Soil moisture readings

[mutleyhehe]

Is it 'normal' for moisture readings to jump up and down so significantly like in the picture below? I calibrated the sensor at 4.30pm, and it correctly spiked to 100%. It's an indoor plant. After a few hours the reading dropped as expected and reached 63% at calibration + 8 hours (C+8).

But then suddenly at C+9 the reading shot back up to 100% and remained there for the next 6 hours, even though I haven't watered the plant since calibration. At hour C+17, it's now coming down to 72%.

[PricelessToolkit]

Hi, No, it's not normal! It can jump 3% with air humidity.

1. Not calibrated with the full range, "driest to wettest"
2. Does not sit tightly in the soil
3. New plant, new soil? "The soil is not yet compacted?"

[mutleyhehe]

Thank you! The plants and soil are not new. I guess I must have messed up the calibration process. I will try again.

Same issue with outdoor plant as well (below report, last watering was done a couple of days ago but the readings are all over the place).

[mutleyhehe]

@PricelessToolkit is there a proper way to reset the calibration? I've re-done the calibration a few times (doing the calibration/trigger button press/release method) but the moisture readings are not changing (or are only being reported very rarely, e.g. after 10 hours instead of hourly). This is happening on 3 units, and they are each in different plant pots/areas.

[PricelessToolkit]

if it hapens to 3 units, i can say with 100% confidence the problem is in calibration. if you re calibrate it correctly it will work!! Full ADC span is 3200-4100 so dryest and wethest if your calibration span is like 3600-3650, it will jump like your sensor.

1. Can you explain how you do colibration

2. Is your firmware a new version? Does it have basic wifi?

3. Your sensors in ESP-NOW or WIFI mode ?

4. Have you text row sensor ?

[mutleyhehe]

Thank you! Here is the information:

1. I put the sensor in the plant plot. Then manually hold CAL button, press TRG briefly and release it, then release CAL before 2 seconds. I see the calibration blue lights in the sequence as you describe in the instructions. I wait for the dry calibration to complete, then water the soil, and wait for the wet calibration to complete.

Do you think I should just use soil in 2 glasses, like you do in the YouTube video?

2. I don't know how to check which firmware I have.
3. I'm using Fast WiFi.
4. I've attached screen of the 3 sensors.

[PricelessToolkit]

ok, I think your dry soil is not dry enough. You can test it by holding the sensor in your hand, tightly no gap! “part that goes into soil” and trigger it and watch “Moisture ADC” value then release it and don’t touch “part that goes into soil” and trigger again, and check the value it's a raw sensor value and it does not change by calibration. So minimum needs to be ~3200, max ~4200; this will give you an idea that the sensor hardware is working. For testing your calibration, compare percentage humidity with ADC value “when jumping” you will find that ADC value isn’t changing much.

[mutleyhehe]

Hi again, I've been playing around with calibrating the sensors. One of them seems to be behaving more normally. The other two (one indoor, one outdoor) are still giving strange results. Maybe let's focus on the indoor one, to troubleshoot.

I don't think it's to do with the soil not being dry enough. I let the plant go without water for 4 days and the soil was very dry and the leaves were are already starting to die out at the tips.

The first screenshot shows the soil moisture flatlining at 40% over the past day or so. The second screenshot shows the ADC results.

The second screenshot also shows the air humidity being 73% and the Lux is 0lx. Neither of these is accurate. I have numerous Aqara humidity sensors and they are showing the room (and the apartment) is 55%. Even the Soilsens in the next room is reporting air humidity of 63%. It's currently still broady daylight and plenty of sunshine in the room.

The third screenshot shows the Lux behaving erratically.

[PricelessToolkit]

Hi, just to clarify a few things:

The luminance value is not a calibrated measurement, it’s just an approximate indicator to give you a basic idea if the sensor is in direct light or not. Please don’t expect it to match a lux meter or give precise daylight readings.

Regarding the moisture percentage:
If the sensor is not calibrated, then the percentage values won’t make sense, they might jump, flatline, or behave unpredictably. That’s expected until calibration is done correctly.

The Moisture ADC value is raw and does not depend on calibration. To test your sensor hardware:

1. Note the ADC value when the soil is completely dry.

2. Note the ADC value again when the plant is watered.

If there's a difference of ~200 or more, your sensor is working fine. The issue is with calibration, not hardware. Most likely, when you calibrated the sensor, the dry and wet ADC values were too close to each other, which causes the incorrect readings.

You can also test the sensor in air:
Just hold the soil probe tightly in your hand (no gaps), trigger a reading and check the ADC value. Then release and trigger again. You should see a clear ~200 ADC change.

I don’t know how to help more if the calibration isn't done properly...

[mutleyhehe]

Hi many thanks! I will try and do some more testing tomorrow if I get a chance. I realise I am being slow to get up to speed with this. I've never used a soil sensor before and this is all quite new to me.

I did actually do the original "handheld" ADC tests you suggested, but I accidentally omitted them from my previous post above. But here were the results:

• when the sensor was in the air, the ADC was 3650
• when I squeezed tightly with my hand, ADC went to 4100
• I put the sensor into the soil, ADC was 3400, and I completed the dry calibration.
• I watered the soil, and after it completed the wet calibration, I saw the ADC was at 3900.

So taking everything into account, if I understand correctly:

• user error with the calibration, because the dry and wet measurement were too close together. Dry should have been around 3200 and wet should have been around 4200.

• the luminence or humidity sensors cannot be relied on to give accurate readings. Particularly since luminence sometimes reports zero lux in broad daylight and the humidity sensor is reporting significantly inaccurate readings.

[PricelessToolkit]

The luminance value is not calibrated, and even if it were, it still wouldn’t be truly accurate. To measure light properly, a sensor needs a light meter white dome, which smart sensors like this don’t use. That kind of setup is only found in professional tools like this one:
https://global.sekonic.com/sekonic-l-308s-flashmate-light-meter/

The humidity sensor is relatively accurate, but it’s important how and where it’s placed. You can’t compare it directly to another sensor in the same room unless both sensors are very close together (1–2 cm). Also, keep in mind this humidity sensor is only 2–3 cm above freshly watered soil, so it naturally reads higher than a wall-mounted sensor across the room.

The greater the ADC difference between wet and dry points, the better the accuracy of your calibrated moisture readings. Different soils give different ADC values, compactness, composition, and moisture retention all affect that.

[PricelessToolkit]

"Particularly since luminence sometimes reports zero lux in broad daylight" - outside ??

Sensitivity is a different thing, plus the sensor has directionality.