Found Most Accurate Humidity Settings After Calibrating 2x LYWSD03MMC & 1x MJWSD05MMC with MgCl and NaCl

[Yoshi-523]

After many months of testing I think I've found the most accurate settings for calibrating my humidity sensor settings in my LYWSD03MMC and MJWSD05MMC meters in the custom firmware provided by pvvx.

I would be most grateful if anyone else has done any humidity testing to offer their results too.

I have found that for me the best humidity settings are:
Humidity Slope Factor: 150.00 Zero Offset -18.00

The default settings (I no longer use) are:
Humidity Slope Factor: 125.00 Zero Offset -6.00

Thankfully the temperature sensors didn't need adjusting as they where all very accurate when tested in relation to multimeter thermocouple and icewater test.

I have 2x LYWSD03MMC ver:2019 (HW: B2.0) with the custom pvvx firmware v5.1 flashed and they each say thet are now HW: B1.7
I also have 1x MJWSD05MMC ver:2022 (HW: 2.3) with the custom pvvx firmware v5.1

All calibration testing was carried out using the temperature and humidity reference chart below. I tested in a sealed air tight tupperware type container with the meters in it and inside a small container of either Magnesium Chloride and distilled water for a 33% humidity calibration standard or Sodium Chloride and distilled water for a 75% humidity calibration standard. I let the container sit for a few days at air temp (generally around 17-22 degrees C) to get a stable reading and then again in the fridge for a number of days (generally around 3-4 degrees C). I also threw in a couple of other cheap and nasty humidity guages that don't have the ability to calibrate, just to help give some more numbers for general reference.

I tried many different numbers in Humidity Slope Factor and Zero Offset. Each time having to wait days after taking out of the fridge or putting back in the fridge or after changing from Magnesium Chloride to Sodium Chloride. The process went on for months. Since I didn't know which numbers really would do what unil I played around with them.

Rather than bore you with my whole massive list of numbers tried and how they compared to reference I'll just post the 2 here that are my preferred settings vs the default:

MJWSD05MMC ver:2022 (HW: 2.3)
(USING DEFAULT SETTINGS):
Humidity Slope Factor: 125.00 Zero Offset -6.00

Left 3-4 days to Warm Up (NaCl which shd be 75%)
Temp: 17.27°C Humidity: 72.06%

2 Days Later Left in Fridge: (NaCl which shd be 75%)
Temp: 4.12°C Humidity: 73.37%

Left Many days to Warm Up (MgCl which shd be 33%)
Temp: 19.0°C Humidity: 36.8%

---

(USING MY PREFERRED SETTINGS:)
Humidity Slope Factor: 150.00 Zero Offset -18.00

Left 3-4 days to Warm Up (NaCl which shd be 75%)
Temp: 17.19°C Humidity: 75.74%

Left for 2 days in Fridge (NaCl which shd be 75%)
Temp: 3.28°C Humidity: 75.06%

Left for 1-2 days in Fridge (MgCl which shd be 33%)
Temp: 3.73°C Humidity: 34.55%

---

1st meter LYWSD03MMC ver:2019 (HW: B2.0) [called HW:B1.7 in pvvx custom firmware]
(USING DEFAULT SETTINGS):
Humidity Slope Factor: 125.00 Zero Offset -6.00

Left for unknown number of days (NaCl which shd be 75%)
Temp: 17.98°C Humidity: 74.74%

2 Days Later Left in Fridge (NaCl which shd be 75%)
Temp: 3.95°C Humidity: 75.83%

Left 2 Days to Warm Up (NaCl which shd be 75%)
Temp: 26.47°C Humidity: 78.03%

Left Many days to Warm Up (MgCl which shd be 33%)
Temp: 19.2°C Humidity: 41%

---

(USING MY PREFERRED SETTINGS:)
Humidity Slope Factor: 150.00 Zero Offset -18.00

Left 3-4 days in Fridge (NaCl which shd be 75%)
Temp: 3.84°C Humidity: 75.69%

Left 1-2 days in Fridge (MgCl which shd be 33%)
Temp: 3.89°C Humidity: 37.74%

---

2nd meter LYWSD03MMC ver:2019 (HW: B2.0) [called HW:B1.7 in pvvx custom firmware]

(USING DEFAULT SETTINGS):
Humidity Slope Factor: 125.00 Zero Offset -6.00

2 Days Left in Fridge (NaCl which shd be 75%)
Temp: 3.88°C Humidity 76.00%

Left 2 Days to Warm Up (NaCl which shd be 75%)
Temp: 26.4°C Humidity: 77%

Left Many days to Warm Up (MgCl which shd be 33%)
Temp: 19.3°C Humidity: 41%

---

(USING MY PREFERRED SETTINGS:)
Humidity Slope Factor: 150.00 Zero Offset -18.00

2 Days Later Left in Fridge (MgCl which shd be 33%)
Temp: 3.78°C Humidity: 35.61%

2 Days Later Left in Fridge (NaCl which shd be 75%)
Temp: 3.69°C, Humi: 74.58%

Uploading Humidity Fixed Points of Binary Saturated Aqueous Solutions (Large List of Chemicals, Temps & Humidity %).pdf…

[pvvx]

The firmware uses the coefficients indicated by the manufacturer of the sensors.

Calibration in media with chemicals forever spoils the humidity sensor. These sensors are not designed for use in pairs of salts and other chemistry.

For the environments you used for calibration, other types of humidity sensors are used. Perhaps simple options with a special membrane that only allows water vapor to pass through and retains salt and alkaline compounds will do...
Otherwise, the sensor is guaranteed to degrade.
A decrease in the humidity slope coefficient and zero offset already indicates a deterioration in the condition of the sensor.
It appears that the conductivity has changed due to the presence of salt on the measuring element. Now you can throw these sensors in the trash.

[Yoshi-523]

Thanks for the feedback pvvx. Can you please help to explain the physics of this a bit more?

So the actual sensors where never in direct contact with any of the magnesium chloride water or the sodium chloride water. I did it based on multiple demonstrations found online for calibrating humitidy guages. So I made sure to keep the magnesium chloride always in vast excess when in the water so the small container (inside the larger airtight tupperware container) always had a larger amount of visible salt crystal with just a layer of water solution above it. I thought that since the salts used are hygroscopic and at such low temperatures (compared to their boiling point) that there would be no significant evaporation of the salt into the surrounding air and rather just absorb water (in much the same way as damprid products are used which contain calcium chloride to absorb moisture from the air). Was I incorrect in my assumtion of this? Are you saying that even when a container has an excess of hygroscopic salt crystal in water that some amount of salt is volitised into the surrounding air causing it to come in contact with the sensor and cause corrosion?

I understand you saying these sensors are set according the manufacturers spec but there has been so many videos on much more expensive humidity guages which have large amounts of variation in humidity readings when placed along side each other so I simply bought these meters as I realised I wanted a meter that had the ability to adjust and calibrate (as I am told every meter should be calibrated at some point to keep it accurate). I read that analog meters are usually needed to calibrate once per year and digital less often. I'm finding that humidity sensors such as ones inbuild into dehumidifiers seem to have crazy numbers indicating that they are either vastly inaccurate or too close to a source of humidity different than the room (ie the condensor fins, or water capture container).

Sorry if this is a long message. It's just that I wanted and still want to get the most accurate humidity readings for an indoor dwelling as I have become chronicall ill from toxic mould following a very large period of rain and severe flooding a few years back. So after moving house multiple times to recover and experiencing issues with mould illness in multiple dwellings, I now need to monitor how to often I need to run a compressor dehumidifier to keep the humidity below 50-55% to avoid mould growing where I live.

[pvvx]

The sensor measures relative humidity and temperature. There is always convection and distribution of air layers by temperature in the room. As a result, two sensors installed on a shelf next to each other can show a temperature difference of several degrees. Accordingly, the relative humidity readings will be different. Only the dew point, calculated based on the temperature and relative humidity readings, will be closer in value.
Even if one sensor gets hot, the dew point value will remain the same, but the relative humidity readings will be completely different.
As a result, there is no point in relying on relative humidity readings. Especially for diagnosing the appearance of mold and other fungi.
You should compare the dew point and the lowest surface temperature where fungi can grow.

[Yoshi-523]

Hi PVVX.
Your Custom firmware is highly commendable and thanks for your fantastic contributions to the community.
As an enthusiast/tinkerer, I’ve been using custom firmware for every smart phone I've owned (starting back with my first HTC Desire HD in 2011 to multiple Samsung models I still use today. I've even been running my Ethernet & Wi-Fi routers for many years now all on open source custom firmware as I feel that the additional features & controls offered to the power user are invaluable.

Considering my last question about if you thought it's possible that water vapour in the airtight Tupperware container may have contained volatised Magnesium chloride and then tested also with Sodom chloride may have volatised small amounts of salt, entering the air and gotten into the sensor compromising the accuracy? I was wondering if firstly you fan confirm if this hypothesis is true?

So far my calibrated Xiaomi meters seem to be giving readings of humidity and temp that seem in the ballpark or possibly more accurate than other meters (obviously I know why you said, they are likely damaged now from small amounts of salt vapour and potential corrosion.

I was wondering if either of these two options you'd recommend:
OPTION 1 : Considering the 3x calibrated meters in question are likely to be damaged from the professional calibrations with salts. So I could try (with the risk of destroying one of my 3x Xiaomi meters exposed to salts for calibration the following:
Take apart unit. Remove lithium coin battery sell. Possibly manually short caps (to discharge them) (or wait for them to self-discharge). Once all power is drained I was thinking of rinsing or spraying down the circuit board and internal sensor with plenty of distilled water to remove any potential trace amounts of salt that may have vaporised into the sensor. After leaving the circuit board and internal sensor for a few days or even a week to fully dry, I can put the batter back in and compare reading with other sensors.

OPTION 2 :
I could buy a complete new
LYWSD03MMC ver:2019 (HW: B2.0) [called HW:B1.7 in PVVC custom firmware
I’d prefer this using as the price in roughly have for the one below and more compact but I seem remember some of your comments on the forum suggesting that may newer models have components removed resulting in poor performance (esp battery life). AliExpress isn’t always the best for accurate descriptions of which model of LYWSD03MMC they are selling OR for about twice the price I could buy another: MJWSD05MMC ver:2022 (HW: 2.3)

I can compare the results with a new undamaged meter alongside the 3x sensors you say are likely damaged from salt calibration. I can then set test the new meter and 3x older potentially damaged meters.
(USING DEFAULT SETTINGS):
Humidity Slope Factor: 125.00 Zero Offset -6.00

I can also re-test new meter alongside the 3x sensors you say are likely damaged from salt calibration.
(USING MY PREFERRED SETTINGS :)
Humidity Slope Factor: 150.00 Zero Offset -18.00

Please let me know what up think would be the best option.

Thanks & best regards,
Yoshi