This is a re-upload of video#207 of June 17, 2018. Some other solutions proposed by the viewers are: - Connect an output pin of the MCU instead of VCC to the sensors and switch it to high only for measurement. This reduces the time the current is flowing - Use other probe materials like steel, nails, or even gold-plated material The "green stuff" is copper-oxide for sure not healthy, but it should not hurt the plants in minor concentrations. The same happens at the Statue of Liberty, BTW. Some viewers also used nail polish to protect the edges. And it seems that sensors with manufacturing errors are in the market. Search for the "Flaura - Smart Plant Pot" channel
I don't know if it will help extend the life of the sensor, but another possible optimisation is to power the sensor from the MCU's pins such that the polarity of the sensor can be flipped in software. Then, when a reading is needed, a polarity can be chosen at random to take the reading. The assumption is that the electrolysis reaction can be reversed by reversing the polarity.
It is not copper oxide. CuO is black and Cu2O is red. It is some mixture of copper hydroxide/carbonate. Maybe Cu2(OH)2CO3 and some other copper compounds also.
I tried to measure moisture 45 years ago with two wires, and found out it worked for about a day. When I asked my teacher, he introduced me into the wonderful world of electrochemistry. Thank you, Jos Berenzten!
Must have been a good teacher, that you still remember this situation! I had a teacher who helped me build an FM transceiver, and you see what I became ;-)
A range of these capacitance sensors were tested on the Flaura - Smart Plant Pot video. The youtube algorithm worked this time! This channel shows that many are made with flaws and he shows the flaws (82% of his test sample), how to choose when buying, and how to fix the flaws if you have one. It is definitely worth watching.
finding a good design and ordering your own PCBs should be trivial, this will save you headdache of searching for good ones and it will times more cheaper.
Dude, change the electrodes for graphite, and also implement fast read instead of continuous read and you'll have less power drain and waaaay less corrosion, i mean a milissecond read in 5 minutes = 300.000x less corrosion compared to continuous.
Great video, I tried explaining to a colleague in our makerspace why the moisture sensors they used for literally all the plants are bad for the soil and everyone was complaining what I was such a downer and that it should work fine because 'everyone does this' and 'why would they be so massively sold if they dont work'. Well, now we know why. Anyway, the moisture sensors all failed within a year and some plants died. Copper oxide is really bad for plant life. I'll be sure to share this video to prove my point, but on top, the best part of this video is that you offer a solid solution. I will get a bunch of these sealed sensors and seal the PCB edges and components with some additional epoxy coating. I might even design my own since you shared the schematics and it looks like this could also work with copper enameled wires since these some pre-insulated and might be easy to wrap around a customly designed 3D printed part (no worries, I use PolyCarbonate for the printed parts so it won't degrade like PLA).
What I can recollect about Soil moisture sensors is that you need to either operate them with an AC signal, or periodically run an AC signal through them to prevent the electrode from destroying itself, particularly sensors like the water mark that uses a gypsum (I think that's what's in there) in a small canister and if you don't there is a migration effect that will eventually destroy the sensor. Sorry, wish I had more details I hope that helps. The capacitive sensors are probably the best for you applications, but the cheap ones could potential work if you treat them right.
Thanks for your additions. We also have to distinguish between measuring moisture and humidity, and I am not sure if gypsum is used to measure moisture.
@@AndreasSpiess 1. The watermark. made by Irrometer, is a soil moisture sensor, it's a green plastic outer shell with a stainless steel screen and there's a white powder in it, I overheard someone say it could be gypsum... maybe it's doped with something? It is probably a company secret. 2. There was also a leaf wetness sensor that was gold plated with a bunch of fingers that wouldn't touch and the moisture in the air would "bridge" that gap. 3. The Last kind of sensor is a Humidity sensor and their capacitive, long ago this was done (in accurately I will state) with a 555 and was tuned in the factory (painful to do) the newest sensors by Senterion DHT (Digital Humidity Temperature) sensors come already calibrated.
@@olfo4100 You don't... this was a low power application running on batteries, but the generated AC is used to prevent migration of material from one pole to the other. Even powering for short bursts in the same direction migrates the material and the sensor looses the ability to read over time... it might take a while to happen.
What if you switch the polarity (but still use DC) from one measure to the next? And, obviously, turn them on for a few milliseconds at a time (when you make a measurement) and not run them 24/7?
@@NicksStuff Yep that's kind of the idea, or you could run a small back current or AC signal after you take the readings, and since soil moisture is a slow moving value so you can measure infrequently, like every 15 or 30 minutes... Depending on your needs.
Hi Andreas! Just a comment about the home made PCB. It is very easy for copper clad FR4 to make 2 cuts through the copper down the center close together with a razor blade. Then lift a corner of the copper with the razor blade and grab it with a small pliers. It can usually be stripped off fairly easily. It should only take a minute or so to do.
Hi Andreas, Always a pleasure to view your videos. I have played with quite a few of the capacitive sensors. I found that with some of them, the 1M ohm resistor that is supposed to be connected between GND and SIG is there but is NOT connected to GND! . Seems to be a design defect as a whole batch I ordered from the famous Chinese supplier had the same problem. As a result, the reading was false. To correct it, I just soldered a 1M ohm resistor between the two pins on the connector. Easier than trying to correct it on the PCB with the surface mounted components.
Great video! I did some research a while back into these and came to a similar conclusion, that both of the main sorts are fatally flawed. Coating the capacitive sensor with a stronger layer (e.g. dipping in epoxy), or pulsing the conductive sensor for very short times were the results I came to. I’m not sure what the corrosion without any current is, but I figure it’s pretty low. A gold-plated sensor might last long even with a continuous current. As for the toxicity, considering the nonzero chance of that plating containing lead, I’d be very hesitant to use an off the shelf Chinese conductive sensor. Don’t want my broccoli to start lowering IQ points. Tin and nickel, and especially copper, can be toxic to plants too, but with the small concentrations after rain washing things it’s likely not a big deal. Actually, some people deliberately corrode copper wire in their soil to increase the amount of copper in the soil. Apparently copper deficiency is common enough in plants that it’s a cheap way of improving plant growth, though I haven’t looked into it myself.
@@galaxya40s95 I suspect that will just double it, though you might find that it replates some of the lost metal back. Some materials won’t corrode, mainly inert metals like platinum and iridium, gold is probably pretty good too. But there’s also some conductive oxides, like indium-tin-oxide (found on a lot of LCD and probably OLED displays), and lead dioxide (as acts as one electrode in a lead acid battery). All but the lead are expensive, and the lead is too toxic to put in your soil.
Just stating the obvious here: Don't add copper to your soil unless you know it is deficient in it. Too much copper is also bad for plants and other wildlife.
@@hamjudo this is definitely true. Especially since in acidic soil the copper will be pretty soluble and easily absorbed. Copper carbonate is (I think) a safer less-soluble form of copper to add to soil, I’ve heard of people converting their used etchant to carbonate before disposing at least. I convert it to copper metal.
I have one project where I decided to solve this problem reading the sensors very quickly, and then turn-off the circuit where the sensors are connected. First the Arduino turn the circuit on, then I made 64 readings in less than a second, and calculate the average value of the readings. After that I turn off the circuit. As I have to make just one or two measures per day, I think this is a good workaround.
Very good tuto, as usual. You are our Master Theacher. Another good solution to protect the capacitive sensor pcb from moisture would be to cover the edges with neil varnish.
Wow, this video is incredibly fascinating! I'm really intrigued by the concept of building a multilevel soil sensor after watching this. However, I'm not quite sure where to begin or how to go about it. It would be amazing to see a video from you detailing the process and offering guidance on how to create one. Your content is always so informative and inspiring!
Grutti Andreas, If you look at the schematic at 8:28 - the exposed "sensor-electrodes" still have a DC on them, albeit being used capacitively. It is a simple addition to place two large-value film-capacitors in series, one on each of the electrodes, to effectively"DC-isolate" these from any active electronics. The plasti-dip or whatever then has to do less work (esp. with pinholes etc. still being able to ruin the sensor). As the two series capacitors are much larger than the soil-dependent capacitance, they do not affect the frequency-modulation.
I agree with your proposal. However, some viewers suggested that the PCB lasts only for a short time too. So maybe a small plastic bag around the sensor would help...
Because of the exposed electronics on that board, the fact that the PCB sides are still exposed, the fact that in some boards the 1M ohm resistor is not connected on one side and the fact that some of those sensors come with an NE555 rather than a TLC555 and the fact I like to experiment with moisture sensors, I have used the PlastiDip method, the plastic bag method (as also suggested by Andreas) and I have rebuilt the circuit, but with the electronics separate from the sensor and the sensor connected via relatively large capacitors. In fact all 3 methods work pretty well. My most favorite method is still to use some carbon rods to measure the soil resistance, but I am well aware that the latter basically measures ions, whereas the capacitive method measures water. If you are really serious about using the capacitive method, I would indeed rebuild the circuit and add 2 capacitors as described, put the electronics in a water tight enclosure, with a short shielded cable to the sensor itself. On top of that you could also switch off the power supply to the circuit and thus to the sensor. But in real honesty....just two carbon rods do fine for home use
@@vazquezjm No, carbon rods, not 'center parts of pencils'. The latter would break when you put them in the soil. Schematics???? really: one goes to the analog in and the other to either ground or Vcc
Excellent! I had to correct my shopping cart after watching this video. Capacitance types as mentioned in Flaura's video and yours will be the choice of sensor.
Using stainless steel will help with the longevity of the sensor. This type of sensing is good for a small container where you don't care too much about the moisture content at different depths. One could use several of these at different depths in order to construct a map that describes the soil's behavior to watering. This information can be used to optimize water usage and similar concept is used in professional sensors.
Stainless steel rods/probes have been used for resistivity and Induced Polarization surveys for a long time. As long as you alternate the polarity you get a good reading and can measure the change in moisture over a large area. Like the video, lots of trash on the market.
I'm wondering if there are also moisture sensors that work on the principle that periodically a small resistor gets heated by electric current and then a circuit measures how long it takes for the resistor to cool down again, which should be dependent on ground moisture levels.
This is how the old gardena probes work, and work well. I don't know about the new smart probe, which is flat instead of round. The smart probe and gateway cost 300€, so expensive to test.
Nice re-up, Herr Spiess. :-) To further enhance this widely used humidity sensor, please have a look for the page titled "Hacking a Capacitive Soil Moisture Sensor (v1.2) for Frequency Output" from the Cave Pearl Project. There you will find 3.3V capability, how to use higher frequencies and some more tips and hacks around this sensor. In short: instead of the sensor with the NE555 chip you should favor the version with the CMOS TLC555 chip.
I have tried all the sensors shown as well. For the capacitive one, it consistently gave "wet" readings even when the soil had completely dried out. It seems like the moisture sticks on the sensor surface, it couldn't dry off quick enough as compared to the moisture in the soil. I had to remove the sensor to wipe it down for it to show the correct readings again, which makes it completely pointless.
@@AndreasSpiess There were also a few occasions when I pulled out the sensor, I could not spot dampness with the naked eye, but after a quick wipe, the readings went back to normal. Could mineral deposits also affect readings, I am not sure. I can say when testing outside soil, the sensor always works perfectly when trying with water. I have since given up trying to automate watering.
Could you propose a method to calibrate the output characteristics for the particular soil type, installation depth, and soil compaction? In other words can we compensate for these effects in order to improve repeatability? Maybe comment on use as tank fluid level sensing. Thank you.
1. I am not sure if they can be calibrated for precision. At least, I did not find any information about it. It is more to show you when you have to water your plants 2. I made a video about measuring water tank levels. Capacitive sensors are used there, too. They are longer.
There is the "tinovi I2C Non-Contact Capacitive Soil Moisture, Temperature sensor" which works very well (full waterproof). I wrote the ESPhome component (avalable in the last ESPhome dev)
Very nice view over the lake. Switzerland is beautiful. I have had the same problem with "open" sensors. I took bicycle spokes and solder it on the sensor. It works OK.
Loving the SHT-35 for humidity, so far at least. They seem to be quite a bit more robust than the old DHT sensors. I run 2 of them 24/7 to add humidity to my plants via a relay-controlled humidifier.
I think that is just good old electrolysis happening there, the direct current will break apart the water into Oxygen and Hydrogen and the residue probably has to do with the impurities or minerals contained in the water that can also react.
Well use missed the two most important principles for measuring soil moisture. 1. Time Domain Reflectometry (TDR). 2. Frequency Domain Reflectometry (FDR) / Capacitance (Frequency) Both are state of the art when measuring soil moisture within scientific context. When I saw your video I hoped that you have managed to build one. 😅
I know that these produce better results. But unfortunately, they are much more expensive and rarely used for the hobby. Even to build one is not easy. Or did you find a project?
That is correct, they are expensive as hell. I hoped that your expertise is sufficient to build one yourself since the principal is well documented. 😅 I started reverse engineering one but never finished. BUT I will do that in the future, at least I will try.
I've never used them. I was expecting you to use Proximity sensor with it's sensitivity set extremely low. I have no idea if it would work but I'm my head it makes sense. Nice video.
Good morning Herr Spiess, I hope your summer holiday is going well, we're missing your fresh content every Sunday morning :) WRT capacitive soil sensors - I coated their edges in 2 layers of nail polish but I was surprised to still find their _surface_ delaminated and corroded after little more than a year in use. Not going to buy again, they're kind of a lost hope.
@@awaagrikh8331 maybe mine also delaminated and corroded earlier, I noticed the readings started drifting after just a few months but I took them out of the pots about a year later since by that time I wasn't paying any attention to them anymore.
NICE HAVE NOT SEEN THIS BEFORE THANKYOU. Have many old stuff from this time and have start growing my own food salads stock for added sustainability in a ever changing eco system, thinks I can grow indoor in window light controlled systems but as you know controlling these systems are harder than people think you for replay sir.
In production horticulture we use a neutron probe, but they are expensive. I have found the capacitive probes have a narrow range and are easily disturbed, somewhat difficult to calibrate IMHO, but definitely a cheaper option!
I remember getting the "chirp" sensors that ran on 3V coin cells back in - well, long ago. What I missed on these capacitive sensors was square wave output - I would assume that would be easier to transmit over longer distances. I've seen I^2C and RS485 versions, but they are pretty expensive.
Some people add an ESP8266 directly to the sensor and use the air for transmission ;-) Transferring the square wave might also be problematic. It could catch some noise over the distance.
@@AndreasSpiess I was thinking about that, but them mildew killed most of the plants so I didn't follow that up back then and forgot about it during the pandemic. My students brought that up as an add-on to the school garden, and now I'm working on it again. As a physicist turned teacher, I am not on good terms with analogue signals over long distances. Measuring pulse durations and calculating a median sound better to me (I might be wrong here, though - maybe I'm just bad at analogue wiring)
Or use a differential analog signal. This increases the complexity a bit on both the sending and receiving side, but because the signal is differential interference will not be a problem. When the cables become very long then the transmission line effects can become non-negligible.
@@AndreasSpiess "use the air for transmission" WiFi eats battery juice quite fast. Be prepared to wire your entire garden with VCC or replace the coin batteries multiple times a day.
I've used a similar circuit on the Farm-Data-Relay-System attached to a Watermark Soil Moisture Sensor. I've used the sensor for years and it works well. I also make Gypson Moisture Sensors that I have described on Printable. They provide similar readings but they tend to dissolve over a period of time. Watermark sensors are very expensive but are worth the money for monitoring my foundation in Texas.
It's super easy to build this kind of sensor yourself anyway. Just get copper rod, solder wires to two pieces, add the 500kOhm input resistor, to one and a ground wire to the other. Add a second wire to the one with the 500kOhm directly to an analog in of your microcontroller. That's it. You can have either the v+ or ground coming from a toggleable port so you can save a bit of life on your sensor.
Hi Andreas, I have tested the capacitive moisture sensors quite extensively in aquaponics. I use a ADS1115. I found considerable variation due to tempurature fluctuation.
Thank you for sharing your experience. Other commenters confirmed your finding. I do not think these sensors are suitable for precise measurements. But they are okay to create signals for watering.
Any electron flow in a damp environment will cause material to migrate either from your plate or to your plate. None contact is the only way to go, or treat the sensor like an LCD display where there it has to be treated the same, an AC signal with no DC component, or you destroy the segment.
I bought the Xiaomi Miflora Sensor wich has also a lot of knockoffs wich are cheaper, Monitoring it for 7 Montths now and it works fine and reliable + has integration features Battery still shows 100% You pay a little extra for the enclosure ( water resistant ) and the ease of use ( not encrypted BLE ) and a nice byfeature is the EC measuring wich gives a rough ( not really useful ) estimate when its time to give a little more fertilizer or if there is too much salt build up Integrated in HomeAssistant with Automated Watering learning based of the dry / wet cycle and times it takes the plant to take up the water added I couldnt have or use more features and its working out of the box, no need to ask for more after fiddling around with diy solutions for years im happy that this project is now running and working
Thx for this topic... It's long time we develope this.. and have so many alternative.. But sometimes, its also depends on media that need to be measured..
I use a very good nail varnish to cover all the moisture sensor v1.2 including the chip and the pins and wires and works fine since 2-3 years now. But there is a missing connection between ground pin and the a ground trace. Simply solder a small diameter wire on both side of the pins header. Cheap China stuff but works.
Cool DIY sensor ;-). But seriously, I think a capacitive DIY sensor can be a good option for makers and you can easily create a square wave signal with a microprocessor board.
I have a commercially made sensor (Ecowitt WH51) that appears to have a continuous loop that goes into the soil. Perhaps under the coating it's not a continuous conductor though. It eems very reliable but now I want to see what is inside.
you can use the cheaper ones without their electronics, and alternate the current thru them. the main problem with all these sensors is their resistance depends on the soil and the minerals present. takes allot of effort to calibrate them on specific soil now try it with changing soilminerals (rain water vs tap)
What about sensing by weight instead of moisture level? A flower pot will have a certain consistent weight. As moisture evaporates, the weight will be reduced. At a certain point, you turn on the water. You might have to re-calibrate from time to time as the plant grows.
Hallo Andreas, der Beitrag kommt genau zur richtigen Zeit :-) bin gespannt, ich habe einen mit zwei V2A Stahlröhrchen "gebaut" :-) Den Kapazitiven Sensor hatte ich auch schon mehrfach verbaut, leider hat sich auch hier nach einiger Zeit die Farbe / Isolierung gelöst und die Kupferbahn aufgelöst, nicht gerade gesund für die Pflanzen
@@AndreasSpiess im Baumarkt gibt es streichbare Faserverstärkte Dichtmasse, die hab ich diese Woche zum Dach abdichten verwendet ich glaube ich streiche mal so ein Teil damit ein und teste es. Damit könnte man auch die Elektronik am oberen Ende abdichten. Ich berichte dir :-)
I have used that cheap sensor on my Christmas tree to warn about low water level for many years without a problem. I realised this design error and implemented very simple correction: I'm not powering it all the time. I use Arduino output pin to power sensor only for a couple of ms only when I do measuring once per each hour.
Hello Andreas, Gardena (Husqvarna) uses a different approach for its irrigation/sprinkler system. Its moisture sensor contains a metal tip that is enclosed in felt. This metal tip heats up periodically (about every 30 minutes) and, based on its temperature decay (cooling time), the degree of soil moisture is calculated. When "wet", the connection between two pins is "closed," when dry the connection stays "open". The two pins are connected to the irrigation controller. Why periodically - to allow for battery operation. Greetings, Stefan.
The first real Arduino project I built used these resistance sensors. I was so proud of myself, until the 3rd and 4th days when the sensors were ruined.
You can use the first pcb very easy with an arduino and no electrolysis. Connect the pcb between to analog inputs. For example A1 and A2. In the first measurement make A1 output high, it will supply now 5V to the sensor, read the analog value from A2. For the second measurement you do the opposite, make A2 digitaal out high and read the analoge value from from A1. Repeat the above. With this way you measure a sort of with ac and the electrodes with not desolve. If the reading of the adc is to low, try setting the reference to 1.1V or other value.
Did you check the RF-emissions of the capacitive sensor with a SA? I think there might be a lot of harmonics causing QRM in the HF bands. Maybe not relevant for the average houshold, but in fact, for example I've turned off my weather-station because of spurious emissions in the ISM-band, disturbing my satellite groundstation. So, maybe as a HAM-operator I'm a little bit over-sensitive to this . Anyway - nice video, thanks!
Andreas - good to see this topic again. I was a disappointed to learn that even capacitive sensors fail over time if they are just coated with resin, as the resin still allows water intrusion via osmosis. I wonder if that includes all forms of epoxy. Perhaps the probes and the base of the instrument body could be coated with "Plasti-Dip"? The plastic bag is a fine idea, I guess, but really the sensors should be fully sealed in plastic until the whole instrument is up out of the soil and well out of the way of splashes, etc. So rather than putting plastic "leggings" on the sensor probes, I'm talking about plastic "pants" that reach to the top of the sensing instrument. I am a bit sloppy when watering plants, yes. :^)
The thicker the protective layer, the less sensitive the sensor is. Which means: use a very fluid resin for coating, so that it will cure in a thin layer.
Hallo aus Österreich, Andreas! Ich mag deine Videos. And "the Swiss accent" :) Wegen des Kupfers... Wenn du, so wie du es gemacht hast, die Elektroden ins Leitungs-/Brunnenwasser steckst (nimmt man ja meist zum Gießen) und mit ein wenig Spannung und einigem Strom beaufschlagst, erzeugts du einfach Knallgas. Sauerstoff und Wasserstoff an je einer Elektrode, das entstehende Wasserstoffgas kann als Blasen an der Oberfläche der Kathode freigesetzt werden. Komplizierter ist es mit Folgendem: Im Leitungswasser sind üblicherweise verschiedene gelöste Salze enthalten, darunter Natrium- bzw. Kalziumionen und Hydrogencarbonationen (HCO₃⁻), die möglicherweise reagieren könnten, um Kupfersalze wie Natriumkuprat (Na₂CuO₂) oder Kupfercarbonat (CuCO₃) zu bilden. Verschiedenen Faktoren wie Wasserzusammensetzung, pH-Wert, Temperatur und Stromstärke beeinflussen die Reaktion. Wenn man Pflanzen mit einer solchen Flüssigkeit bewässert, bzw das im feuchten Erdreich auftritt, könnte dies potenziell schädlich sein, da Kupfer in höheren Konzentrationen toxisch für Pflanzen sein kann. Kupfer ist für Pflanzen notwendig, aber nur in geringen Mengen. Bei zu hoher Konzentration kann es jedoch die Aufnahme von anderen wichtigen Nährstoffen behindern und oxidativen Stress verursachen. Die Kalkablagerungen könnten auch den Boden pH-Wert erhöhen, was wiederum die Nährstoffverfügbarkeit für Pflanzen beeinflussen könnte. Ein erhöhter pH-Wert kann dazu führen, dass bestimmte Nährstoffe weniger leicht aufgenommen werden können. Es ist wichtig zu beachten, dass die Auswirkungen auf Pflanzen von verschiedenen Faktoren abhängen, einschließlich der Konzentrationen der beteiligten Ionen, der Dauer der Bewässerung und der Empfindlichkeit der Pflanzenart.
One guy said put a 1M ohm resistor between signal out & ground on the 1.2v sensor, does that make sense? All mine, 10 of them, give me random values & don't respond when I'm water 😢
Arduino's can drive and measure capacitive changes between plates directly. There even an offical library for reading such sensors. It's also *easy* to create such plates oneself. If aluminum or titanium is used it's also possible to anodize the surface to insulate them. Far more even and resilient surface than varnish Nb the thickness of such an anodized layer adjusts the value/range of such a sensor
Even the capacitive sensor isn't good, with exposed FR4 edges. I looked into this once and Steven's Hydraprobe looks good. It gives moisture, temp, salinity over RS-485 using 50MHz RF. The cost is $695 though (!).
@@AndreasSpiess FWIW, HydraProbe isn't TDR, it uses radar to measure dielectric impedance. It measures the energy storage and energy loss of the soil to determine the moisture content. This reduces the influence of salinity and temperature compared to other solutions, allowing it to work without calibration. It's really neat and I have bought one, but I won't get to test it until early 2024. The cost is silly, but nothing is too good for my lawn. :p Mostly I just want to play with the cool tech. My DIY irrigation controller calculates ETo to determine watering durations. I'll use the soil moisture data to double check that, eg don't water if the soil has enough moisture even if ETo thinks it needs watering.
The two-legged sensor with a gold plating most better than of cheaper lead. I've tested ENIG version about 2 month, if not scratch them, no corrosion will be even with fertilizers. And power on must be very short as possible for measuring.
Great video Andreas! I've been having problems with the capacitive sensor, making my house main differential circuit breaker go off as soon as the sensor contacts the soil. I'm assuming there's some kind of leakage? Will try the plastic bag to isolate the sensor and see if it helps
Hi Andreas, did you ever try to put a Bosch BME sensor into a cavity underground like a small platic pot with the opening facing down? This would give you the relative humidity of the air in equilibrium with the soil.
I remember reading about this problem in the 1980's (probably in Electronics Australia Magazine). If I recall correctly, there was an IC manufacturer who produced a chip specifically for this application. It applied a small AC signal to the probe for measurement to prevent electrolysis. I see lots of moisture probes advertised now and just assumed they would use the AC method. I forget how much junk is sold out there.
I wonder if the resistance sensors would be better with graphite / carbon probes? Carbon is pretty inert in water electrolysis, unlike copper. Whilst they wouldn't last forever, they would significantly outlast copper traces. Also carbon is not toxic to plants whereas high concentrations of copper can be.
@@AndreasSpiess the most effective frequency was particularly interesting. I suppose this would have to be experimented with, if one was to make a level sensor where the tank material (dielectric) was quite thick (8mm). I love your videos. Always nicely informative and without annoying background (sometimes foreground!) music 👍😁
I have tested the capacitive soil moisture sensors extensively and they are not reliable whatsoever to measure anything accurately in soil, the only working solution is to make a resistive sensor that receive power only for a very short time with prongs in stainless steel or silver solder. That was my solution to manage my sprinkler system in Florida. Anyway, there is no need to measure the moisture in real time, 2 or 3 times a day is more than enough, humidity does not leave spontaneously the soil.
Another option to extend the lifespan of sensors with exposed copper conductors is to make measurements in a descrete manner. I.e. not constantly but once every 30 minutes for example.
And keep the time that it is switched on as short as possible. Usually ~10µs should be enough. Then even more frequent measurements should be possible. What could also help is to lower the voltage that it has available. According to wikipedia, water electrolysis requires a minimum of 1.23V. Measuring with max 0.5V should therefore eliminate corrosion. That may be worth testing.
Another technique can be to only power the sensor while the moisture content is being measured. No reason to read the sensor constantly, once an hour or so is more than enough. Cut the power to the sensor to only a few seconds per day will extend the life of the sensor.
If price is not an issue, i have uses smt100 and smt50 from truebner on a research project. They currently have an uptime of 4 years (but they cost 100 and 50euros respectedly)
The capacitive approach is a simplified form of impedance spectroscopy. The real fun with that is when you sweep the frequency and take a spectrum instead.
This was interesting, but might be a little out of context. This would probably be fine as a physical liquid level measure, but that's not how dirt and moisture work in most agricultural situations. If your plants roots are sopping wet, your plant is likely drowning. Usually there is much less liquid in soil, and that may make the readings much different. Could you repeat the experiment with the last sensor, using different types of soil, and adding water like a grower would? The best test scenario would be in a large planting container exposed to the outside air conditions and to the sun, but protected from direct rain. That way, there is a large amount of soil, there is air, heat, sunlight, wind, to dry the soil, and you can add water in small amounts, as if from a sprinkler or irrigation system. The moisture in the soil will need to equalize across the bulk of material, rather than just detect full saturation. Water would arrive on top, and then get wicked through the soil. Having a short or small container might give false readings from water pooling in the bottom, saturating the lower part of the sensor. The sensor needs to be affected by the correct amounts of moisture in the soil, not just "dry_ and" saturated" conditions. That would simulate regular growing medium better, and give a better idea as to the signals seen and the response time to detect actual soil conditions. _____ As an added thought, this looks like it would be a good wired system for a central garden plot. Consider a yard, though, or multiple garden plots, distributed areas, etc. Having wires running 10s or 100s of feet, in multiple directions, on opposite sides of structures, etc., and with foot traffic or vehicle traffic in the area. This would be hard to set up, calibrate, and maintain. A larger scale solution would call for wireless communications, sensor differentiation, logging, and probably power management. This is a job for SUPERMA-- Sorry, I got carried away. This is a job for: I don't know... It's got to be be cost, able to leap large distances, be identified, easy on the batteries, etc. LoRa? I don't know the costs or component size. 422mhz? ATtiny85? RaspiNano? BLE? I can think of a couple approaches. I can envision a mesh network for ease of adding sensors and extending the coverage of the network, but that seems complicated. I can see having a base station and just relying on low power rf transmitters, logging and sleep routines, and periodic polling fob the base station. I believe that BLE can operate as a mesh network. That's enough to think about now. I think I want more coffee.
When I made this video, I was looking for studies as you describe. They should be made by agriculture engineers, not electronics engineers. But unfortunately, I did not find any. That may be different now. These sensors could be more precise. I think they can be used to generate watering signals in typical home environments. If you want to avoid cables, maybe you google "LoRa moisture sensors". Their range can be a few km. You also find many videos about the LoRa technology on this channel.
@@AndreasSpiess Thank you. I am thinking more along the lines of a home owner. We aren't on a large plot, but with a growing space at the front of the yard, two near the house in front, four spread through the back yard, the thought of monitoring all of that using wires is a problem. I guess in the end, I would pick maybe two or three, but even then, that's a lot of wiring spread across hundreds of feet in different directions. As a manual (handheld) tool, that might be a more practical solution for me, but then we're back to doing things manually. Still, it's a neat video.
This is a re-play of an old video. In the meantime, it seems there are different variants available, some that do not work. I added the info in a pinned comment.
I tried several capacitive sensors-cheap one, DF Robot one and two other Chinese brands. None can survive more than 3 to 6 months even if I coated it. Soil is salty due to fertilizer usage so it greatly reduce the life span of the sensors.
sonoff ms01 is looking pretty good, it sends data trough I2C, it is on my to do list to test one in esphome and if satisfied switch from all other ones .... but it is not on top of my list ....
@@AndreasSpiess looking for some "reliable" to integrate with esphome for home assistant. Will look at it, if you know about more please let me know, because i am in fight with soil moisture sensors from spring.
I use hot-galvanized nails as probes. These are the gray, rough looking ones. Few years now, no corrosion. Avoid electro-galvanized (shiny) - they wont last.
So you made an unintentional electrolyzing cell. You are deplating the anode(+) side and plating the cathode(-) side. As well as liberating some hydrogen on one electrode and oxygen on the other. One way to make them last longer is to swap the polarity every other time you read it and only activate it when you want to read it. Nickel plating the electrodes will also help some. 😊 And for the capacitive version you might use some spray on conformal coating.
Hello Andreas I use the last sensor (capacitive on), there is a but you forgot. The value measured is not stable since it is impacted by the temperature. It creates quite big differences. I ve not solved this yet.
I agree it would be interesting to compare the readings of the (quite expensive) gypsum sensors with the ones of these cheap sensors. This is more for an agricultural channel, I think.
The first sensors can benefit from probes made of stainless steel. But all have a major flaw in that they use a lot of power and they are not the best thing for IOT standalone sensors. Also resin coatings in pcb board suffer in the long run from water permeating through the pores because of osmosis. This causes blisters filled with acid under the coating. The latter I suppose one could accept it as a wear part in the system.
Ah the classic ox redox reaction. You might even be creating metal complexes in the water (gets very complicated to know if its plain rust or a complex of Fe)
From what i gather the coated sensor used way more power so they last months instead of over a year battery wise. IOm strill wanting to have a google come compatible sensor over wifi. i dont want to have a tuya or zigby hub google home keeps things simple and working well without the need of a hub making replacing things easy and pain free.
Small problem: Clean water id more or less an insulator. What you can measure is the concentration of different salts wich gives a very unprecise picture of the soil moisture.
@@AndreasSpiess Yes, but what does the salt content in the water means for the dielectric properties? I have experimented with a small cage made of net buried in the soil. Inside a DHT 22 to measure the air humidity and tried to relate it to the soil mosture.
This is a re-upload of video#207 of June 17, 2018. Some other solutions proposed by the viewers are:
- Connect an output pin of the MCU instead of VCC to the sensors and switch it to high only for measurement. This reduces the time the current is flowing
- Use other probe materials like steel, nails, or even gold-plated material
The "green stuff" is copper-oxide for sure not healthy, but it should not hurt the plants in minor concentrations. The same happens at the Statue of Liberty, BTW.
Some viewers also used nail polish to protect the edges.
And it seems that sensors with manufacturing errors are in the market. Search for the "Flaura - Smart Plant Pot" channel
Ah, I knew I recognized the sensor in a plastic bag! :)
For resistive probes AC reduces corrosion i have found..The control circuitry is more complex however.
I don't know if it will help extend the life of the sensor, but another possible optimisation is to power the sensor from the MCU's pins such that the polarity of the sensor can be flipped in software. Then, when a reading is needed, a polarity can be chosen at random to take the reading. The assumption is that the electrolysis reaction can be reversed by reversing the polarity.
It is not copper oxide. CuO is black and Cu2O is red. It is some mixture of copper hydroxide/carbonate. Maybe Cu2(OH)2CO3 and some other copper compounds also.
Why did you re-upload the 5 year old video?
I tried to measure moisture 45 years ago with two wires, and found out it worked for about a day. When I asked my teacher, he introduced me into the wonderful world of electrochemistry. Thank you, Jos Berenzten!
Must have been a good teacher, that you still remember this situation! I had a teacher who helped me build an FM transceiver, and you see what I became ;-)
A range of these capacitance sensors were tested on the Flaura - Smart Plant Pot video. The youtube algorithm worked this time! This channel shows that many are made with flaws and he shows the flaws (82% of his test sample), how to choose when buying, and how to fix the flaws if you have one. It is definitely worth watching.
finding a good design and ordering your own PCBs should be trivial, this will save you headdache of searching for good ones and it will times more cheaper.
Thank you for the info. I added the link to the video in the description.
Do you have a source, where to get components, that fulfill those 3 criteria?
Dude, change the electrodes for graphite, and also implement fast read instead of continuous read and you'll have less power drain and waaaay less corrosion, i mean a milissecond read in 5 minutes = 300.000x less corrosion compared to continuous.
Great video, I tried explaining to a colleague in our makerspace why the moisture sensors they used for literally all the plants are bad for the soil and everyone was complaining what I was such a downer and that it should work fine because 'everyone does this' and 'why would they be so massively sold if they dont work'. Well, now we know why. Anyway, the moisture sensors all failed within a year and some plants died. Copper oxide is really bad for plant life.
I'll be sure to share this video to prove my point, but on top, the best part of this video is that you offer a solid solution. I will get a bunch of these sealed sensors and seal the PCB edges and components with some additional epoxy coating. I might even design my own since you shared the schematics and it looks like this could also work with copper enameled wires since these some pre-insulated and might be easy to wrap around a customly designed 3D printed part (no worries, I use PolyCarbonate for the printed parts so it won't degrade like PLA).
Maybe a small plastic bag would do the trick instead of coating or 3D printing. It could be replaced if needed.
@@AndreasSpiess I'm thinking, maybe using graphite as an electrode in this case might actually make a lot of sense.
What I can recollect about Soil moisture sensors is that you need to either operate them with an AC signal, or periodically run an AC signal through them to prevent the electrode from destroying itself, particularly sensors like the water mark that uses a gypsum (I think that's what's in there) in a small canister and if you don't there is a migration effect that will eventually destroy the sensor. Sorry, wish I had more details I hope that helps. The capacitive sensors are probably the best for you applications, but the cheap ones could potential work if you treat them right.
Thanks for your additions. We also have to distinguish between measuring moisture and humidity, and I am not sure if gypsum is used to measure moisture.
@@AndreasSpiess 1. The watermark. made by Irrometer, is a soil moisture sensor, it's a green plastic outer shell with a stainless steel screen and there's a white powder in it, I overheard someone say it could be gypsum... maybe it's doped with something? It is probably a company secret. 2. There was also a leaf wetness sensor that was gold plated with a bunch of fingers that wouldn't touch and the moisture in the air would "bridge" that gap. 3. The Last kind of sensor is a Humidity sensor and their capacitive, long ago this was done (in accurately I will state) with a 555 and was tuned in the factory (painful to do) the newest sensors by Senterion DHT (Digital Humidity Temperature) sensors come already calibrated.
@@olfo4100 You don't... this was a low power application running on batteries, but the generated AC is used to prevent migration of material from one pole to the other. Even powering for short bursts in the same direction migrates the material and the sensor looses the ability to read over time... it might take a while to happen.
What if you switch the polarity (but still use DC) from one measure to the next?
And, obviously, turn them on for a few milliseconds at a time (when you make a measurement) and not run them 24/7?
@@NicksStuff Yep that's kind of the idea, or you could run a small back current or AC signal after you take the readings, and since soil moisture is a slow moving value so you can measure infrequently, like every 15 or 30 minutes... Depending on your needs.
Hi Andreas! Just a comment about the home made PCB. It is very easy for copper clad FR4 to make 2 cuts through the copper down the center close together with a razor blade. Then lift a corner of the copper with the razor blade and grab it with a small pliers. It can usually be stripped off fairly easily. It should only take a minute or so to do.
Good idea! Thanks for the tip!
Hi Andreas, Always a pleasure to view your videos. I have played with quite a few of the capacitive sensors. I found that with some of them, the 1M ohm resistor that is supposed to be connected between GND and SIG is there but is NOT connected to GND! . Seems to be a design defect as a whole batch I ordered from the famous Chinese supplier had the same problem. As a result, the reading was false. To correct it, I just soldered a 1M ohm resistor between the two pins on the connector. Easier than trying to correct it on the PCB with the surface mounted components.
I heard of this flaw, too, and added a link to a video where it is shown in the description
Great video! I did some research a while back into these and came to a similar conclusion, that both of the main sorts are fatally flawed. Coating the capacitive sensor with a stronger layer (e.g. dipping in epoxy), or pulsing the conductive sensor for very short times were the results I came to. I’m not sure what the corrosion without any current is, but I figure it’s pretty low. A gold-plated sensor might last long even with a continuous current.
As for the toxicity, considering the nonzero chance of that plating containing lead, I’d be very hesitant to use an off the shelf Chinese conductive sensor. Don’t want my broccoli to start lowering IQ points. Tin and nickel, and especially copper, can be toxic to plants too, but with the small concentrations after rain washing things it’s likely not a big deal. Actually, some people deliberately corrode copper wire in their soil to increase the amount of copper in the soil. Apparently copper deficiency is common enough in plants that it’s a cheap way of improving plant growth, though I haven’t looked into it myself.
Thank you for your additional information! I agree with the dipping.
Another way to prolong the life of the electrodes is to switch which one is positive and negative with each measurement.
@@galaxya40s95 I suspect that will just double it, though you might find that it replates some of the lost metal back.
Some materials won’t corrode, mainly inert metals like platinum and iridium, gold is probably pretty good too. But there’s also some conductive oxides, like indium-tin-oxide (found on a lot of LCD and probably OLED displays), and lead dioxide (as acts as one electrode in a lead acid battery). All but the lead are expensive, and the lead is too toxic to put in your soil.
Just stating the obvious here: Don't add copper to your soil unless you know it is deficient in it. Too much copper is also bad for plants and other wildlife.
@@hamjudo this is definitely true. Especially since in acidic soil the copper will be pretty soluble and easily absorbed. Copper carbonate is (I think) a safer less-soluble form of copper to add to soil, I’ve heard of people converting their used etchant to carbonate before disposing at least. I convert it to copper metal.
I have one project where I decided to solve this problem reading the sensors very quickly, and then turn-off the circuit where the sensors are connected. First the Arduino turn the circuit on, then I made 64 readings in less than a second, and calculate the average value of the readings. After that I turn off the circuit. As I have to make just one or two measures per day, I think this is a good workaround.
Indeed, this is a good idea to reduce corrosion.
Very good tuto, as usual. You are our Master Theacher.
Another good solution to protect the capacitive sensor pcb from moisture would be to cover the edges with neil varnish.
Thank you for your kind words! And I agree with the nail varnish.
Wow, this video is incredibly fascinating! I'm really intrigued by the concept of building a multilevel soil sensor after watching this. However, I'm not quite sure where to begin or how to go about it. It would be amazing to see a video from you detailing the process and offering guidance on how to create one. Your content is always so informative and inspiring!
You can do it if you want!
Bedankt
I appreciate your support!
Grutti Andreas,
If you look at the schematic at 8:28 - the exposed "sensor-electrodes" still have a DC on them, albeit being used capacitively.
It is a simple addition to place two large-value film-capacitors in series, one on each of the electrodes, to effectively"DC-isolate" these from any active electronics. The plasti-dip or whatever then has to do less work (esp. with pinholes etc. still being able to ruin the sensor). As the two series capacitors are much larger than the soil-dependent capacitance, they do not affect the frequency-modulation.
I agree with your proposal. However, some viewers suggested that the PCB lasts only for a short time too. So maybe a small plastic bag around the sensor would help...
Because of the exposed electronics on that board, the fact that the PCB sides are still exposed, the fact that in some boards the 1M ohm resistor is not connected on one side and the fact that some of those sensors come with an NE555 rather than a TLC555 and the fact I like to experiment with moisture sensors, I have used the PlastiDip method, the plastic bag method (as also suggested by Andreas) and I have rebuilt the circuit, but with the electronics separate from the sensor and the sensor connected via relatively large capacitors. In fact all 3 methods work pretty well. My most favorite method is still to use some carbon rods to measure the soil resistance, but I am well aware that the latter basically measures ions, whereas the capacitive method measures water.
If you are really serious about using the capacitive method, I would indeed rebuild the circuit and add 2 capacitors as described, put the electronics in a water tight enclosure, with a short shielded cable to the sensor itself. On top of that you could also switch off the power supply to the circuit and thus to the sensor.
But in real honesty....just two carbon rods do fine for home use
@@edw4699amazing explanation. Two "carbon rods" = center part of pencils? Could you share some schematics?
@@vazquezjm No, carbon rods, not 'center parts of pencils'. The latter would break when you put them in the soil.
Schematics???? really: one goes to the analog in and the other to either ground or Vcc
@@vazquezjmThat would be graphite.
Excelente trabajo. Un saludos desde España. Excellent work. Greetings from Spain.
Thank you! Greetings to Spain.
Thanks for pointing out the “if you let it sit for six months or so” issue - helped solve my mystery and fixed my shower. 👍
Always here to help ;-)
Thanks!
You are welcome. And thank you for your support!
Excellent!
I had to correct my shopping cart after watching this video.
Capacitance types as mentioned in Flaura's video and yours will be the choice of sensor.
Glad you saw the videos before hitting the wrong order button ;-)
Using stainless steel will help with the longevity of the sensor. This type of sensing is good for a small container where you don't care too much about the moisture content at different depths. One could use several of these at different depths in order to construct a map that describes the soil's behavior to watering. This information can be used to optimize water usage and similar concept is used in professional sensors.
An interesting idea!
Stainless steel rods/probes have been used for resistivity and Induced Polarization surveys for a long time. As long as you alternate the polarity you get a good reading and can measure the change in moisture over a large area. Like the video, lots of trash on the market.
I'm wondering if there are also moisture sensors that work on the principle that periodically a small resistor gets heated by electric current and then a circuit measures how long it takes for the resistor to cool down again, which should be dependent on ground moisture levels.
This is how the old gardena probes work, and work well. I don't know about the new smart probe, which is flat instead of round. The smart probe and gateway cost 300€, so expensive to test.
@@Mark-Harding That's a rather big pricetag 🥴
Nice re-up, Herr Spiess. :-) To further enhance this widely used humidity sensor, please have a look for the page titled "Hacking a Capacitive Soil Moisture Sensor (v1.2) for Frequency Output" from the Cave Pearl Project. There you will find 3.3V capability, how to use higher frequencies and some more tips and hacks around this sensor. In short: instead of the sensor with the NE555 chip you should favor the version with the CMOS TLC555 chip.
A very comprehensive write-up. Thank you for the link. I will include it in the video description
I have tried all the sensors shown as well. For the capacitive one, it consistently gave "wet" readings even when the soil had completely dried out. It seems like the moisture sticks on the sensor surface, it couldn't dry off quick enough as compared to the moisture in the soil. I had to remove the sensor to wipe it down for it to show the correct readings again, which makes it completely pointless.
Interesting. So the sensors worked, but the contact with the soil somehow was not good. This is the first time I have heard this.
That is very true. In short, it is not install and forget. Choosing where to install it is also important.
Maybe try to add a layer of sponge or ink absorbing paper on it will do the trick? It's worth trying if you have the time..
@@AndreasSpiess There were also a few occasions when I pulled out the sensor, I could not spot dampness with the naked eye, but after a quick wipe, the readings went back to normal. Could mineral deposits also affect readings, I am not sure. I can say when testing outside soil, the sensor always works perfectly when trying with water. I have since given up trying to automate watering.
@@awaagrikh8331 You are right, I think different types of soil/soil mix might also have an impact.
Could you propose a method to calibrate the output characteristics for the particular soil type, installation depth, and soil compaction? In other words can we compensate for these effects in order to improve repeatability? Maybe comment on use as tank fluid level sensing. Thank you.
1. I am not sure if they can be calibrated for precision. At least, I did not find any information about it. It is more to show you when you have to water your plants
2. I made a video about measuring water tank levels. Capacitive sensors are used there, too. They are longer.
@@AndreasSpiess please share the link
Very insightful, A few years back someone shared this corrosion problem, i think the guy used graphite rods to solve it.
Graphite rods are probably also a good idea.
Your videos are always wonderful, instructional and your Swiss accent is adorable too
Thank you for your kind words!
There is the "tinovi I2C Non-Contact Capacitive Soil Moisture, Temperature sensor" which works very well (full waterproof). I wrote the ESPhome component (avalable in the last ESPhome dev)
This sensor looks good! And I like people writing ESPhome components. Thank you!
With your explaination it is very clear the one to use. Thank you for good work.
You are welcome!
Very nice view over the lake. Switzerland is beautiful. I have had the same problem with "open" sensors. I took bicycle spokes and solder it on the sensor. It works OK.
I wonder how you were able to solder to steel spokes. I did not know it works.
A very timely video.
I'm working on a tool, a portable moisture sensor.
Subscribed.
Thanks!!
Welcome aboard the channel!
Loving the SHT-35 for humidity, so far at least. They seem to be quite a bit more robust than the old DHT sensors. I run 2 of them 24/7 to add humidity to my plants via a relay-controlled humidifier.
I am not sure if humidity sensors work to measure moisture. Do you use them for moisture?
I don't, but you did mention humidity sensors in the video. :) So slightly on topic.
I think that is just good old electrolysis happening there, the direct current will break apart the water into Oxygen and Hydrogen and the residue probably has to do with the impurities or minerals contained in the water that can also react.
I think, you are right. Other commenters also pointed in this direction.
Well use missed the two most important principles for measuring soil moisture. 1. Time Domain Reflectometry (TDR).
2. Frequency Domain Reflectometry (FDR) / Capacitance (Frequency)
Both are state of the art when measuring soil moisture within scientific context. When I saw your video I hoped that you have managed to build one. 😅
I know that these produce better results. But unfortunately, they are much more expensive and rarely used for the hobby. Even to build one is not easy. Or did you find a project?
That is correct, they are expensive as hell. I hoped that your expertise is sufficient to build one yourself since the principal is well documented. 😅 I started reverse engineering one but never finished. BUT I will do that in the future, at least I will try.
I've never used them. I was expecting you to use Proximity sensor with it's sensitivity set extremely low.
I have no idea if it would work but I'm my head it makes sense.
Nice video.
Just try it and you know if it works ;-)
Good morning Herr Spiess, I hope your summer holiday is going well, we're missing your fresh content every Sunday morning :)
WRT capacitive soil sensors - I coated their edges in 2 layers of nail polish but I was surprised to still find their _surface_ delaminated and corroded after little more than a year in use. Not going to buy again, they're kind of a lost hope.
Yours can survive up to 1 years. Mine is less than 3 months even if I coated it.
Good to know. Thank you for the feedback. So we need to find another cover like plastic.
@@awaagrikh8331 maybe mine also delaminated and corroded earlier, I noticed the readings started drifting after just a few months but I took them out of the pots about a year later since by that time I wasn't paying any attention to them anymore.
Great video. Thankyou for the upload!!!!
Glad you enjoyed it!
NICE HAVE NOT SEEN THIS BEFORE THANKYOU. Have many old stuff from this time and have start growing my own food salads stock for added sustainability in a ever changing eco system, thinks I can grow indoor in window light controlled systems but as you know controlling these systems are harder than people think you for replay sir.
Growing your own stuff can be very rewarding. Indoors it is probably a bit easier because of the lack of several pests
In production horticulture we use a neutron probe, but they are expensive. I have found the capacitive probes have a narrow range and are easily disturbed, somewhat difficult to calibrate IMHO, but definitely a cheaper option!
I agree with your judgment. They are ok to provide a signal to water the plants than to do studies. As you write, their values depend on many factors.
I remember getting the "chirp" sensors that ran on 3V coin cells back in - well, long ago. What I missed on these capacitive sensors was square wave output - I would assume that would be easier to transmit over longer distances. I've seen I^2C and RS485 versions, but they are pretty expensive.
Some people add an ESP8266 directly to the sensor and use the air for transmission ;-) Transferring the square wave might also be problematic. It could catch some noise over the distance.
@@AndreasSpiess I was thinking about that, but them mildew killed most of the plants so I didn't follow that up back then and forgot about it during the pandemic. My students brought that up as an add-on to the school garden, and now I'm working on it again. As a physicist turned teacher, I am not on good terms with analogue signals over long distances. Measuring pulse durations and calculating a median sound better to me (I might be wrong here, though - maybe I'm just bad at analogue wiring)
Or use a differential analog signal. This increases the complexity a bit on both the sending and receiving side, but because the signal is differential interference will not be a problem.
When the cables become very long then the transmission line effects can become non-negligible.
@@AndreasSpiess "use the air for transmission" WiFi eats battery juice quite fast. Be prepared to wire your entire garden with VCC or replace the coin batteries multiple times a day.
I've used a similar circuit on the Farm-Data-Relay-System attached to a Watermark Soil Moisture Sensor. I've used the sensor for years and it works well. I also make Gypson Moisture Sensors that I have described on Printable. They provide similar readings but they tend to dissolve over a period of time. Watermark sensors are very expensive but are worth the money for monitoring my foundation in Texas.
Glad to read that you use the FDR system (I once made a video about it)
It's super easy to build this kind of sensor yourself anyway. Just get copper rod, solder wires to two pieces, add the 500kOhm input resistor, to one and a ground wire to the other. Add a second wire to the one with the 500kOhm directly to an analog in of your microcontroller.
That's it. You can have either the v+ or ground coming from a toggleable port so you can save a bit of life on your sensor.
Good point!
Hi Andreas, I have tested the capacitive moisture sensors quite extensively in aquaponics. I use a ADS1115. I found considerable variation due to tempurature fluctuation.
Thank you for sharing your experience. Other commenters confirmed your finding. I do not think these sensors are suitable for precise measurements. But they are okay to create signals for watering.
This was incredibly helpful. Thank you for this!
You are welcome!
Any electron flow in a damp environment will cause material to migrate either from your plate or to your plate.
None contact is the only way to go, or treat the sensor like an LCD display where there it has to be treated the same, an AC signal with no DC component, or you destroy the segment.
I agree!
I bought the Xiaomi Miflora Sensor wich has also a lot of knockoffs wich are cheaper, Monitoring it for 7 Montths now and it works fine and reliable + has integration features Battery still shows 100% You pay a little extra for the enclosure ( water resistant ) and the ease of use ( not encrypted BLE ) and a nice byfeature is the EC measuring wich gives a rough ( not really useful ) estimate when its time to give a little more fertilizer or if there is too much salt build up
Integrated in HomeAssistant with Automated Watering learning based of the dry / wet cycle and times it takes the plant to take up the water added
I couldnt have or use more features and its working out of the box, no need to ask for more after fiddling around with diy solutions for years im happy that this project is now running and working
I agree these Miflora sensors are a good choice. Thank you for sharing your experience!
Thx for this topic...
It's long time we develope this.. and have so many alternative..
But sometimes, its also depends on media that need to be measured..
I agree. These sensors are made for typical soil only.
I use a very good nail varnish to cover all the moisture sensor v1.2 including the chip and the pins and wires and works fine since 2-3 years now. But there is a missing connection between ground pin and the a ground trace. Simply solder a small diameter wire on both side of the pins header. Cheap China stuff but works.
Thanks for sharing your experience. I left a video where the quality problems are explained.
Cool DIY sensor ;-). But seriously, I think a capacitive DIY sensor can be a good option for makers and you can easily create a square wave signal with a microprocessor board.
I agree. But the simplicity of a 555 timer is hard to beat ;-)
I have a commercially made sensor (Ecowitt WH51) that appears to have a continuous loop that goes into the soil. Perhaps under the coating it's not a continuous conductor though. It eems very reliable but now I want to see what is inside.
Inside are two not connected traces that form a capacitor.
@@AndreasSpiess That's my suspicion after watching your video. It was very helpful to understanding how it works.
you can use the cheaper ones without their electronics, and alternate the current thru them. the main problem with all these sensors is their resistance depends on the soil and the minerals present. takes allot of effort to calibrate them on specific soil now try it with changing soilminerals (rain water vs tap)
I agree with your findings.
What about sensing by weight instead of moisture level? A flower pot will have a certain consistent weight. As moisture evaporates, the weight will be reduced. At a certain point, you turn on the water. You might have to re-calibrate from time to time as the plant grows.
That is a good idea. I once made a video on how to use load cells.
I testesd one of the capacitive V1.2 wrapped in film for resin UV resin detection and was surprised it worked.
Thank you for sharing your experience!
Hallo Andreas, der Beitrag kommt genau zur richtigen Zeit :-) bin gespannt, ich habe einen mit zwei V2A Stahlröhrchen "gebaut" :-) Den Kapazitiven Sensor hatte ich auch schon mehrfach verbaut, leider hat sich auch hier nach einiger Zeit die Farbe / Isolierung gelöst und die Kupferbahn aufgelöst, nicht gerade gesund für die Pflanzen
Ja, die scheinen auch nicht ewig zu halten. Vielleicht würde eine Plastiktasche Abhilfe schaffen. Die Sensoren müssen ja die Erde nicht berühren.
@@AndreasSpiess im Baumarkt gibt es streichbare Faserverstärkte Dichtmasse, die hab ich diese Woche zum Dach abdichten verwendet ich glaube ich streiche mal so ein Teil damit ein und teste es. Damit könnte man auch die Elektronik am oberen Ende abdichten. Ich berichte dir :-)
I have used that cheap sensor on my Christmas tree to warn about low water level for many years without a problem. I realised this design error and implemented very simple correction: I'm not powering it all the time. I use Arduino output pin to power sensor only for a couple of ms only when I do measuring once per each hour.
This is a good improvement and reduces corrosion!
Could you drive the original sensors with an AC voltage to avoid the electrolisys and then rectify the output signal with a diode?
That would complicate the design, but it is possible and would reduce the corrosion.
Hello Andreas, Gardena (Husqvarna) uses a different approach for its irrigation/sprinkler system. Its moisture sensor contains a metal tip that is enclosed in felt. This metal tip heats up periodically (about every 30 minutes) and, based on its temperature decay (cooling time), the degree of soil moisture is calculated. When "wet", the connection between two pins is "closed," when dry the connection stays "open". The two pins are connected to the irrigation controller. Why periodically - to allow for battery operation. Greetings, Stefan.
Thank you for the description. I never read about such a method. But obviously, it works!
The first real Arduino project I built used these resistance sensors. I was so proud of myself, until the 3rd and 4th days when the sensors were ruined.
This went fast! Usually, they live a bit longer.
You can use the first pcb very easy with an arduino and no electrolysis.
Connect the pcb between to analog inputs. For example A1 and A2.
In the first measurement make A1 output high, it will supply now 5V to the sensor, read the analog value from A2.
For the second measurement you do the opposite, make A2 digitaal out high and read the analoge value from from A1.
Repeat the above.
With this way you measure a sort of with ac and the electrodes with not desolve.
If the reading of the adc is to low, try setting the reference to 1.1V or other value.
Thank you for the info. Indeed, you can reduce the time current flows with this method
Did you check the RF-emissions of the capacitive sensor with a SA? I think there might be a lot of harmonics causing QRM in the HF bands. Maybe not relevant for the average houshold, but in fact, for example I've turned off my weather-station because of spurious emissions in the ISM-band, disturbing my satellite groundstation. So, maybe as a HAM-operator I'm a little bit over-sensitive to this . Anyway - nice video, thanks!
No, I did not check it for QRM. I have enough other sources, and I am building a remote station because of that ;-)
Nice!, good research👍🏻👍🏻
Thank you!
Andreas - good to see this topic again. I was a disappointed to learn that even capacitive sensors fail over time if they are just coated with resin, as the resin still allows water intrusion via osmosis. I wonder if that includes all forms of epoxy.
Perhaps the probes and the base of the instrument body could be coated with "Plasti-Dip"? The plastic bag is a fine idea, I guess, but really the sensors should be fully sealed in plastic until the whole instrument is up out of the soil and well out of the way of splashes, etc. So rather than putting plastic "leggings" on the sensor probes, I'm talking about plastic "pants" that reach to the top of the sensing instrument.
I am a bit sloppy when watering plants, yes. :^)
Thanks for sharing your experience. I do not have longtime experience with the different sorts of epoxy. Maybe others experimented with it already.
The thicker the protective layer, the less sensitive the sensor is. Which means: use a very fluid resin for coating, so that it will cure in a thin layer.
Hallo aus Österreich, Andreas! Ich mag deine Videos. And "the Swiss accent" :)
Wegen des Kupfers...
Wenn du, so wie du es gemacht hast, die Elektroden ins Leitungs-/Brunnenwasser steckst (nimmt man ja meist zum Gießen) und mit ein wenig Spannung und einigem Strom beaufschlagst, erzeugts du einfach Knallgas. Sauerstoff und Wasserstoff an je einer Elektrode, das entstehende Wasserstoffgas kann als Blasen an der Oberfläche der Kathode freigesetzt werden.
Komplizierter ist es mit Folgendem:
Im Leitungswasser sind üblicherweise verschiedene gelöste Salze enthalten, darunter Natrium- bzw. Kalziumionen und Hydrogencarbonationen (HCO₃⁻), die möglicherweise reagieren könnten, um Kupfersalze wie Natriumkuprat (Na₂CuO₂) oder Kupfercarbonat (CuCO₃) zu bilden. Verschiedenen Faktoren wie Wasserzusammensetzung, pH-Wert, Temperatur und Stromstärke beeinflussen die Reaktion.
Wenn man Pflanzen mit einer solchen Flüssigkeit bewässert, bzw das im feuchten Erdreich auftritt, könnte dies potenziell schädlich sein, da Kupfer in höheren Konzentrationen toxisch für Pflanzen sein kann. Kupfer ist für Pflanzen notwendig, aber nur in geringen Mengen. Bei zu hoher Konzentration kann es jedoch die Aufnahme von anderen wichtigen Nährstoffen behindern und oxidativen Stress verursachen.
Die Kalkablagerungen könnten auch den Boden pH-Wert erhöhen, was wiederum die Nährstoffverfügbarkeit für Pflanzen beeinflussen könnte. Ein erhöhter pH-Wert kann dazu führen, dass bestimmte Nährstoffe weniger leicht aufgenommen werden können. Es ist wichtig zu beachten, dass die Auswirkungen auf Pflanzen von verschiedenen Faktoren abhängen, einschließlich der Konzentrationen der beteiligten Ionen, der Dauer der Bewässerung und der Empfindlichkeit der Pflanzenart.
Danke für deine ausführlichen Erklärungen. Chemie war nie meine Stärke ;-)
One guy said put a 1M ohm resistor between signal out & ground on the 1.2v sensor, does that make sense? All mine, 10 of them, give me random values & don't respond when I'm water 😢
I do not know :-( You have to try.
Arduino's can drive and measure capacitive changes between plates directly. There even an offical library for reading such sensors.
It's also *easy* to create such plates oneself. If aluminum or titanium is used it's also possible to anodize the surface to insulate them. Far more even and resilient surface than varnish Nb the thickness of such an anodized layer adjusts the value/range of such a sensor
Thank you for your additional info!
Even the capacitive sensor isn't good, with exposed FR4 edges. I looked into this once and Steven's Hydraprobe looks good. It gives moisture, temp, salinity over RS-485 using 50MHz RF. The cost is $695 though (!).
I agree. These TDR sensors are much better, I agree. But for controlling the watering of 10 flower pots probably a bit expensive ;-)
@@AndreasSpiess FWIW, HydraProbe isn't TDR, it uses radar to measure dielectric impedance. It measures the energy storage and energy loss of the soil to determine the moisture content. This reduces the influence of salinity and temperature compared to other solutions, allowing it to work without calibration. It's really neat and I have bought one, but I won't get to test it until early 2024.
The cost is silly, but nothing is too good for my lawn. :p Mostly I just want to play with the cool tech. My DIY irrigation controller calculates ETo to determine watering durations. I'll use the soil moisture data to double check that, eg don't water if the soil has enough moisture even if ETo thinks it needs watering.
The two-legged sensor with a gold plating most better than of cheaper lead. I've tested ENIG version about 2 month, if not scratch them, no corrosion will be even with fertilizers. And power on must be very short as possible for measuring.
Thank you for the info!
Great video Andreas! I've been having problems with the capacitive sensor, making my house main differential circuit breaker go off as soon as the sensor contacts the soil. I'm assuming there's some kind of leakage? Will try the plastic bag to isolate the sensor and see if it helps
The 5V power supply should insulate your devices from the net. So pay attention!
@@AndreasSpiesshmm, that could be the issue... will try with a different PSU. Thanks for replying!
Soil has much less moisture, so are these sensors sensitive enough to measure it?
Yes, they are sensitive enough to alarm you when you have to water plants (probably not cacti ;-)
Hi Andreas, did you ever try to put a Bosch BME sensor into a cavity underground like a small platic pot with the opening facing down? This would give you the relative humidity of the air in equilibrium with the soil.
I learned that measuring humidity and moisture is a different thing. So I never tried it.
I remember reading about this problem in the 1980's (probably in Electronics Australia Magazine). If I recall correctly, there was an IC manufacturer who produced a chip specifically for this application. It applied a small AC signal to the probe for measurement to prevent electrolysis.
I see lots of moisture probes advertised now and just assumed they would use the AC method. I forget how much junk is sold out there.
I did not know that this issue is so old.
@@AndreasSpiess I guess I'm showing my age here, haha.
Keep the great videos coming, they're always interesting and I learn a lot from you. Cheers.
@@desparky It cannot be the age. I am 66 ;-)
I wonder if the resistance sensors would be better with graphite / carbon probes? Carbon is pretty inert in water electrolysis, unlike copper. Whilst they wouldn't last forever, they would significantly outlast copper traces. Also carbon is not toxic to plants whereas high concentrations of copper can be.
Others used graphite rods. I do not know how their handling is. The ones I know are very brittle.
Thank you! This reminds me of the capacitive tank level sensors. Same principle.
Yes, just longer ;-)
@@AndreasSpiess the most effective frequency was particularly interesting. I suppose this would have to be experimented with, if one was to make a level sensor where the tank material (dielectric) was quite thick (8mm).
I love your videos. Always nicely informative and without annoying background (sometimes foreground!) music 👍😁
Hi andreas can the capacitive sensor be used as rain sensor?
I do not know how you would do it. The just do what I described.
I have tested the capacitive soil moisture sensors extensively and they are not reliable whatsoever to measure anything accurately in soil, the only working solution is to make a resistive sensor that receive power only for a very short time with prongs in stainless steel or silver solder. That was my solution to manage my sprinkler system in Florida. Anyway, there is no need to measure the moisture in real time, 2 or 3 times a day is more than enough, humidity does not leave spontaneously the soil.
Tin the copper! Or use silver solder...
Thank you for sharing your experience! Shortening the measuring time for sure reduces corrosion.
Another option to extend the lifespan of sensors with exposed copper conductors is to make measurements in a descrete manner. I.e. not constantly but once every 30 minutes for example.
I agree.
And keep the time that it is switched on as short as possible. Usually ~10µs should be enough. Then even more frequent measurements should be possible. What could also help is to lower the voltage that it has available. According to wikipedia, water electrolysis requires a minimum of 1.23V. Measuring with max 0.5V should therefore eliminate corrosion. That may be worth testing.
Another technique can be to only power the sensor while the moisture content is being measured. No reason to read the sensor constantly, once an hour or so is more than enough. Cut the power to the sensor to only a few seconds per day will extend the life of the sensor.
A good idea!
there is a prefect soluiton from 70's: gypsum block. the response time is slow but extremly accurate.
Thank you for the info. I found that these sensors are pretty expensive.
@@AndreasSpiess easy to make u just need two graphite electrode and gypsum. and a temperature sensor if u want to be really pro.
If price is not an issue, i have uses smt100 and smt50 from truebner on a research project. They currently have an uptime of 4 years (but they cost 100 and 50euros respectedly)
I recently saw them on on a fair. They seem to be solid.
Durex is a brand to protect some sensors with a rubber layer
Thank you for the information!
The capacitive approach is a simplified form of impedance spectroscopy. The real fun with that is when you sweep the frequency and take a spectrum instead.
Indeed so. I worked with non-linear dielectric spectroscopy for years measuring yeast viability.
What would I discover in the spectrum? Is there any information available?
A super cheap workaround is running DC for 10ms every minute or so. You still get your readings and the sensor lasts much longer.
You are right. This helps to avoid electrolysis.
This was interesting, but might be a little out of context.
This would probably be fine as a physical liquid level measure, but that's not how dirt and moisture work in most agricultural situations.
If your plants roots are sopping wet, your plant is likely drowning.
Usually there is much less liquid in soil, and that may make the readings much different.
Could you repeat the experiment with the last sensor, using different types of soil, and adding water like a grower would?
The best test scenario would be in a large planting container exposed to the outside air conditions and to the sun, but protected from direct rain. That way, there is a large amount of soil, there is air, heat, sunlight, wind, to dry the soil, and you can add water in small amounts, as if from a sprinkler or irrigation system. The moisture in the soil will need to equalize across the bulk of material, rather than just detect full saturation.
Water would arrive on top, and then get wicked through the soil. Having a short or small container might give false readings from water pooling in the bottom, saturating the lower part of the sensor. The sensor needs to be affected by the correct amounts of moisture in the soil, not just "dry_ and" saturated" conditions.
That would simulate regular growing medium better, and give a better idea as to the signals seen and the response time to detect actual soil conditions.
_____
As an added thought, this looks like it would be a good wired system for a central garden plot.
Consider a yard, though, or multiple garden plots, distributed areas, etc.
Having wires running 10s or 100s of feet, in multiple directions, on opposite sides of structures, etc., and with foot traffic or vehicle traffic in the area.
This would be hard to set up, calibrate, and maintain.
A larger scale solution would call for wireless communications, sensor differentiation, logging, and probably power management.
This is a job for SUPERMA-- Sorry, I got carried away. This is a job for: I don't know... It's got to be be cost, able to leap large distances, be identified, easy on the batteries, etc. LoRa? I don't know the costs or component size. 422mhz? ATtiny85? RaspiNano? BLE?
I can think of a couple approaches.
I can envision a mesh network for ease of adding sensors and extending the coverage of the network, but that seems complicated.
I can see having a base station and just relying on low power rf transmitters, logging and sleep routines, and periodic polling fob the base station.
I believe that BLE can operate as a mesh network.
That's enough to think about now. I think I want more coffee.
When I made this video, I was looking for studies as you describe. They should be made by agriculture engineers, not electronics engineers. But unfortunately, I did not find any. That may be different now.
These sensors could be more precise. I think they can be used to generate watering signals in typical home environments.
If you want to avoid cables, maybe you google "LoRa moisture sensors". Their range can be a few km. You also find many videos about the LoRa technology on this channel.
@@AndreasSpiess Thank you. I am thinking more along the lines of a home owner. We aren't on a large plot, but with a growing space at the front of the yard, two near the house in front, four spread through the back yard, the thought of monitoring all of that using wires is a problem.
I guess in the end, I would pick maybe two or three, but even then, that's a lot of wiring spread across hundreds of feet in different directions.
As a manual (handheld) tool, that might be a more practical solution for me, but then we're back to doing things manually.
Still, it's a neat video.
Wasn't some of the capacitive sensors missing a critical component?
This is a re-play of an old video. In the meantime, it seems there are different variants available, some that do not work. I added the info in a pinned comment.
I tried several capacitive sensors-cheap one, DF Robot one and two other Chinese brands. None can survive more than 3 to 6 months even if I coated it. Soil is salty due to fertilizer usage so it greatly reduce the life span of the sensors.
That is what others reported, too. Maybe a small plastic bag instead of a coating would help? The sensors do not need to be in contact with the soil.
sonoff ms01 is looking pretty good, it sends data trough I2C, it is on my to do list to test one in esphome and if satisfied switch from all other ones .... but it is not on top of my list ....
Xiaomi offers a similar one, too. They are a good solution if you do not want to tinker.
@@AndreasSpiess looking for some "reliable" to integrate with esphome for home assistant. Will look at it, if you know about more please let me know, because i am in fight with soil moisture sensors from spring.
I use hot-galvanized nails as probes. These are the gray, rough looking ones. Few years now, no corrosion. Avoid electro-galvanized (shiny) - they wont last.
Good tip. Thanks!
Would using two stainless steel probes solve the corrosion problem?
There are several proposals in the comments. One mentioned special nails that avoid the corrosion problem.
So you made an unintentional electrolyzing cell. You are deplating the anode(+) side and plating the cathode(-) side. As well as liberating some hydrogen on one electrode and oxygen on the other. One way to make them last longer is to swap the polarity every other time you read it and only activate it when you want to read it. Nickel plating the electrodes will also help some. 😊
And for the capacitive version you might use some spray on conformal coating.
Thanks for your tips! Conformal coating for sure helps.
Or with suitable choice of material... Any dc current flow will anodize a plate and stop the conduction.
Hello Andreas I use the last sensor (capacitive on), there is a but you forgot. The value measured is not stable since it is impacted by the temperature. It creates quite big differences. I ve not solved this yet.
I do not think these sensors are accurate, too. They are good to remind you to water your plants, not for agricultural studies...
i would love to see you include a "Gypsum" soil moisture sensor in your comparison!
I agree it would be interesting to compare the readings of the (quite expensive) gypsum sensors with the ones of these cheap sensors. This is more for an agricultural channel, I think.
The first sensors can benefit from probes made of stainless steel. But all have a major flaw in that they use a lot of power and they are not the best thing for IOT standalone sensors. Also resin coatings in pcb board suffer in the long run from water permeating through the pores because of osmosis. This causes blisters filled with acid under the coating. The latter I suppose one could accept it as a wear part in the system.
Maybe the capacitive sensors can be protected by plastic foil since they do not need contact with the soil.
Who would've thought that the verable 555 timer would get a new lease on life being used in a humidity sensor? That chip just keeps on giving.
Indeed, this is a very versatile chip. (And it was invented by a Swiss ;-)
Good explanation as always.
Thanks for watching!
Ah the classic ox redox reaction. You might even be creating metal complexes in the water (gets very complicated to know if its plain rust or a complex of Fe)
Thanks for the explanation!
I wonder if an air humidity sensor placed under an upturned cup on the soil surface would be of any use. It could be an interesting experiment.
Maybe you try it? The sealing has to be very good, I assume. Otherwise, you probably measure the humidity of the room.
From what i gather the coated sensor used way more power so they last months instead of over a year battery wise.
IOm strill wanting to have a google come compatible sensor over wifi. i dont want to have a tuya or zigby hub google home keeps things simple and working well without the need of a hub making replacing things easy and pain free.
I have ordered few for my greenhouse after watching the video
:-)
Danke die für den super Beitrag! 73 von HB9TIA!
Gern geschehen! 73
Small problem: Clean water id more or less an insulator. What you can measure is the concentration of different salts wich gives a very unprecise picture of the soil moisture.
For the resistive sensors, you are right. The capacitive ones use water as a dielectric material.
@@AndreasSpiess Yes, but what does the salt content in the water means for the dielectric properties? I have experimented with a small cage made of net buried in the soil. Inside a DHT 22 to measure the air humidity and tried to relate it to the soil mosture.
@@andersgardo6777 And did it work?
@@AndreasSpiess I got some different values but i don't know how relevant they are. Lack of time to go deeper.
those two leg sensors basically suck, but you can prolong their life by measuring soil moisture only when needed, e.g. couple of times per day
I agree.
Interesting, just what I found using in my greenhouse. Thanks.
I hope your plants do well if they get sufficient water with your automation ;-)
true , the more water between the legs , the smaller the output 02:45
I agree.