This is a placeholder topic for “Monk Makes Plant Monitor - Capacitive Moisture Meter - Temperature & Relative Humidity” comments.
If you're like me and have trouble keeping your calatheas and pepperomias happy and not overwatering them to root rot ruin, why not take a gander at…
Read moreI’m giving this soil moisture sensor a test. On the bench, using wet tissue paper it responds quickly and appears to track wetness well as the tissue dries up. In a pot plant it doesn’t work so well. Soil that seemed fairly moist to me registered 0% for wetness. So I gave it a real good water. I’m taking readings every 60s and the data is mainly oscillating between 100% and 0% for a few minutes, then mainly 0% with the occasional 23% or 15% thrown in. I’ll run the test overnight, but has anybody else seen these kind of results? BTW the temperature and humidity are good and I use the ‘j’ command to return the entire json payload…
So overnight, for about 8 hours it oscillated between 0 and 100% and then settled near zero. The temp and humidity data still seems fine. Wetness is shown in the right chart in yellow.
The above chart only shows 1 sample every 15 min. The below data shows 1 min samples as an example of the oscillations.
Thu 19 Dec 19:21:03 | SoilSensor/stat/SOIL : {"wetness":5, "humidity":60.55, "temp":24.09}
Thu 19 Dec 19:22:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":57.26, "temp":24.08}
Thu 19 Dec 19:23:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":60.04, "temp":24.05}
Thu 19 Dec 19:24:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":56.15, "temp":24.07}
Thu 19 Dec 19:25:03 | SoilSensor/stat/SOIL : {"wetness":100, "humidity":56.36, "temp":24.07}
Thu 19 Dec 19:26:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":56.14, "temp":24.05}
Thu 19 Dec 19:27:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":56.80, "temp":24.04}
Thu 19 Dec 19:28:03 | SoilSensor/stat/SOIL : {"wetness":100, "humidity":57.50, "temp":24.01}
Thu 19 Dec 19:29:03 | SoilSensor/stat/SOIL : {"wetness":100, "humidity":55.46, "temp":24.02}
Thu 19 Dec 19:30:03 | SoilSensor/stat/SOIL : {"wetness":100, "humidity":54.88, "temp":23.99}
Thu 19 Dec 19:31:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":54.48, "temp":24.02}
Thu 19 Dec 19:32:03 | SoilSensor/stat/SOIL : {"wetness":100, "humidity":58.09, "temp":23.96}
Thu 19 Dec 19:33:03 | SoilSensor/stat/SOIL : {"wetness":16, "humidity":56.57, "temp":23.97}
Thu 19 Dec 19:34:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":56.28, "temp":23.96}
Thu 19 Dec 19:35:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":52.94, "temp":23.94}
Thu 19 Dec 19:36:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":56.85, "temp":23.89}
Thu 19 Dec 19:37:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":57.05, "temp":23.87}
Thu 19 Dec 19:38:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":55.41, "temp":23.87}
Thu 19 Dec 19:39:03 | SoilSensor/stat/SOIL : {"wetness":100, "humidity":58.66, "temp":23.86}
Thu 19 Dec 19:40:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":59.36, "temp":23.82}
Thu 19 Dec 19:41:02 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":56.80, "temp":23.85}
Thu 19 Dec 19:42:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":61.14, "temp":23.80}
Thu 19 Dec 19:43:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":56.85, "temp":23.82}
Thu 19 Dec 19:44:03 | SoilSensor/stat/SOIL : {"wetness":100, "humidity":59.45, "temp":23.77}
Thu 19 Dec 19:45:03 | SoilSensor/stat/SOIL : {"wetness":10, "humidity":55.86, "temp":23.79}
Thu 19 Dec 19:46:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":56.19, "temp":23.78}
Thu 19 Dec 19:47:03 | SoilSensor/stat/SOIL : {"wetness":9, "humidity":56.46, "temp":23.78}
Thu 19 Dec 19:48:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":56.77, "temp":23.76}
Thu 19 Dec 19:49:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":56.67, "temp":23.75}
Thu 19 Dec 19:50:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":57.34, "temp":23.73}
Thu 19 Dec 19:51:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":54.25, "temp":23.75}
Thu 19 Dec 19:52:03 | SoilSensor/stat/SOIL : {"wetness":100, "humidity":56.04, "temp":23.71}
Thu 19 Dec 19:53:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":55.52, "temp":23.72}
Thu 19 Dec 19:54:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":56.88, "temp":23.69}
Thu 19 Dec 19:55:03 | SoilSensor/stat/SOIL : {"wetness":0, "humidity":55.14, "temp":23.67}
Thu 19 Dec 19:56:03 | SoilSensor/stat/SOIL : {"wetness":8, "humidity":57.18, "temp":23.64}
Thu 19 Dec 19:57:03 | SoilSensor/stat/SOIL : {"wetness":100, "humidity":57.80, "temp":23.62}
Thu 19 Dec 19:58:03 | SoilSensor/stat/SOIL : {"wetness":30, "humidity":57.27, "temp":23.60}
Hey @MarkMakies,
Thanks for bench-testing this! This will be great info going forward.
It’s very odd that it responds well to wet tissue, but not to a plant pot. Can you send an image of your physical setup testing the device in a pot?
This is the second pot I tried, same result. Might be that the potting mix is too loose. I’m about to try with more compacted soil.
Hi Mark
Just butting in a bit here
Would the sensor be a bit close to the edge of the pot by any chance. I am not sure if it should make the difference you are seeing but you never know just what the moisture distribution out to the edge of the pot is like.
Cheers Bob
Third pot test. The first two were plastic and this one is a composite material. Same type of potting mix. I pushed the probe in, then removed and poured water down the slot before reinserting. The mix is loose as well like the others. This pot is outside, whilst the other two were in the greenhouse which is a lot hotter. First two tests used USB plug pack for power, this one is using a power brick. These are the only differences I can think of.
To my surprise, this third test was an absolute pass, settling in only a few minutes to a stable value of 37.
12:20:02.252 MQT: SoilSensor/stat/SOIL = [{"wetness":18, "] ...
12:21:02.254 MQT: SoilSensor/stat/SOIL = [{"wetness":76, "] ...
12:22:02.287 MQT: SoilSensor/stat/SOIL = [{"wetness":53, "] ...
12:23:02.289 MQT: SoilSensor/stat/SOIL = [{"wetness":48, "] ...
12:24:02.289 MQT: SoilSensor/stat/SOIL = [{"wetness":45, "] ...
12:25:02.291 MQT: SoilSensor/stat/SOIL = [{"wetness":44, "] ...
12:26:02.295 MQT: SoilSensor/stat/SOIL = [{"wetness":43, "] ...
12:27:02.251 MQT: SoilSensor/stat/SOIL = [{"wetness":41, "] ...
12:28:02.254 MQT: SoilSensor/stat/SOIL = [{"wetness":39, "] ...
12:29:02.258 MQT: SoilSensor/stat/SOIL = [{"wetness":39, "] ...
12:30:02.260 MQT: SoilSensor/stat/SOIL = [{"wetness":38, "] ...
12:31:02.262 MQT: SoilSensor/stat/SOIL = [{"wetness":38, "] ...
12:32:02.266 MQT: SoilSensor/stat/SOIL = [{"wetness":38, "] ...
12:33:02.272 MQT: SoilSensor/stat/SOIL = [{"wetness":38, "] ...
12:34:02.273 MQT: SoilSensor/stat/SOIL = [{"wetness":37, "] ...
12:35:02.276 MQT: SoilSensor/stat/SOIL = [{"wetness":37, "] ...
12:36:02.281 MQT: SoilSensor/stat/SOIL = [{"wetness":37, "] ...
12:37:02.281 MQT: SoilSensor/stat/SOIL = [{"wetness":37, "] ...
Hi Robert,
First I had it real close to the edge, then moved it inwards with no difference in test 1 & 2
Hey all,
Good idea regarding the moisture distribution @Robert93820! The device from DFR is actually to position it as close as possible to the edge of the pot, with the “front” of the device facing outward.
I did notice @MarkMakies, the WLED on the face of the device looks to be off in both images, contrary to the DFR test images which see it either red or green. See below. Is this normal? I wasn’t aware of an “off” state for the LED in this case.
Re: the power changes in the third test, I’ll look into power consumption. I would have thought the bank was perfectly fine, but its possible it may have caused an issue.
How hot is the greenhouse exactly? I would think that could be the cause of the issue, depending on the actual temperature.
LED is off because I turned it off with the ‘l’ command.
Today the green house is about 32 degrees and the last test at 24 degrees.
No, power bank was never an issue - it’s the test that gave the good results.
Hey Mark,
Thanks for clearing up the LED.
I must’ve swapped the power bank test around when reading, all good.
I can’t imagine those temperatures could be causing issues, particularly for a device designed to work in hot/humid/wet environments.
Typically I would consider testing this at our warehouse, but I don’t believe we have any planters anywhere.
Hi Zach
If Max current is 15mA it could well be the issue. These devices are meant to charge another device in the absence of a mains charger.I believe “end of charge” is detected by a reduced output current requirement.
At 15mA max it is quite probable the power bank was switching on and off. I don’t know how many times this would happen before switching off completely.
This subject has come up many times and usually results in problems caused by using a device for a job it was never designed to do.
Cheers Bob
Just noticed this. Surprising, as in the past it has usually proved to be THE issue.
Another EDIT.
Is this a 5V power “supply” or a phone charger of some flavour ie Samsung etc. If a phone charger it is possible it would like to “see” a phone at the end of the cable to function properly.
The possibilities are pretty endless and without knowing EXACTLY what all these devices are I think most answers are a bit of guesswork.
Hey Zach & Bob,
Problem now identified (So I thought)., thanks to this discussion.
It’s a power issue, but what?: Plug pack 4.45V, power brick 4.70V. The plant monitor is powered at 3.3V via the dev boards’ LDO regulator, so I wouldn’t think it would be an issue but looking at the power into the plant monitor up we see:
Yellow plot when connected to brick (which works) red plot connected to 12W USB plug pack (which doesn’t work). The specs say it should work down to 2.4V. I’m sampling at 2GSa/a so I don’t think the dip goes lower then what I can see on the scope. So it really shouldn’t be an issue.
So I soldered a 4u7 tantalum and 100n monolithic across the plant monitor supply pads, believing problem is solved, but got the same results. I’ve tried different USB cables/plug packs, up to 45W, but again same result. Then back to power brick reference and it’s perfect.
So I’m really stumped. Considering it measures capacitance, could the fact that using a USB plug packs, which is grounded, is now be measuring capacitance between the real ground and one of the ‘probes’ in the pot? I had an isolation transformer somewhere around here but haven’t been able to find it to test this theory.
I’ll take that theory back, there is no ground pins on my plug packs, duh!
4 different plug packs tried: Belkin 12W, Apple 10W, Anko QC 15W 3A, Jackson 45W.
Hey @MarkMakies,
Interesting… I have two questions:
You mentioned the scope is measuring power to the plant monitor, is this power instability mirrored on the 5V power to the dev board?
Does the timing of the dip correlate with something? Does it happen randomly, or periodically?
You mentioned the scope is measuring power to the plant monitor, is this power instability mirrored on the 5V power to the dev board?
I didn’t test
Does the timing of the dip correlate with something? Does it happen randomly, or periodically?
It periodic, maybe once or twice a second, but I don’t think it’s related to the plant monitor as I’m only polling it once every 60s. More likely something on dev board (DOIT ESP32 DevKit 1) as I’ve noticed this before,and traced it to Wi-Fi transmits.
My next step is just to use my own regulator. I just found a MIC29510 3.3V LDO (350mV) in my stash, it’s overkill at 1.5A, I’ll let you know how it goes soon.
Mark.
Are you measuring with 10:1 probe on the scope or 1:1. If 1:1 the bandwidth might be reduced. I think you have a larger pulse than what is showing as you are triggering reliably. If you look carefully you will see the trigger point is lower than the scope trace gets. I am assuming negative going slope for the trigger.
If you extend a line down from the top trigger indicator where it crosses the trace is the trigger point. As you can see this occurs where the 2 traces are the same amplitude (approx). This corresponds to approx 3.15V. Yet the trigger level is set to 2.9V (right hand indicator).
You seem to be triggered reliably so this suggests the signal is actually about 2.9V at the trigger point with the further likelihood the actual pulse is reducing to quite a bit below 2.9V.
What flavour scope is that. Are there any more values on screen which might be of assistance.
As I have said several times in the past getting full appreciation of what you are looking at is having some knowledge of the relevant limitations of the instrument in use. All instruments are useful if you know these limitations.
I am not saying I am absolutely correct here but I don’t think I am far out.
You could re scale that trace by picking the trigger point then using this as 2.9V rescale the trace to suit. Use percentages don’t just move the trace down. You will probably find the lower level to get to about your 2.4V or even a bit lower. Would be an interesting exercise.
Cheers Bob
PS. If you get to the bottom of this you will possibly find you have fixed quite a few future problems that have not even happened yet.
Hey Bob,
Thanks, I’ll give it a try as the problem is getting stranger.
Exactly the same results: 4 different USB plug packs all fail, old power brick doesn’t miss a beat. I’m getting out the scope now.
Bob,
10x probes, pretty much followed your suggestions I think. Top plot is incoming 5V to regulator, bottom plot is 3v3 out to board. Exactly as you suggested the problem is worse than I thought.
BUT 3v3 supply into module through soldered regulator looks good. 17mV ripple and no lower than 3.26V, so I now can’t explain why it is failing. I’m just about to repeat same with power brick for reference.
And you can get our latest projects and tips straight away by following us on:
