, etc. If they took their offset corrections a bit further and allowed for calculations based on their relative humidity, that would be super helpful and it would allow users to correct readings that are "off-the-charts" when humidity is at 99%.

I had considered returning the WH41 and just getting the Purple Air sensor but now that I have been able to apply a calculation to the reading using weewx, I'll hang on to the unit. I'll bet Ecowitt could sell a bunch more of these sensors if they corrected their unit for high humidity levels. Why don't they? PurpleAir does.

[davidefa]

Very interesting reading ( need 'some' time to digest it ).

Only one point: they applied the correction to 24h mean values ( from a cursory reading ), I'd like to apply it also to 'instantaneous' values.
Another consideration: I underestimated the value of 24h mean values.
I mean till now I considered only peak values ( which are important ). I think I should pay more attention to total/integrated dust exposure ( sort of 'dust  dosimeter' ) and evaluate both acute and chronic dust pollution

[billfor]

Interesting, with my WH41 the humidity does not affect it as much as rain. Not sure how I could correct for that.  We just had the hurricane run through here and it went bonkers.

[ajay]

Interesting, with my WH41 the humidity does not affect it as much as rain. Not sure how I could correct for that.  We just had the hurricane run through here and it went bonkers.

That is interesting. I wonder if you have the Honeywell sensor or the Plantower sensor. Mine works quite differently. Typically when it rains it drops the count down to practically nothing (which it should as the air gets washed out.) It's when visibility is reduced by thick fog that mine goes bonkers. We had a good soaking rain this morning and it was down to about 5ug/m3 with humidity at about 96-97%. Only with fog (99%) does it go up to above 60 ug/m3.

[davidefa]

May these different behaviour connected to the sensor mounting ( for example under a large patio or inside a small solar shield )?
Mine is still inside ( home made pms7003 ), I'd like to mount outside but don't know where to position properly

[billfor]

Support says "the tiny water droplets in the air works like a lens or mirror, and this can lead to the optical counting sensor showing a very high value. This is typical to such sensor. "

[watsonm]

I have mounted my PMS5003 in my  garage having drilled a couple of holes for it to poke through and used a cowling to protect it.

I also added code to correct for humidity


//Correction algorithm thanks to help of Zbyszek Kiliański (Krakow Zdroj)
pm2p5 = pm2p5/(1.0+0.48756*pow((humidity/100.0), 8.60068));

pm10 = pm10/(1.0+0.81559*pow((humidity/100.0), 5.83411));


Probably not really required since the units are low end but made me feel good !


 [ You are not allowed to view attachments ]

[zippm]

@watsonm love this installation, very professional.

Very interested in how to factor humidity into the ecowitt WH41 results, which is what I'm using now. I'm in the pacific northwest where humidity does sometimes get off the charts. My unit actually says that it's the WH41B, but I have no idea what they changed.

As far as calibration goes, I'm using home assistant as a "middleman" to present easy graphs for the family before ingesting from that to a time series database (I'm using prometheus and grafana). I'm using home assistant's template object, which lets you take one datasource, transform it, and expose it as a new one, for other metrics as a way to apply formulas (for example to calculate AQI) and calibration offsets where they aren't supported.

[watsonm]

My installation is one that is using my own software so I can easily add the correction code.  This is then sent to a broker where it is read by
another software package for integration into my weather data.