Nice catch on the resistor and cap. Sadly, others who have received these sensors in the same condition, may not know how to find out what the value of that cap is. Also, they may not realize that the resistor isn't even connected at one end. The fix is so simple once one knows. In like manner, I use a lot of the DHT22 temperature/humidity sensors. They mostly work well, but sometimes, I've had some that seem to "bounce" a lot. Until recently (in Jan2023) I had never found out why. No seller has this documented, but the fix to stabilize these sensors is.... yup, you guessed it, a resistor! In this case, it's a 10K connected from the 3.3V pin to the Data pin! The sensors work best when connected to 3.3V instead of 5V, as I've also found out. When I first discovered this, it didn't make a lot of sense to me, because to me, this was pulling UP the data line. The data line is "digital" not analogue. But, it works! Since adding this, I find that the sensors don't bounce anymore. I do wish that we had some GOOD manufacturers of all these sensors and devices, rather than the rip-off trash we receive. Oh well...... Now I need to find a good schematic of a circuit to use to make my own capacitive water level sensor. Fun times.
Hi, The value of the cap is 1uF. However, it is connected on the board. It's only the resistor is not connected on the defective ones. There is a schematic on the link mentioned above showing all the components. I tried the DHT22 aa couple of times, Both times they ran for a while and seemed to die with no obvious explanation. Your problem with the pullup resistor is not unusual, even with digital outputs. Low voltage devices sometimes have open drain outputs. The resistor helps to give a certain amout of bias between off and on. The Capacitive moisture sensor can be used for water level, depends how deep the water is. Circuit would use Arduino or Raspberry Pi Pico. For large volumes, I use ultrasonic distance sensors. I can easily measur contants of a 10,000 litre irrigation tank. Best Regards, Stewart.
Very correctly identified, the output needs to be explicitly loaded because of that stupid capacitor. Having a resistor between the Analog out and the the ground helps discharge this. This problem exists in 2.0 as well. I had multiple sensors connected and they seemed to interfere with each other. When one was pulled down, all others would go down automatically! Spent hours figuring out the problem Glad you were able to identify it.
Huge thanks, new to all this and have been struggling for week thinking it was something I was doing wrong. Added the resistor and happy days, working and it's responding as fast as all the other project videos online.
Excellent and very useful information. I had the same problem and got frustrated getting it fixed. I build my own circuit and merged it with the probe (I removed all the SMD components). All the version has the same issues.
I sand down the edges of the board so there are no sharp edges. I then put heat shrink wrap on the cable where it connects to the board. Then I dip the whole thing in varnish. Leave to dry. Repeat 2 or 3 times to get a good thick coat on the board.
I am so glad I came across your video. And that by cheer coincidence as I thought something went wrong with my ADC (MCP3008) setup. Since September I have been working on a little auto gardening hobby project for my Raspberry Pi. The capacitive soil moisture level had always been a struggle for me as I could not explain those readings (usually running towards 0 after a few seconds). I even went so far to try another ADC (ADS1115). Only after testing with an old YL-26 that I had lying around and getting more reliable results with that one, I started questioning the capacitive soil moisture sensor 1.2. It's a shame so many people have wasted so much time on trouble shooting. Thanks again for sharing!
I have had a lot of grief with these sensors and have bought several from different suppliers, and there's a number of different faults. Here's what I have found: 1) v2.0 is no different to v1.2. The circuit is the same, and they contain the same faults. Don't be fooled by this feeble marketing trick. 2) The resistor R1 is not connected to ground on some sensors, but is on others. Where it's not connected, response is slow as shown in the video. The 1M Ohm resistor solves the problem. 3) Some of them are fitted with the wrong version of the 555 Timer chip (IC2). The supply voltage out of the regulator is either 3.0v or 3.3v, but some sensors are fitted with 555 that needs at least 5v. 4) Some sensors do not have the 3v regulator fitted. Instead, there is a diode, so if you're running off a battery, you'll get incorrect readings as the battery voltage declines. 5) The reading when the sensor is perfectly dry should be close to the supply voltage (in mV) and when immersed in water, should be close to zero. On most sensors that give me relatively sensible results, I get around 2500mV dry and 950mV in water. However, I did have some giving 3700mV. To use them effectively, you'd have to calibrate them and adjust the software accordingly. Unfortunately, most vendors just buy whatever is cheapest, so you don't always get the ones they show in their photographs. It's luck of the draw whether you get a working one or not. I had some arrive that were labelled "Capacltlve Soll Molsture Sensor v2.0". They couldn't even reproduce the label correctly, and they had R1 disconnected so reacted very slowly. Another RUclipsr I follow tested over 30 different sensors from different suppliers and found 82% to be faulty. Buyer beware!
Thank you for this post. Had the same issue with my replacement sensor. The previous one was not working at all. What's this about doing in varnish? Where can I find more info on that?
Hi Kristoffer. Thanks for the comment. The idea behind the varnish is to protect the board and components from moisture. The edge of the PCB is porous and rough. The theory is that water may soak through the edges and into the PCB. Not sure if this is true of fibreglass, but why risk it? Sand the edges of the board so they are rounded and have no sharp corners. Dip the board into varnish and allow the excess to drip off. Allow to dry, then repeat the process until you are confident the board and components are sealed.
Thanks for making this video, very interesting. I have bought some of the V2.0, how can I tell if I have the TLC555 or the NE555? There are no markings to say which one I have. Am I right in saying that if it powers up at 3.3v then I have the TLC555?
Hi. Thanks for commenting. You may find there are markings on the chip that are not very clear. Use a torch and a magnifier to look closely. Also, if there is a regulator it is probably a 3.3 volt regulator, which means it is almost certainly a TLC555.
recently bought 5 of those on amazon (DE) all nicely packaged an in good optical condition. One seemed to work for a few minutes all the others had the same problem you discribed. ill try your fix tomorrow and report 4 of 5 are very slow - when connecting the resistor between signal and ground its going to show 0 One of them works with your fix but range goes down to 450-200 from 900-100
That new range sounds more realistic to me. You just need to calibrate your software to work with those figures. With the ones that show zero, check the comment from Zwick Nagl above. He found that some of these devices are fitted with the wrong type of 555 chip. Best Regards.
@@So-Now found another seller on Ebay offering same thing but with one changed component. He wasnt aware of that but after a quick conversation he told me that his ones are 98% working fine. Bought 5 - all working with a Range ~300-2500 and super fast - awesome. Unfortunately i already send the others back to test if changing the u2 Component to to a 0 ohm Resistor/ wire (if thats what im seeing here) The PCB is also half Thickness and printed Text is double in width of the old ones. https//imgur.com/a/WXfB8iz
@@madnessfire9081 Hi again. Not sure what you are saying here. The range you say is ~300-2500. Was that a typo? Anyhow, it seems there may be several problems affecting 'some' of these devices. I am currently adding a page to my web site (will post here when it's done), just for this sensor , because its easier to update my site than it is to update a RUclips video. Plus, I am learning more about these sensors every day as others are updating me on their findings. Seems this device is in demand at the moment and any new information will be very helpful. Thanks again for commenting and keep me posted if you find anything new. Best Regards.
Hi Emmanuel, Thanks for the comment. In my case, I sanded the edges, then I dipped it in undercoat paint, let it dry, then into varnish a couple of times..
A little while ago I purchased six from Banggood and all seem to be working as expected, grabbed one from Amazon due to the quicker delivery and that one was faulty. Hopefully this means the problems been solved and is just an issue with older batches.
Thanks everybody who has commented or subscribed so far. Sorry for not replying sooner. I had a few comments and for some reason, they all vanished for a couple of weeks. They suddenly re-appeared in the last few days, along with some new ones. Thanks again.
Hey, Stewart nice video you have there! do you think the capacitive soil moisture will work if it is submerged to oil? or just the moisture from the oil?
Good question. I could only guess as I have not tried it. It would also depend on what type of oil. Mineral oil, vegetable oil. I would expect they would have slightly different capacitance.
Good question. The resistor is just one component this circuit needs in order to work. The 555 pulses on and off, charging up C4. Without the resistor, C4 would charge up and reach the supply voltage and stay there. The 1M resistor is required to allow this capacitor to discharge, making the output variable. The ratio of the charge/discharge time is governed by the capacitance of the traces on the sensor. Hope this helps.
Stewart Cash the other issue I have with this unit is it is so voltage dependant for good and trustworthy readings. If I upload the same sketch to a nano then the voltage regulation is different and leads to me having to recalibrate from scratch. If using from cells then heaven forbid a drop in voltage. This is to be expected with this type of circuit - just frustrates me no end - great videos you do though, you don’t hold back - I like that.
Hi Dave, Yes. Depending on accurate supply voltage, processor, ADC, even component tolerances on each individual sensor, the actual reading will vary. I have found that each sensor needs to be calibrated totally dry, then submerged in water, to get top and bottom limits. This may also need to be done more than once. For accurate results, your software/firmware must include some form of calibration which can be repeated. Hope this helps.
Stewart Cash Hi Stewart, Yer, I have found this frustrating to say the least. I wrote the code for a simple water feeder and allowed a 8 second filling stage when it required and then waited 30 seconds for the water to penetrate and repeat if required. All worked like a dream, till I moved the code to an ATMEGA chip on a home made board. Then it just kept watering as the results were way out for the voltage differences - I’m now doing the same on the esp32 and have much much better stability and results. Please keep up your very informative work - got to love electronics, got to love programming - put the two together and you have many headaches but mountains of joy. Thanks Dave
Thank you very much for this video! I have a few of these, all with the same issue. How did you get to the bottom of finding the missing connection on R1, beyond measurements. Can you visibly see it on your pcb? On mine, they physically don't look like they are missing the connection but the behaviour suggests that the connection is in fact missing. Thanks again!
Hi Jacob, Thanks for the comment and subscribing. I found it just by testing. You cannot see the tracks on the board because the solder mask is black. However it seems that the large ground track, either side of the connector, has been broken by adding the connector at the design stage. If you scrape off a little solder mask each side of the connector and solder a jumper wire across, this also fixes the the problem without adding a resistor. I have ordered a few more and will sacrifice one to prove my theory. I will make a video if I am correct. Best Regards.
Hi Stewart, Thank you for the video. My v1.2 sensor has 0 ohm resistor instead of voltage regulator. Moreover I am using it with nodemcu v3 with easy_esp installed. I found that when I supply sensor with esp 3.2v then ne555 has 360kHz and the output varies between 0.8v in the water and 2v in the air. But when I use 4.5v form usb then ne555 has round 413kHz if I recall it correctly and output between 1.8 and 3.2v. I've double checked and nodemcu v3 recognise 3.13 as 1024 (max value). I haven't damaged it yet but I'd rather be careful. But the most important part is that when I collect all this data in domoticz it's fluctuating +- 2%. I was trying to find some good tips in the Internet but without success. I don't know what might cause the fluctuations.
Hi Marcin. Thanks for the comment. I have been told by other viewers about the zero Ohm link in place of the regulator. This is because the manufacturers have replaced the TLC555 with a NE555. The TLC555 can run at 3.2 volts. The NE555 must have 4.5 to 5.5 volts. You would have to supply it with around 5 volts but then you could not drive it from a 3 volt processor. I have never measured the frequency of the output, but the resistor and capacitor would tend to smooth that out. The difference would be, higher frequency, slightly faster rise and fall time.
Thanks Stewart , really valuable feedback. I purchased 4 of the V1.0 sensors from DFROBOT and all the parts are present and appear to conform to their circuit diagram. I'm planning to remove the 3 pin socket and directly attach a screened cable pigtail, then spray coat the whole board with several applications of CRC urethane seal coat. This will hopefully provide sufficient protection for direct burial below the lawn. Did you observe any adverse impact on the sensor sensitivity due the additional sealing you applied ?.
Hi Mike. Thanks for the comment. On several of these devices, I have removed the connector and replaced it with 2 metres of screened cable. I then carefully sand the edges of the board, then dip the whole device in paint or varnish. I am lucky enough to have a 3D printer, so my last step is to print a small clam-shell case to cover the component part of the board. I then snap this shut over the board with epoxy resin to seal it all shut. Hope this helps.
Thanks Stewart, that’s useful information and gives me confidence that I’m going in the right direction. As I plan on using four buried sensors to monitor the lawn on each side of my house, I’ve been trying to decide whether to separately extend back the sensor voltage on separate cables to connect with a data logger, or use a bus interconnection and digitise the moisture signal at the sensor with a microcontroller. After investigating, it appears that professional grade sensors use an digital SDI-12 interface, which appears reasonably easy to implement with a cheap 8 bit microcontroller and an bus driver circuit. I’ve decided to go with digital interconnection of the sensors, so I’ll be putting the sensor electronics, microcontroller and bus driver circuit in a small enclosure and ‘potting’ the assembly with urethane sealant. Thanks for getting back to me.
@@mikegofton1 Hi again Mike. I was just re-reading through the comments and came across this one again. How far have you got with your project? I was thinking, maybe you could look at using a CAN-Bus interface. We used to use it on vehicle based systems (as it was designed for) however, we also used it on a mini underwater drone tether. Up to 1000 metres cable length, multiple devices on a single twisted pair. Could maybe use network cabling if you can get enough current down the other pairs to power your devices. A couple of driver chips and some fairly simple to understand libraries for the Arduino. Maybe worth looking at. Best Regards, Stewart.
@@So-Now I'm using an arduino pro mini to interface to a 3 wire serial bus (5V half duplex 1200 bps asynch, same as the physical level for SDI-12 ), and JSON to transport messages rather than the SDI-12 protocol messages. There is a SDI-12 protocol arduino library, however it uses software 'bit banging' - I prefer to use UART hardware. Most SDI-12 sensors are also professional grade and expensive, which seemed like overkill for my garden. I've built the hardware / written software, and will test the system this week sometime.
@@So-Now Hi. i bought a version 1.2 sensor (or rather, I bought version 2.0, but the seller deceived and sent 1.2), you can see the U2 element (on three contacts) on the video at 2:50, it is removed on my sensor, the upper and right contacts are connected by a resistor (or something similar), tell me if your advice will help to fix problems with the sensor? Thanks.
Hi. I never tried it, but results might not be very good. The output depends on the capacitor charging up and being discharged by the 1Mohm resistor. You will not damage anything by trying it. Let us know.
How can I connect it sir? Do I have to solder it to the component itself ? Can I connect it to the wires? Sorry because I don't have that much information or technical knowhow about this. Thank you for your response!
thank you for this fix there is still people out there selling these as i order some from eBay and AliExpress and both were 1.2 V same as yours but Im going to fix them now but i bought new ones with V2.0 thanks
No. But I could imagine it happening. The soil moisture sensor is a capacitive stick. Which is basically how we make antennae (antennas for the Yanks), aerials. It would therefore act as a magnet for any stray electromagnetic fields floating around. If the grow light runs at any frequency, if it has some dimming control for instance, it could very well cause some problems. One thing to try is using screened cable to connect to the sensor. Most often, interference finds it's way in through wires (again, acting as aerials). I would be interested to know what type of interference you are seeing. Is the grow light LED or flourescent? Do you have an oscilloscope? Thanks for commenting. It's good to see there is still some interest. Best Regards, Stewart.
Confirmed: missing link between 1Mr resistor and gnd. Due to faulty layout so a lot will have the same. Grrr.... Also NE555 instead of TLC555 => no startup @ 3.2V. 2nd Grrr... Thanks a lot for your video !
Thanks for this tip! I had the situation that the sensor was in the ground and would read 0 all the time. Whenever I touched the plant, it would start working. I guess I was the 1M resistor to ground :-)
@@So-Now Honestly, I haven't fixed it yet. It just seems to make sense that your method may resolve the issue for me as well. I'll let you know the output once I get hold of a 1M resistor.
does this also fix the problem that the measured voltage decreases (eg. from 1.26V to 1.18V) when the sun shines on the sensors ? I want to use them outside but that behaviour makes it a bit more difficult
Thanks for commenting. I have not heard of that particular fault, though it sounds likely. Any of the components on the board could be affected by heat. I would have to test by heating one up.
So I just played with it a little bit and it seems like that T4 (transformer I guess ?) is causing this problem. The sensor changes the value after a small shadow created by my finger covers T4. I hope I can fix it by just covering it with something. Edit: I forgot to mention that I have V2.0
@@abc7297 Yes, I cover these with a few layers of paint and varnish to protect the components from moisture. That should shelter them from sunlight a little. Recently I 3D printed a casing to protect the components.
Hi Jason. Thanks for commenting. Comments are important to me as they give me some idea of how much interest my videos are attracting. I seem to get slightly more comments for this video at this time of year, at least in my part of the World. Probably more people gearing up for the summer growing season. It is interesting that you are the only person so far to have mentioned the anomaly. It does seem to show slightly different results after the second dip/removal cycle. Without connecting it to some more sophisticated test gear, I couldn't answer your question. Perhaps, as it is a capacitive device, some capacitance builds over time. Perhaps some stray electromagnetic energy nearby? The Arduino analogue inputs can sometimes take several reads to settle. One viewer mentioned interference from LED grow lights. It was close enough for my purposes, so I didn't look into it further. As with any sensor, a few test cycles to let the circuit settle would be wise. I think I also added some averaging algorithm to my code to help smooth things out too. Best Regards, Stewart.
Hi and thanks. I am working on fixing the link at the moment. May be a few days yet. I made the clip on my 3D printer. It is designed to hang on the side of a plant pot. If you have a 3D printer, I can send you the files. Alternatively, e-mail me your postal address and I can send you one. stewartcash(aT)hotmail(dOt)com
I`ve bought 6 pcs of these ones. Though the sensors` behavior`s very strange: while a sensor`s being ejected from water, its readings`re going up rapidly; but while the sensor is being put from the air to water, its` readings`re going down very slowly. I would actually fix these sensors using this recipe with a resistor, but I wonder about my sensors` unusual behavior... How can you help me?
Not sure I can help because I don't have one of the types you are using. So I cannot test. If you have a multi meter, test for continuity from one end of the resistor to ground. Also check to see which 555 chip is installed on the board. Please see so-now.com/electronics/capsense.php for all of the info we have so far.
@@Starter381 Did you get anywhere with your question? I have purchased a few of these sensors, but version 2.0, thinking they have been upgraded so will not have that issue... and I was wrong. Very slow response when submerged and fast response when pulled out. The circuit is the same and using TL555C, no dislodged components.
@@wojciechhaase, it seems @Stewart Cash is ignoring my sent email. Although I didn't tried his fix recipe yet. Hope it works though, because no another information about this problem is present around the Internet.
Got three of them, 1.2 & 2x2.0, one without a regulator, one without a resistor, 2/3 not working from the start, soldering the 1MOhm resistor helped still it died 2-3 days later 🤣 Good products
I did order a set of 5 of the same sensor from Amazon US and I still struggle to get a stable reading out of the box. The ADC readings in the air are constantly oscillating from 902 to 857 and in the water, ADC reads a range from 660-632 and it never stabilize. I did not try to shunt anything yet as I am searching why the ADC readings are playing yo-yo while any video I find anywhere else have stable readings straight away. I tried with Arduino Uno and with Arduino mini Pro and the yo-yo occurs every time with all the sensors. Any tip is welcome. They all are fitted with a NE555 and a 0 Ohm resistor instead of the voltage regulator. C1 is present. C4 is the only component to not respond at all when testing.
@@So-Now They are sold as 5V sensors so I use them at 5V. I modified one with the shunt and the resistor, it swings less but still is not dead steady, I had to use a smooth library in order to get somehow steady readings.
Hi again Thierry Due to the sensitivity of the ADC in the processor you are using, the variation could be nothing to do with the actual sensor. Stray capacitance, radio interference etc. could cause the ADC to show varying results. Try your code without the sensor attached, then try with a length of wire attached to the analogue pin. See what results you get. You may need to do some software averaging to get more meaningful results. Best Regards, Stewart.
Hello Stewart, thank you. I have bought some of these sensors from Amazon (DE). It seems they mine are even more broken than yours. They have the same missing connection between the 1MOhm resistor and GND. The fix is easy, as you stated. Additionally, my sensors are populated with an NE555 instead of a TLC555. The problem with the NE555 is, that it should be supplied with at least 5V VCC, they do not work with 3.3V Therefore, I have removed the voltage regulator (U2) and then I have connected the spare pins 2 and 3 (top and right) so that the NE555 gets a correct supply. For the modifications see: ruclips.net/video/Odtz8g5A-bA/видео.html Of course, you can not use such a modified sensor with 3.3V you need 5V. The readings at 5V are higher, but the sensors must be calibrated anyway. Hope this helps.
@@So-Now Certainly the link from R4 to GND is required but I have tried my sensor which has the NE555 at 3.3V and it works OK. Just the values are reduced. I know the datasheet for the NE555 says it needs 5V but mine is working. I get 1.2V in a glass of water and 2.12V in air. YMMV.
ruclips.net/video/pFQaFnqpOtQ/видео.html here are the correct version with C1 and R1 on pcb. C1 is bypass capacitor of the 3.0V regulator, it should not be big concern for the accuracy. According to the schematic, R1 is in place on pcb. But top end of R1 is not connected to ground due to the copper trace was broken. That is why R1 does not work. Simply solder or jump wire the top end of R1 to GND and it will work great. You set baud at 115200, while others set at 9600. But the reading is still delayed a bit. Do you have any idea to speed the sensor response faster?
WOW !! That is fascinating. I would never have even thought of that. I should say 'LOL' because when I read your comment, I literally did 'Laugh Out Loud'. Thanks for sharing. That is a vital piece of information.
Hi Mariusz, Thanks for taking the time to comment. The Chinese manufacturing mentality was to get a circuit, mass produce it and flood the market with cheap clones. There will be hundreds of thousands of these laying around waiting to be sold. All with the same defect. At least the fix is easy. Best Regards, Stewart.
@@So-Now Question: besides just putting the resistor across the signal pin and ground, did you also add the capacitor as well or did it just work well without it?
@@BlondieHappyGuy Hi. Thanks for commenting. The capacitor is required. But it is also connected on the board. Only the resistor is not connected properly. Adding the resistor or wire link is sufficient. Best Regards, Stewart.
Nice catch on the resistor and cap. Sadly, others who have received these sensors in the same condition, may not know how to find out what the value of that cap is. Also, they may not realize that the resistor isn't even connected at one end.
The fix is so simple once one knows.
In like manner, I use a lot of the DHT22 temperature/humidity sensors. They mostly work well, but sometimes, I've had some that seem to "bounce" a lot. Until recently (in Jan2023) I had never found out why.
No seller has this documented, but the fix to stabilize these sensors is.... yup, you guessed it, a resistor!
In this case, it's a 10K connected from the 3.3V pin to the Data pin! The sensors work best when connected to 3.3V instead of 5V, as I've also found out.
When I first discovered this, it didn't make a lot of sense to me, because to me, this was pulling UP the data line.
The data line is "digital" not analogue. But, it works! Since adding this, I find that the sensors don't bounce anymore.
I do wish that we had some GOOD manufacturers of all these sensors and devices, rather than the rip-off trash we receive. Oh well......
Now I need to find a good schematic of a circuit to use to make my own capacitive water level sensor.
Fun times.
Hi,
The value of the cap is 1uF. However, it is connected on the board. It's only the resistor is not connected on the defective ones.
There is a schematic on the link mentioned above showing all the components.
I tried the DHT22 aa couple of times, Both times they ran for a while and seemed to die with no obvious explanation.
Your problem with the pullup resistor is not unusual, even with digital outputs. Low voltage devices sometimes have open drain outputs.
The resistor helps to give a certain amout of bias between off and on.
The Capacitive moisture sensor can be used for water level, depends how deep the water is. Circuit would use Arduino or Raspberry Pi Pico.
For large volumes, I use ultrasonic distance sensors. I can easily measur contants of a 10,000 litre irrigation tank.
Best Regards,
Stewart.
Very correctly identified, the output needs to be explicitly loaded because of that stupid capacitor. Having a resistor between the Analog out and the the ground helps discharge this. This problem exists in 2.0 as well. I had multiple sensors connected and they seemed to interfere with each other. When one was pulled down, all others would go down automatically! Spent hours figuring out the problem Glad you were able to identify it.
Huge thanks, new to all this and have been struggling for week thinking it was something I was doing wrong. Added the resistor and happy days, working and it's responding as fast as all the other project videos online.
Excellent and very useful information. I had the same problem and got frustrated getting it fixed. I build my own circuit and merged it with the probe (I removed all the SMD components). All the version has the same issues.
What about waterproofing for outdoor use? Does one have to also cover (besides the electronics), the backside of the board terminal connectors?
I sand down the edges of the board so there are no sharp edges.
I then put heat shrink wrap on the cable where it connects to the board.
Then I dip the whole thing in varnish. Leave to dry. Repeat 2 or 3 times to get a good thick coat on the board.
@@So-Now thanks!
@@So-Now And that doesn't reduce the sensitivity of the sensor?
@@BlondieHappyGuyHi.
It does not seem to affect it noticably.
Best Regards,
Stewart.
I am so glad I came across your video. And that by cheer coincidence as I thought something went wrong with my ADC (MCP3008) setup. Since September I have been working on a little auto gardening hobby project for my Raspberry Pi. The capacitive soil moisture level had always been a struggle for me as I could not explain those readings (usually running towards 0 after a few seconds).
I even went so far to try another ADC (ADS1115). Only after testing with an old YL-26 that I had lying around and getting more reliable results with that one, I started questioning the capacitive soil moisture sensor 1.2.
It's a shame so many people have wasted so much time on trouble shooting.
Thanks again for sharing!
I have had a lot of grief with these sensors and have bought several from different suppliers, and there's a number of different faults. Here's what I have found:
1) v2.0 is no different to v1.2. The circuit is the same, and they contain the same faults. Don't be fooled by this feeble marketing trick.
2) The resistor R1 is not connected to ground on some sensors, but is on others. Where it's not connected, response is slow as shown in the video. The 1M Ohm resistor solves the problem.
3) Some of them are fitted with the wrong version of the 555 Timer chip (IC2). The supply voltage out of the regulator is either 3.0v or 3.3v, but some sensors are fitted with 555 that needs at least 5v.
4) Some sensors do not have the 3v regulator fitted. Instead, there is a diode, so if you're running off a battery, you'll get incorrect readings as the battery voltage declines.
5) The reading when the sensor is perfectly dry should be close to the supply voltage (in mV) and when immersed in water, should be close to zero. On most sensors that give me relatively sensible results, I get around 2500mV dry and 950mV in water. However, I did have some giving 3700mV. To use them effectively, you'd have to calibrate them and adjust the software accordingly.
Unfortunately, most vendors just buy whatever is cheapest, so you don't always get the ones they show in their photographs. It's luck of the draw whether you get a working one or not.
I had some arrive that were labelled "Capacltlve Soll Molsture Sensor v2.0". They couldn't even reproduce the label correctly, and they had R1 disconnected so reacted very slowly.
Another RUclipsr I follow tested over 30 different sensors from different suppliers and found 82% to be faulty. Buyer beware!
oh man you saved me! I was just testing mine and getting super inconsistant results.
Thank you for this post. Had the same issue with my replacement sensor. The previous one was not working at all.
What's this about doing in varnish? Where can I find more info on that?
Hi Kristoffer. Thanks for the comment.
The idea behind the varnish is to protect the board and components from moisture.
The edge of the PCB is porous and rough. The theory is that water may soak through the edges and into the PCB.
Not sure if this is true of fibreglass, but why risk it?
Sand the edges of the board so they are rounded and have no sharp corners.
Dip the board into varnish and allow the excess to drip off. Allow to dry, then repeat the process until you are confident the board and components are sealed.
Thanks for making this video, very interesting. I have bought some of the V2.0, how can I tell if I have the TLC555 or the NE555? There are no markings to say which one I have. Am I right in saying that if it powers up at 3.3v then I have the TLC555?
Hi. Thanks for commenting.
You may find there are markings on the chip that are not very clear.
Use a torch and a magnifier to look closely.
Also, if there is a regulator it is probably a 3.3 volt regulator, which means it is almost certainly a TLC555.
recently bought 5 of those on amazon (DE) all nicely packaged an in good optical condition. One seemed to work for a few minutes all the others had the same problem you discribed. ill try your fix tomorrow and report
4 of 5 are very slow - when connecting the resistor between signal and ground its going to show 0
One of them works with your fix but range goes down to 450-200 from 900-100
That new range sounds more realistic to me. You just need to calibrate your software to work with those figures.
With the ones that show zero, check the comment from Zwick Nagl above.
He found that some of these devices are fitted with the wrong type of 555 chip.
Best Regards.
@@So-Now found another seller on Ebay offering same thing but with one changed component.
He wasnt aware of that but after a quick conversation he told me that his ones are 98% working fine.
Bought 5 - all working with a Range ~300-2500 and super fast - awesome.
Unfortunately i already send the others back to test if changing the u2 Component to to a 0 ohm Resistor/ wire (if thats what im seeing here)
The PCB is also half Thickness and printed Text is double in width of the old ones.
https//imgur.com/a/WXfB8iz
@@madnessfire9081
Hi again.
Not sure what you are saying here. The range you say is ~300-2500. Was that a typo?
Anyhow, it seems there may be several problems affecting 'some' of these devices.
I am currently adding a page to my web site (will post here when it's done), just for this sensor , because its easier to update my site than it is to update a RUclips video.
Plus, I am learning more about these sensors every day as others are updating me on their findings.
Seems this device is in demand at the moment and any new information will be very helpful.
Thanks again for commenting and keep me posted if you find anything new.
Best Regards.
@@So-Now Sorry - forgot to mention that i use a esp32 with ADC Resolution of 4095bit 0-3.3v - ill take sone Measurements tomorrow
What did you cover it with
Hi Emmanuel, Thanks for the comment.
In my case, I sanded the edges, then I dipped it in undercoat paint, let it dry, then into varnish a couple of times..
A little while ago I purchased six from Banggood and all seem to be working as expected, grabbed one from Amazon due to the quicker delivery and that one was faulty.
Hopefully this means the problems been solved and is just an issue with older batches.
Thanks everybody who has commented or subscribed so far.
Sorry for not replying sooner.
I had a few comments and for some reason, they all vanished for a couple of weeks.
They suddenly re-appeared in the last few days, along with some new ones.
Thanks again.
THANK YOU!!! 3 of mine had that exact issue! Purchased months apart... 🤯
Hey, Stewart nice video you have there!
do you think the capacitive soil moisture will work if it is submerged to oil? or just the moisture from the oil?
Good question.
I could only guess as I have not tried it.
It would also depend on what type of oil. Mineral oil, vegetable oil.
I would expect they would have slightly different capacitance.
What I would like to know is, why does such a simple 1M resistor make so much difference to this circuit? Great video and so well taken apart.
Good question.
The resistor is just one component this circuit needs in order to work.
The 555 pulses on and off, charging up C4. Without the resistor, C4 would charge up and reach the supply voltage and stay there.
The 1M resistor is required to allow this capacitor to discharge, making the output variable.
The ratio of the charge/discharge time is governed by the capacitance of the traces on the sensor.
Hope this helps.
Stewart Cash the other issue I have with this unit is it is so voltage dependant for good and trustworthy readings. If I upload the same sketch to a nano then the voltage regulation is different and leads to me having to recalibrate from scratch. If using from cells then heaven forbid a drop in voltage. This is to be expected with this type of circuit - just frustrates me no end - great videos you do though, you don’t hold back - I like that.
Hi Dave,
Yes. Depending on accurate supply voltage, processor, ADC, even component tolerances on each individual sensor, the actual reading will vary.
I have found that each sensor needs to be calibrated totally dry, then submerged in water, to get top and bottom limits.
This may also need to be done more than once.
For accurate results, your software/firmware must include some form of calibration which can be repeated.
Hope this helps.
Stewart Cash Hi Stewart,
Yer, I have found this frustrating to say the least. I wrote the code for a simple water feeder and allowed a 8 second filling stage when it required and then waited 30 seconds for the water to penetrate and repeat if required. All worked like a dream, till I moved the code to an ATMEGA chip on a home made board. Then it just kept watering as the results were way out for the voltage differences - I’m now doing the same on the esp32 and have much much better stability and results.
Please keep up your very informative work - got to love electronics, got to love programming - put the two together and you have many headaches but mountains of joy.
Thanks
Dave
Very helpful as those sensors are so unreliable as they are. Your simp-e fix cured mine. Thanks. Subbed too.
Thank you very much for this video! I have a few of these, all with the same issue. How did you get to the bottom of finding the missing connection on R1, beyond measurements. Can you visibly see it on your pcb? On mine, they physically don't look like they are missing the connection but the behaviour suggests that the connection is in fact missing. Thanks again!
Hi Jacob, Thanks for the comment and subscribing.
I found it just by testing. You cannot see the tracks on the board because the solder mask is black.
However it seems that the large ground track, either side of the connector, has been broken by adding the connector at the design stage.
If you scrape off a little solder mask each side of the connector and solder a jumper wire across, this also fixes the the problem without adding a resistor.
I have ordered a few more and will sacrifice one to prove my theory. I will make a video if I am correct.
Best Regards.
Hi Stewart,
Thank you for the video. My v1.2 sensor has 0 ohm resistor instead of voltage regulator. Moreover I am using it with nodemcu v3 with easy_esp installed. I found that when I supply sensor with esp 3.2v then ne555 has 360kHz and the output varies between 0.8v in the water and 2v in the air. But when I use 4.5v form usb then ne555 has round 413kHz if I recall it correctly and output between 1.8 and 3.2v. I've double checked and nodemcu v3 recognise 3.13 as 1024 (max value). I haven't damaged it yet but I'd rather be careful.
But the most important part is that when I collect all this data in domoticz it's fluctuating +- 2%. I was trying to find some good tips in the Internet but without success. I don't know what might cause the fluctuations.
Hi Marcin.
Thanks for the comment.
I have been told by other viewers about the zero Ohm link in place of the regulator.
This is because the manufacturers have replaced the TLC555 with a NE555.
The TLC555 can run at 3.2 volts. The NE555 must have 4.5 to 5.5 volts.
You would have to supply it with around 5 volts but then you could not drive it from a 3 volt processor.
I have never measured the frequency of the output, but the resistor and capacitor would tend to smooth that out.
The difference would be, higher frequency, slightly faster rise and fall time.
Thanks Stewart , really valuable feedback.
I purchased 4 of the V1.0 sensors from DFROBOT and all the parts are present and appear to conform to their circuit diagram.
I'm planning to remove the 3 pin socket and directly attach a screened cable pigtail, then spray coat the whole board with several applications of CRC urethane seal coat.
This will hopefully provide sufficient protection for direct burial below the lawn.
Did you observe any adverse impact on the sensor sensitivity due the additional sealing you applied ?.
Hi Mike.
Thanks for the comment.
On several of these devices, I have removed the connector and replaced it with 2 metres of screened cable.
I then carefully sand the edges of the board, then dip the whole device in paint or varnish.
I am lucky enough to have a 3D printer, so my last step is to print a small clam-shell case to cover the component part of the board.
I then snap this shut over the board with epoxy resin to seal it all shut.
Hope this helps.
Thanks Stewart, that’s useful information and gives me confidence that I’m going in the right direction. As I plan on using four buried sensors to monitor the lawn on each side of my house, I’ve been trying to decide whether to separately extend back the sensor voltage on separate cables to connect with a data logger, or use a bus interconnection and digitise the moisture signal at the sensor with a microcontroller. After investigating, it appears that professional grade sensors use an digital SDI-12 interface, which appears reasonably easy to implement with a cheap 8 bit microcontroller and an bus driver circuit. I’ve decided to go with digital interconnection of the sensors, so I’ll be putting the sensor electronics, microcontroller and bus driver circuit in a small enclosure and ‘potting’ the assembly with urethane sealant.
Thanks for getting back to me.
@@mikegofton1 Hi again Mike.
I was just re-reading through the comments and came across this one again.
How far have you got with your project?
I was thinking, maybe you could look at using a CAN-Bus interface.
We used to use it on vehicle based systems (as it was designed for) however, we also used it on a mini underwater drone tether.
Up to 1000 metres cable length, multiple devices on a single twisted pair.
Could maybe use network cabling if you can get enough current down the other pairs to power your devices.
A couple of driver chips and some fairly simple to understand libraries for the Arduino.
Maybe worth looking at.
Best Regards,
Stewart.
@@So-Now I'm using an arduino pro mini to interface to a 3 wire serial bus (5V half duplex 1200 bps asynch, same as the physical level for SDI-12 ), and JSON to transport messages rather than the SDI-12 protocol messages. There is a SDI-12 protocol arduino library, however it uses software 'bit banging' - I prefer to use UART hardware. Most SDI-12 sensors are also professional grade and expensive, which seemed like overkill for my garden.
I've built the hardware / written software, and will test the system this week sometime.
@@So-Now Hi. i bought a version 1.2 sensor (or rather, I bought version 2.0, but the seller deceived and sent 1.2), you can see the U2 element (on three contacts) on the video at 2:50, it is removed on my sensor, the upper and right contacts are connected by a resistor (or something similar), tell me if your advice will help to fix problems with the sensor? Thanks.
Greetings Sir,
Will it still work if I use a 10 k ohm resistor?
Thank You
Hi.
I never tried it, but results might not be very good.
The output depends on the capacitor charging up and being discharged by the 1Mohm resistor.
You will not damage anything by trying it. Let us know.
How can I connect it sir?
Do I have to solder it to the component itself ?
Can I connect it to the wires?
Sorry because I don't have that much information or technical knowhow about this.
Thank you for your response!
thank you for this fix there is still people out there selling these as i order some from eBay and AliExpress and both were 1.2 V same as yours but Im going to fix them now but i bought new ones with V2.0 thanks
New info added to: so-now.com/electronics/capsense.php
Soon to become a major new motion picture/RUclips video.
Have you come across the problem with a grow light interfering with a capacitive soil moisture sensor?
No. But I could imagine it happening.
The soil moisture sensor is a capacitive stick. Which is basically how we make antennae (antennas for the Yanks), aerials.
It would therefore act as a magnet for any stray electromagnetic fields floating around.
If the grow light runs at any frequency, if it has some dimming control for instance, it could very well cause some problems.
One thing to try is using screened cable to connect to the sensor.
Most often, interference finds it's way in through wires (again, acting as aerials).
I would be interested to know what type of interference you are seeing.
Is the grow light LED or flourescent?
Do you have an oscilloscope?
Thanks for commenting. It's good to see there is still some interest.
Best Regards,
Stewart.
Confirmed: missing link between 1Mr resistor and gnd. Due to faulty layout so a lot will have the same. Grrr....
Also NE555 instead of TLC555 => no startup @ 3.2V. 2nd Grrr...
Thanks a lot for your video !
Hi. Thanks for the comment.
NE555 will not run at 3.2v. See:
so-now.com/electronics/capsense.php
for more info.
Thanks for this tip! I had the situation that the sensor was in the ground and would read 0 all the time. Whenever I touched the plant, it would start working. I guess I was the 1M resistor to ground :-)
Thanks for the comment.
It seems this problem has many different symptoms.
Which method did you use to fix it?
@@So-Now Honestly, I haven't fixed it yet. It just seems to make sense that your method may resolve the issue for me as well. I'll let you know the output once I get hold of a 1M resistor.
You saved my project! Hero!
You made my day! That solved my problem I was puzzeling on since weeks ...
does this also fix the problem that the measured voltage decreases (eg. from 1.26V to 1.18V) when the sun shines on the sensors ?
I want to use them outside but that behaviour makes it a bit more difficult
Thanks for commenting.
I have not heard of that particular fault, though it sounds likely.
Any of the components on the board could be affected by heat.
I would have to test by heating one up.
So I just played with it a little bit and it seems like that T4 (transformer I guess ?) is causing this problem.
The sensor changes the value after a small shadow created by my finger covers T4.
I hope I can fix it by just covering it with something.
Edit: I forgot to mention that I have V2.0
@@abc7297 Yes, I cover these with a few layers of paint and varnish to protect the components from moisture.
That should shelter them from sunlight a little.
Recently I 3D printed a casing to protect the components.
Is there a reason why it didn’t get the same result twice when you submerged it the 2nd time once it was repaired?
Hi Jason.
Thanks for commenting. Comments are important to me as they give me some idea of how much interest my videos are attracting. I seem to get slightly more comments for this video at this time of year, at least in my part of the World. Probably more people gearing up for the summer growing season.
It is interesting that you are the only person so far to have mentioned the anomaly.
It does seem to show slightly different results after the second dip/removal cycle.
Without connecting it to some more sophisticated test gear, I couldn't answer your question.
Perhaps, as it is a capacitive device, some capacitance builds over time. Perhaps some stray electromagnetic energy nearby? The Arduino analogue inputs can sometimes take several reads to settle.
One viewer mentioned interference from LED grow lights.
It was close enough for my purposes, so I didn't look into it further.
As with any sensor, a few test cycles to let the circuit settle would be wise.
I think I also added some averaging algorithm to my code to help smooth things out too.
Best Regards,
Stewart.
Update info link doesn't work anymore :-( where did you get that clip to hold the sensor?
Hi and thanks. I am working on fixing the link at the moment. May be a few days yet.
I made the clip on my 3D printer. It is designed to hang on the side of a plant pot.
If you have a 3D printer, I can send you the files.
Alternatively, e-mail me your postal address and I can send you one.
stewartcash(aT)hotmail(dOt)com
@@So-Now thanks, I do have a 3d printer so the file would be awesome. I'll email you. Thanks again 😊
I`ve bought 6 pcs of these ones. Though the sensors` behavior`s very strange: while a sensor`s being ejected from water, its readings`re going up rapidly; but while the sensor is being put from the air to water, its` readings`re going down very slowly. I would actually fix these sensors using this recipe with a resistor, but I wonder about my sensors` unusual behavior... How can you help me?
Not sure I can help because I don't have one of the types you are using. So I cannot test.
If you have a multi meter, test for continuity from one end of the resistor to ground.
Also check to see which 555 chip is installed on the board.
Please see so-now.com/electronics/capsense.php for all of the info we have so far.
@@So-Now I had emailed you to stewartcash@hotmail.com and supplied more information. The chip type is 9 1M NXNG4. What does it mean?
@@So-Now sorry for bothering you, did you receive my email?
@@Starter381 Did you get anywhere with your question? I have purchased a few of these sensors, but version 2.0, thinking they have been upgraded so will not have that issue... and I was wrong. Very slow response when submerged and fast response when pulled out. The circuit is the same and using TL555C, no dislodged components.
@@wojciechhaase, it seems @Stewart Cash is ignoring my sent email. Although I didn't tried his fix recipe yet. Hope it works though, because no another information about this problem is present around the Internet.
Thank you! You just got yourself a new subscriber.
You just saved my project! Thank you for posting this!!!
Thanky you, simple fix actually. Save me a lot of time.
Now I don`t need to throw them out the window ! Nice
merci pour cette video 👍
Got three of them, 1.2 & 2x2.0, one without a regulator, one without a resistor, 2/3 not working from the start, soldering the 1MOhm resistor helped still it died 2-3 days later 🤣 Good products
Great information, thank you!
I did order a set of 5 of the same sensor from Amazon US and I still struggle to get a stable reading out of the box. The ADC readings in the air are constantly oscillating from 902 to 857 and in the water, ADC reads a range from 660-632 and it never stabilize. I did not try to shunt anything yet as I am searching why the ADC readings are playing yo-yo while any video I find anywhere else have stable readings straight away. I tried with Arduino Uno and with Arduino mini Pro and the yo-yo occurs every time with all the sensors. Any tip is welcome. They all are fitted with a NE555 and a 0 Ohm resistor instead of the voltage regulator. C1 is present. C4 is the only component to not respond at all when testing.
What voltage are you supplying the sensor with?
@@So-Now They are sold as 5V sensors so I use them at 5V. I modified one with the shunt and the resistor, it swings less but still is not dead steady, I had to use a smooth library in order to get somehow steady readings.
Hi again Thierry
Due to the sensitivity of the ADC in the processor you are using, the variation could be nothing to do with the actual sensor.
Stray capacitance, radio interference etc. could cause the ADC to show varying results.
Try your code without the sensor attached, then try with a length of wire attached to the analogue pin.
See what results you get.
You may need to do some software averaging to get more meaningful results.
Best Regards,
Stewart.
Very informative. Thanks a lot
Thanks for taking the time to comment.
Always good to know my video is useful.
Thank you. In 10 units I got 3 with dislodgd capacitors...
Very cool, thank you.
Thank you
Thank you. It works.
Hello Stewart,
thank you. I have bought some of these sensors from Amazon (DE). It seems they mine are even more broken than yours.
They have the same missing connection between the 1MOhm resistor and GND. The fix is easy, as you stated.
Additionally, my sensors are populated with an NE555 instead of a TLC555. The problem with the NE555 is, that it should be supplied with at least 5V VCC, they do not work with 3.3V
Therefore, I have removed the voltage regulator (U2) and then I have connected the spare pins 2 and 3 (top and right) so that the NE555 gets a correct supply.
For the modifications see: ruclips.net/video/Odtz8g5A-bA/видео.html
Of course, you can not use such a modified sensor with 3.3V you need 5V. The readings at 5V are higher, but the sensors must be calibrated anyway.
Hope this helps.
Yes, that is a great help. Thanks.
I will add that info to the web page I am working on.
Best Regards.
@@So-Now Certainly the link from R4 to GND is required but I have tried my sensor which has the NE555 at 3.3V and it works OK. Just the values are reduced. I know the datasheet for the NE555 says it needs 5V but mine is working. I get 1.2V in a glass of water and 2.12V in air. YMMV.
it works thank you
It's great
ruclips.net/video/pFQaFnqpOtQ/видео.html
here are the correct version with C1 and R1 on pcb.
C1 is bypass capacitor of the 3.0V regulator, it should not be big concern for the accuracy.
According to the schematic, R1 is in place on pcb. But top end of R1 is not connected to ground due to the copper trace was broken. That is why R1 does not work. Simply solder or jump wire the top end of R1 to GND and it will work great.
You set baud at 115200, while others set at 9600. But the reading is still delayed a bit. Do you have any idea to speed the sensor response faster?
And i was thinking "Why it works on my osciloscope??" then i watched your video and saw that my osciloscope has 1MOhm probe...
WOW !! That is fascinating. I would never have even thought of that.
I should say 'LOL' because when I read your comment, I literally did 'Laugh Out Loud'.
Thanks for sharing. That is a vital piece of information.
Similar result with my dmm. Working with measuring the voltage with the dmm, but won‘t with my arduino
Two years later, they are still selling this crap…
Hi Mariusz, Thanks for taking the time to comment.
The Chinese manufacturing mentality was to get a circuit, mass produce it and flood the market with cheap clones.
There will be hundreds of thousands of these laying around waiting to be sold. All with the same defect.
At least the fix is easy.
Best Regards,
Stewart.
@@So-Now Question: besides just putting the resistor across the signal pin and ground, did you also add the capacitor as well or did it just work well without it?
@@BlondieHappyGuy Hi. Thanks for commenting.
The capacitor is required. But it is also connected on the board. Only the resistor is not connected properly.
Adding the resistor or wire link is sufficient.
Best Regards,
Stewart.
I've got 5 of them ffs lmfao 🤣
LOL.
All faulty ?
Shouldn't take you long to fix 'em.
Best warm up that soldering iron.
danke danke
Bitte schoen
Yes or you may replace r1 by a 500k
anonymous of arduino