EDIT: A portion of this video has been trimmed out because of an error in the shown formula of capacitance. I apologize. Thanks to "All in the Familiy" for having spotted the error. I have to say that I was not in comfort with that formula from start, at the point I regret not having cut out the whole part from the video in the first place. That happened because I had a mismatch with theoretical calculated values and the measured values and being in hurry to publish the video and being this a simple showcase video I didn't keep it on hold until that point was cleared. Sorry. More in the reply I've given to "All in the Familiy"'s comment below. Set apart this, the remainder of the content in the video still hold and after almost four years of operation I have to say that this sensor worked flawlessy, so I think it remains an useful piece of information for DIY electronics. I will try to do more about this topic. Thanks for sticking with me. Remark. Because the dielectric constant of water changes with temperature this sensor provides a reliable measure as long as the water has a limited variation in temperature. In the case the water has variable temperature (say from 20C to 90C / 68-194F) a pressure sensor is better suited. (Pipe-pressure, the same that I'm gonna make for gauging the level of the sludge into a manhole.)
You mentioned doing another sensor for your "sewage manhole". Did you ever get to this project? I am very interested in doing something similar but encasing the capacitor plates in molded plastic.
You also could add a temp sensor to the bottom of the capacitor sensor and then design a bias ckt to make the comparator circuit automatically adjust (compensate) for temperature variations....
I built similar tubes a few years ago using copper tape and pvp pipe. It was for an aquaponics setup and designed so that the exposed portion was the hollow center and could be easily cleaned with a pipe cleaner brush. I appreciate your coverage of the maths in this and your rapid pcb production technique is spot on!
A VERY INTERESTING FULL EXPLANATION THAT I HAVE NEVER FOUND ELSEWHERE, EVEN AT THE PROFESSIONAL AND COMMERCIAL SITES (AS THEY DO NO LIKE TO LET THE PUBLIC KNOW THE SECRETS OF THE INVOLVED PROCEDURE). THANK YOU VERY, VERY, MUCH.
Thank you again I designed a new type of level meter that fit my device. It's totally different from yours, but I got a hint from your video. Your new attempt has been good teacher for me.
I initially thought about making a sensor like this, but as a level transmitter for my arduino. I planned on using a 555 in a monostable configuration in order to read its actual capacitance value. Then after would have to calibrate it in order to correlate its level to its reading. [Edit] I shared your video to GreatScott. He had a project recently to monitor a water tank level and used an ultrasound sensor in the end. But I wanted him to be aware of your creation :)
I used a common electrical cable with two internal wires as a sensor, I melted the tip of the outer sheath to isolate the two cables, and as an electronics I used a common 555 whose oscillation frequency is read by arduino, works perfectly even as a sensor of soil moisture :)
@@syedumarjunaid5499 www.electroniq.net/555-timer-circuits/square-wave-generator-using-555-timer.html - the probe is a dual isolated cable connected instead of the 'frequency select' capacitors, consider that a cable about 80 cm long has about 80 pF.
Nice one. Can you please answer some, 1. What is the maximum and minimum voltage at f/v converter ? 2. How this f/v converter constructed with just RRC passive components ? 3. What is the maximum output drive current of NAND gate ? 4. What is the Hysteresis voltage? 5. Does MCP3302 work in place of lm311 (from the data sheet which you have choosen to calculate input impedance of IC, “typical (or) maximum” values of offset voltage and bias current ) ? Thanks again.
Hi sir, can you explain me circuit how the capacitance is measured in the fuel tank the value need to measure in the microcontroller the capacitance value ranges from 50pf to 500pf.
hi , I liked the the capacitance sensor that you made. Would it be possible for you to share the electronic circuit diagram with specifications of components needed? I am looking for making similar type of application for fuel level gauging. If you have any suggestions you are welcome.
Hi, were you able to get the electronic circuit diagram with the specifications of the components? I am a bit stuck with a project similar to this and it would be much appreciated. Thanks.
Very nice, very nice video with a lot of useful info and also the necessary theory behind. I wish you would have spent a little longer on the math and on the circuit components. This said, why not make the circuit with a proper CAD and use the transfer method? Also, I think disconnecting AC whit a relay when the tank is full is a poor failover method. Why not create some sort of passive overflow system that safely discards the excess water? It could be a series of pipes that passively discharge extra water or simply positioning the tank on a drain. Last question: will water collected from a condenser be slightly acidic? Is aluminum a good material for this or would something like a stainless steel be better?
Sorry for not covering deeper the math involved, but that would be beyond the scope of this channel, also I am not a matematician ;) The passive method for discharging the excess of water would it were possible wouldn't had rendered this circuit necessary. Also, since the water is produced only if the AC is on, even in the case of power outage this would not cause any problems making the circuit intrinsecally safe. I don't know if the water become slightly acidic, I did never tested it. The aluminum of the plates is covered with resin so there is no direct contact of the aluminum plates with the collected water.
Okay, it works, but why did you go to the trouble of building what is essentially a capacitive liquid-level sensor capable of measuring the full range of liquid levels in the tube, when all you needed was a top-level discrete switch to turn on a warning? My experience with these circuits was as an aircraft mechanic. As I remember, the capacitive liquid level sensor was developed in Britain late in World War II to fulfill the need for an aircraft fuel-level sensing system that was accurate in flight and had no moving parts. It is still widely used today, although ultrasonic level sensors are now used in new designs. One of the most reliable and trouble-free systems I encountered was used in Convair airliners equipped with twin turboprop engines, the Convair 580. Based on a vacuum-tube amplifier, it used a simple capacitance-resistive bridge circuit to drive a servo motor inside the instrument that positioned the indicating needle at the scale position for the sensed fuel quantity, and also drove a potentiometer wiper that rebalanced the bridge circuit for every indicator movement. The capacitance input was provided by the total of all the in-tank, tubular capacitors connected in parallel, plus a separate reference capacitance signal from a submerged compensator. It worked amazingly well, and better than some of the digital capacitance circuits found on later aircraft models.
That sensor with pot feedback based on a vacuum tube, you nicely described is interesting indeed. I chose the capacitive sensor for three reasons: it is easier to adjust the alarm level, it's robustness because it has no moving parts and it cannot be damaged when it is pulled off the tank (when it has to be emptied), and to have fun 😊
Nice video inspired me to do it. I tried to make one. And modified the probe. 1. Probe capacitance varying 160pf to 205pf, so i calibrated RC ckt to tune with same, modified R 150kohm to 42kohm. 2. From f/v output I wanted to get 4 different outputs, so placed 4 individual comparators to glow 4 LEDs for 4 stages of water level 10%, 35%, 65% and 95%. And the comparators output given to RF encoder (HT12E) then 433mhz transmitter. Please do let me know if my considerations would go wrong. Thanks!.
I think your design should work correctly, in particular if the comparators have a decently high input impedence, which usually it is the case in particular if you use the LM311. Of course you can change with other comparators, maybe a package that contains four. Please note that the schematic shown in the video has an error: I forgot to draw a line from pin1 to pin4 of the LM311. The correct schematic is visible on my website. Greetings from the Alps.
Look at the "scale with a can" video, the principle is pretty much the same. EDIT: The video is about weighing with a tin can, here it is the link: th-cam.com/video/meluEWZpsXY/w-d-xo.html
I'm thinking of making another one of these using a slightly different principle. The last one I made was really difficult to set up (was from EDN online magazine and cut down a little for a water level monitor for a boat, copper tubing around a plastic pipe with an insulated electrode down the centre, making a co-axial capacitor). I was thinking using a fixed frequency, and then having an output that tells you when the capacitor is charging or not 1=charging, 0=not charging to give a PWM output that is then filtered and can give a wider range output without too much amplification. Capacitor discharged quickly at rising edge of frequency generator, maximum frequency needed determined by capacitor and charging resistor time constant, at the time constant's highest value (variable charging resistor to adjust for maximum).
Just out of curiosity, may you provide the link to that article? The coaxial version of the sensor can be seen in my video at 4:21 . I don't get how your solution works: since the capacitor is always charging from the raising edge to the falling one, and alwasy discharging from the falling edge to the raising one, if you look at the current (to see the charges' flow) then this current will lasts for the full positive cycle (and the full negative cycle on the other way). I am not saying it does not work but likely I missed something in your explanation, perhaps is your idea to cap the charge cycle at a given voltage by setting a threshold so that you can look at how long the capacitor takes to charge (up to that given voltage), switch to discharge, and hence generating a sort of PWM? That would be equivalent to the "method 1" I described in the video (at 2:29). Anyway, good luck with your project, also let us know about the results. Thanks for having commented. Cheers.
@@AccidentalScience Sure, the EDN article is this one: www.edn.com/design/test-and-measurement/4423158/Precision-capacitance-meter I used an LM3914 bar graph to show water level. I had trouble with getting the right range with the EDN circuit - I think it is designed for much lower capacitance and smaller variation. I guess I wasn't too clear, I was going to let the capacitor charge through a resistor and discharge it at regular intervals with a small pulse, like in your weight sensor video, but without a constant current and just a resistor (that you can vary too). The idea was when the capacitor is charging (below threshold) a 1 is output and 0 above threshold to make a PWM signal controlled by the discharging repetition frequency. I'm not sure if it's functionally equivalent to yours that uses PDM (pulse density modulation).
I've since found a circuit that would work in a similar manner to what I was thinking, although I've already designed my own version (with it's less than ideal behavior, but can compensate for that). I'm also trying a sensor with the elements wound in a spiral around a small diameter tube, hoping to get a measurable capacitance change between them as covered by the fluid (similar construction to an open wire line as used for RF). As the fluid is only on the outside (the tube can be blocked or a rod used) then it should be ok for measuring fluid with 'sludge'. PWM type sensor - here used for a capacitive humidity sensor: www.digikey.com/en/articles/techzone/2011/oct/humidity-sensors-and-signal-conditioning-choices
Sorry for the late reply, I have noticed today that your last comments were kept held by YT for review ...it's a plague common to other commentators, that I have to clear manually. About the circuit. I've seen the one on E D N, it is based on a fixed frequency and the capacitance is read by comparing a fixed time charge (the branch R14 and CA1D in their schematic) and the variable time charge (due to the test capacitance C?) on the branch R11 CA1C that is conditioned by the availability of the current that split at the node R8-R11 through the BAV99 "sucked" by the test capacitance. Thus, at the input of the opamp would appear kind of a race between the fixed time charge and the variable one from the two branches. If I did not missed something, because CA1C is quite large (10nF) the test cap should be in a comparable range otherwise the difference would need a large amplification, they set to AV=10. Also, the two CA1x capacitors should be matched for the best results, but that should be secondary for the functioning of the circuit. It seems a complex circuit to me, and frankly I don't like it. One thing you have to keep in mind is the design of the layout or the PCB, as the devil hide in details. Beware of every stray capacitance, and leakage from the surface of the circuit because the common (cheap) FR4 adsorb non trivial amount of humidity from the air (and after the etching process), so after cleaning the PCB with (warm) isopropil or ethilic 99,9% alchool (which can grab up to about 30% of water in it), it should be backed at 60C for 30' before covering it (still warm) with a good quality conformal coating. Also, the spiral idea do not convince me, it could become an inductor that may cause weird behaviours, pick up RF or LF EM ...and cause headakes :) To have a good capacitance simply reduce the distance between the plates. I've been very busy the last three weeks, but I will upload an other video on capacitive sensing. Cheers.
I want use this project for measure vegetable oil level on a bottle. I have a question: if the oil is dielectric, Can I use the electrodes without epoxy resin? I'm thinking make the electrodes with stainless steel.
Epoxy covering (actually I used polyester) is not aimed to protect against corrosion but to provide insulation. Look for a food grade covering or just use bare metal as oil is not conductive. However oil has a quite different dielectric coefficient and both the plates and the circuit would require a redesign for such an application.
sorry folks that is incorrect, the formula for capacitors with stacked dielectric materials is C=Eo*A / ((d1/Er) + (d2/Ew) + (d3/Er))), Eo=8.85E-12F/m, Ew=10, Er=3, A=area,(lxw), d1=thickness resin plate1=1mm, d2=thickness air/water=3mm, d3=thickness resin plate 2=1mm,. With these values the calculated capacitance with air is less then 1pF namely 0.8pF, with water the capacity is 3.1Pf.
First off, thanks for your comment. You somewhat put the finger where the spot hurts, since that is the formula I used at first when I tested this circuit and probe but then I used an alternative formula because it didn't fit the experimental results. Well, you got the wrong result because you put 1 mm thickness for the resin layer when it is more like .1 mm (not sure about the exact thickness), but that does detract nothing to your point: that should be the correct formula and it is basically where the formula for series capacitors comes. Because intuitively a stack of dielectrics sould behave as a series of capacitors. But as I've said the results didn't fit experimentally. My first theoretical results (using the formula you mentioned) were: 9.9pF dry and 219pF full of water (apart a small portion); while the actual measurements were 34pF dry (measured with a capacimeter) and 2.34nF in water (calculated with the circuit). They were too off. At first I thought to have some leakage, a resistor in parallel or maybe a series resistor caused by a bad junction. But after thoroughly checking with the multimeter I didn't find any significative resistance that could have changed the response of the circuit and the resulting capacitance of the probe *that much*. So I searched online and found a paper where they said something related to the density of charges, giving a formula where the relative dielectric constant (permittivity) of multiple layers was simply multiplied to the permittivity in vacuum. Unfortunately I didn't keep a reference. However that formula gave results that was close to the experimental data so I trusted that formula. Though I never felt in comfort with that and I should have skipped the theoretical part in the video. This is what I've done now: I trimmed off that part! That said, still something do not add up properly because the experimental data were way too off from the theoretical data. As soon as I'll have a bit of time I will try to go deeper into this thing. I can't pin your comment, but definitely I will include it in some way into my already pinned comment. Thank you again.
Hi , excellent video . I got two major issues when I made cylindrical capacitive diesel sensor. Please can some one answer following queries. : 1, it suddenly changes reading ,mostly to 100% or even more than that. It may be because of " charge accumulation on central rod ". Mind it I did not coat with resin/acrylic. 2, it changes reading with temperature . e.g. increases in afternoon and decrease in night. 3, wire b/w probe n circuit also causes reading drift.
Missing coating makes the probe resistive. Diesel is not a conductor but it could mess up the capacitive vs. Resistive ratio. Not sure though, it would require a thorough investigation. Wires cause parasite impedance so the probe must be directly connected to the circuit: the two parts must be seen as one.
You could calibrate an existing sending unit from a fuel tank. Attached to a gauge it would give an accurate reading of the fluid level. Or, just a float attached to a switch that triggers a warning light when the fluid level in the tank is too high or too low.
Thank you. The thickness of the aluminium strips is irrilevant, and as long as it is a good conductor the material is irrilevant too, soo even foil copper can be used.
My experience is that when titanium is used for the ungrounded probe... It's size, and the thickness of any titanium oxide/anodised layer has big effects on the capacitance. My understanding of this while "tap water" is a poor conductor... Compared to titanium oxide it's not that bad a conductor... So in effect the capacitor is formed between inside of plate, and water. Ie across the titanium oxide layer which 1) is very thin 2)has a fairly high dielectric constant.
Hi, I really like your video and appreciate your efforts. I have some questions and i hope you would like to answer these. 1. Can I make 4.5 feet long capacitance probe? 2. How long wire from probe terminal to circuit is recommended. 3. What is accuracy if water changes from clear water to dirty (contains mud or other particles) water, does it give same reading? 4. Is it mandatory to use Acrylic Color on top of Crystal resin coating? 5. Can we use any other material than Aluminum for probes? 6. If Arduino have ADS, will these capacitance reading be readable in Arduino? Regards,
1. Yes, requires tuning. 2. As short as possible. This is one of the trickiest part when designing these kind of circuits. 3. I don't know, I had to make a test but never had the time to do it. 4. No. But insulation and the dielectric of the insulator is of paramount importance. 5. Yes, provided it is a good conductor of electricity. 6. What ADS stands for? If it was meant to refer to ADC, as long as the ADC input is decoupled from the primary stage of the circuit (after the F to V conversion) it does not pose any problem. However for a professional design I would not use an ADC, instead I would directly gauge the frequency. Cheers.
Hi, I just saw your video and it is great. You are really an engineer with hands on. Is it possible to speak to you to discuss a problem that I am having at the moment? Thank you
Leave your question here, I'll see what I can do. Instead, if you need a consultancy drop an email with a detailed description of your problem, and I'll give you a quote. Email address can be found in the about section of this channel.
Very informative video! I have a 20gal chlorine tank used for well water treatment and would like to use an ESP 8266 to send an MQTT message to my node-red/raspberry Pi system so as to flash one of the lights in my living room which is connected to a Sonoff Basic running Tasmota. Flashing the ESP 8266-01 or some other variant to send a MQTT message at some level {approx 20% full} might tax my limited programming skills however.
I want to replicate this project. Can you please answer few of my queries? It would also help other, those who visit this video! 1. Can I attach the probe separately with a wire and keep the circuit at a distant? 2. Which wire should I use? Twisted pair or Single core shielded? 3. Can I use this same circuit for moisture count in soil?
Answers: 1. No. The closer the probe is to the circuit the better the sensor will work. I attached the probe with two short leads for testing purposes only, but those leads was source of noise and troubles that I compensated in some way ...you know, experience helps here :) 2. see 1. 3. No, the circuit and the probe have to be redesigned for that specific purpose.
Very good! I would just like to ask which integrated circuit used 8 legs that appears on the left side of the scheme in the video, when mounted the board could not identify, is a 555 oscillator? grateful for the help
It is a 393 dual comparator (one of the two is used only). It is mounted in the etched board, not on the breadboard (where you can see an other opamp because I've been simultaneously testing an other circuit not discussed in the video).
It's the first NAND gate. Looking at the configuration you can guess that it's used as an inverting gate. Being a Schmitt trigger is fundamental to operating properly. Happy new year!
Can you change this circuit for 4ft and 5ft lenght tank and give me components list and PCB layout because I don't have oscilloscope that's why I can not do this calculation. I will waiting for your prompt reply. Best regards Umar
OK but I'm trying to do this outdoors for rainwater containers. Complex circuitry breaks down due to humidity. I tried to enclose my circuits in jars and cover the jars from the rain but humidity still gets in. Only simple conductive circuits seem to work reliably -- nothing with timers or oscillators has worked for me yet. Any chance you could make one of these with components that don't care about moisture? Thanks!
This is already meant to be moisture proof. You have to spray the circuit with (insulating) lacquer. Or put it in a box and fill the box with epoxy. Conductive circuits suffer of corrosion and moisture find a way to get in easier than with capacitive that can be completely insulated.
@@AccidentalScience Thank you for your reply. My question really was whether some components have better moisture resistance than others and whether they can be used preferentially to build a circuit such as this. I think the most effective way to keep moisture out of the circuit is to submerge it in a non-conducting liquid that's heavier than water. That way the circuit stays dry even if there are small cracks in the container. But these are just brute force approaches. I was wondering if there was something more elegant available.
@@SashaXXY all electronic components have more or less the same problem with moisture. The only standard technique used to prevent this is lacquer or epoxy coating aka conformal coating.
If the level sensor will connected to the pcb using about 1/2 meter thin coaxial wiring could this cause problems to the high impedance input of oscillator's schmitt trigger gate?
Hi... ¿How can I find information about that oscillator and pulse generator? Now I know I can simulate it in a computer program or built mine in house but in both cases if I want to model it mathematically I would have to measure... In resume: could somebody tell me where can I see theory about oscillator and pulse generator?
While I don't have a specific answer right now, a good starting point is to understand the theory of operation. Some more information is available on the relative page on my website, but there is literature (I think it should be available online) that covers oscillators and other signal manipulation through logic gates. I have an old book titled "Elettronica Digitale" (don't remember the publisher though and AFAIK it's only in Italian) that provides good coverage on the topic. Maybe searching for a similar title you may find something.
@@AccidentalScience Yeah... Thnx for your answer... Another question I have is: Did you review your book for design that circuit? I can understand how the oscillator part works by seeing connections in NAND, resistors and capacitor because it is the first part but I can understand pulse generator and the other parts by simple inspection 😓
You mentioned doing another sensor for your "sewage manhole". Did you ever get to this project? I am very interested in doing something similar but encasing the capacitor plates in molded plastic.
Wanted to ask if the volume of the water will have any effect on the capacitance? What I mean is if you use a bigger tank with the same height, will it needs to be calibrated
Can anyone please explain why you can't just measure the change in resistance between two closely separated metal plates? It seems that the circuitry (and maths) involved would be much simpler! Obviously, you wouldn't want it running all the time because you'd corrode the electrodes, but using something like an Arduino you only need to power it for a few milliseconds when you're actually measuring.
Obviously you can. Provided you have tap water, or salt water. With water coming from HVAC which is far less conductive things become a little bit complex. More, water changes its conductivity because variations in dissolved minerals way more than its dielectric constant. That said I made a water level sensor using stainless steel rods of various length and a simple row of LEDs that worked nicely for years. In a non-linear fashion like that stainless steel rods are way more convenient. Maybe I will make a video about this point. Thanks for commenting.
@@balajiramesh6814 oh gotcha. Water have a typical dielectric constant (AKA relative permittivity) of 80. This is a dimensionless number being a ratio. See here for more: en.wikipedia.org/wiki/Relative_permittivity The relative permittivity (or dielectric constant) does not change by varying the immersion of the sensor's probe, what it changes is the ratio between air and water between the plates, which makes the capacitance to change because the dielectric constant of air is ~1. The relation between depth of immersion and capacitance can be calculated by the ratio of the segment that is not immersed using the dielectric cons. of air, and the segment that is immersed using the dielectric const. of water. Through this means you can make a relationship between depth in mm and capacitance, or frequency (after the oscillator stage) or voltage (after the VF converter stage).
ESP 8266. You obviously need to make a program in order to connect and see the output on a smart phone. The program has to read the analog signal (at the point indicated in the schematic) and generate some kind of user interface to display the information, i.e. a web page, and operate as a wifi access point.
Your question is a good point and to try to give an answer we should wonder what are the reasons limescale/minerals build up: 1) it could be caused by sedimentation, in this case there is no problem because the gravity will attract the particles down, outside the pipe; 2) it could be caused by chemical concentration, at the point the particles stick to the surfaces, this typically happens when high temperatures are involved, and it is usual in resistors used to heat up boilers, however the stripes are not heated (and with no high temperature spots the particles stay diffused). However after long time some particles will stick anyway but far from the point to cause the pipe be chocked; 3) it could be caused by electrical attraction: if the plates (the stripes) were connected to a continuous electric field that could attract polar minerals eventually. However here not only the electric field is very small but it is also alternating so there is no risk to attract polar particles. About the thin layer of minerals/limescale that may build up in long times, mentioned at the point 2 above, one thing that we may wonder is if that layer may modify the dielectric properties of the sensor, thus shifting the returned measured value. I have no answer for this last point because I think the only way to give an answer is to try experimentally. However since the sensor as shown in the video is quite simple I think that for the purpose of its use it should be fine anyway, it is however possible to design a different kind of sensor that have intrinsic compensations for this kind of "interferences". Finally it is also possible to make the sensor to work in a completely harsh environment (say with plenty of minerals/limescale or other particles suspended in the fluid) by changing the design of the stripes. Thank you for the question. May be I could make a video about this topic.
Thank you for the detailed answer. A reliable, maintenance free liquid sensor for a sump pump is what I am looking for. LevelGuard seems to make what I need but it is too expensive. They use a proprietary sensor which I would like to replicate. If you are interested in collaborating in this project please let me know how I can contact you. Thank you.
Well, let's talk, my email is in the about section here in my channel. When you have done please leave a comment here because once in a while the mail server lacks to send me notifications.
@@AccidentalScience Is it possible if the pressure sensor uses a barometric pressure sensor. Then the resulting pressure will be proportional to the water level(?)
Can i use this sensor for more than one meter and can u explain the way and the code for using this sensor with arduino thank u so mach for this great information
The sensor as shown in the video is "monolitic" that means that the probe (the two plates) and the circuit form a single sensor, so you can use more than one to connect to a single unit meant to work as a metering station. Since the sensor measure the capacitance between the plates there is almost no risk of interference if more than one sensor is plunged into the same tank, provided a minimum distance (say some cm) between the sensors. I do not love much Arduino because I prefer to code directly or to use my framework that makes multiple event programming much easier, so I am not able to direct you for specific code, however essentially you have to setup the ADC and by polling the ADC in the main loop to read out the analog signal coming from the sensor, and then use this data to do whatever you want in the program. With the NDC framework the things go similarly: you select the extension of the main loop and in the device loop file enter the code to poll the ADC, in this case it is easier as the ADC is already read out and you simply need to pick up the value using getIn(ANALOG0). For more information about the NSC framework look among my videos.
Hello, sorry for the delayed reply. As mentioned in the video it depends by the circuit. For a NAND oscillator the constant value typically is approx. 2.1 to 2.3 times R C or f = 0.47 * 1/(RC). I cannot provide more details because as said above it depends up how the circuit is designed, not to mention the fact that I am not a guru at math :)
Good question. The distance between the two plates is large enough to avoid capillarity, this was a thing that I taken into consideration when I designed this sensor. Accuracy however was not the primary goal, the instrument has shown an accuracy of circa 3mm over 250mm or ~1% (measured against the water column); the final stage of the circuit, the trigger that excite the output alarm, is far less accurate tough.
Thanks for answering. By the way, I was trying to take a similar approach. But my water container doesn't support insertion, so I am trying to use two aluminum plates to measure the water between。
NC : not connected. Closest value for 2k7 is 3k not 2k, not sure it would be a negligible difference if you change that resistors, I should check the circuit but I have not time to do it.
Yes it can. The tuning resistor at the input has to be recalibrated to the permittivity of vegetable oil (I don't know how much it is so unless you know just do some trial and error), and keep the plates apart enough to avoid the oil to clog the gap. Good luck.
@@AccidentalScience Wow! Thank you for your very quick reply! I am not an electronics buff, but I will attempt to build the equipment. It would help me if the circuit part was available as an add-on to an Arduino. I am building a computerised log stove that will also burn used veg oil using a special burner shpe and a drip feed. The aim is to make it so safe that it can be left running without attendance.
hi,great work i have one question i want to make a sensor like this for my own lab first i wanted to use arduino to read the data(by means of charge and discharge time) but i think its better to use the frequency method but i cant make the exact circuit can you send a technical drawing of the circuit? i want to connect the point you mentioned in video to arduino thank you very much
The point indicated in video is the analog output, the connection indicated is symbolic as this output have to be buffered with an opamp to preserve the high impedance. To read the frequency you need to connect at the output of the last gate (74HC132). It is mandatory to use exactly THAT component otherwise the circuit need to be retuned (and likely wouldn't work properly). You will read a pulse that vary in frequency, as you can see in the video.
@@David-xr4pd Why? I have indicated where you can connect an input of the Arduino to read the signal as frequency instead of analog: the output of the last gate or (for a reversed logic) the output of the previous gate (the pulse generator, see the schematic in the video at 7:20). Also looking at the waveform read from the scope you can infer the frequency and the pulse width. Also you have to choose an appropriate input in the Arduino board as it has to cope the frequency and width of the pulse.
@@AccidentalScience Hi, I have connected exactly to the analog read of the arduino but it only shows a constant value of 513 without any change in its value with variation in the capacitance. While, I have matching the circuit for a week, even the oscillator output frequency varies with change in the sensor water level but the analog read value doesnt changes. I made the spice model of the circuit till the output of the last 74hc132 but there also it is not a constant voltage while it is a frequency pulse.
Hi, Can we use Not gate instead of Nand Gate ST ? Why Resistance value is kept as specific 560 Ohm in Pulse generator section. what would be the impact of if we increase its value ?
Yes you can use a NOT gate 74HC14 even though it is a little bit slower than the NAND, reading from NXP datasheet. This however shouldn't undermine the circuit. Changing the resistor in the Pulse generator section would change the charging pulse width, you can of course change the resistor *and* change the capacitor so that the RC result is the same (approx. 6us). However this must follow the series resistor (82ohm in my design) to keep the discharing pulse width accordingly (approx. 1us). Because this would also be influenced by the input capacitance of the gate, and stray capacitances of the PCB, sticking with relatively small resistance values helps to avoid problems.
@@AccidentalScience Thank you for explaining. I am testing the circuit under different temperature conditions (from 25 Degree Cel. to 65 degree cel.). The value of 10 K pot at LM311 needs to change according to change in temperature. Please suggest how to compensate / modify the circuit for wide range of temp.
@@shiveshsainisavitsaini507 it would require another circuit because water's dielectric changes with temperature. Too complex to give an answer here, sorry.
@@shiveshsainisavitsaini507 you simply need another sensor that picks a fixed capacitive segment that must be always immersed. Then use the deviation from a known capacitance-temperature to compensate the deviation from the measured capacitance of the water sensor. This is the most accurate method of compensation. An alternative solution is a thermistor immersed into the water connected to an opamp with adjustable gain and floor to compensate the voltage from the sensor using a second opamp as an adder and then use its output to feed the comparator already in the sensor (the lm311 or for this application even a lm193 or half 393).
Why where theoretical and actual capacitance different? Water is often slightly electrically conductive... If water has considerably better conduction than insulation... the water can be approximated as a conductor... And arrangement modeled as two capacitors in series... With inner most plates being formed out water. And the dialectic being the insulation This can be shown by making one plate out of say copper, and the other out of aluminium. Applying a fixed DC current will anodise the aluminium plate, showing water will conduct until the anodising layer builds up. The anodised layer is a dielectric of thickness proportional to anodising voltage. Once built the anodised layer stays there. .
@@GeogyStephen Current? The probe is insulated. There is no current flowing between the plates, save instantaneous flow due to the moving charges back and forward because the oscillator, and that *alternating* current is negligible, I don't remember the exact values but in the order of microamps I guess.
I don't know the pinout of the Arduino board (also I think it depends by the board model) but essentially you have to follow the diagram shown in the video to connect to the analog input of the ESP8266, in place of the ESP you'll connect the analog input of your Arduino board (for those that have an ATmega328, any pin that goes to any PA0 to PA7 Mega's pins).
Hmmm, you can detect two different fluids with capacitive sensing, but I am not sure what do you mean for "a" capacitor and "different polarities", may you elaborate?
The schematic with chips should be visible in the video. I don't remember right now. Anyway in a month or so my website should be up and I will put there the schematics.
@@AccidentalScience i am doing it with a 555 astable multivibrator to convert the capacitance to frequency, then am using the LM331 to convert frequency to voltage which is sent to an arduino analog pin
@@kevin-kw8bv 555 is slow and have low impedance, which is far from ideal for this application. I don't say it won't work but it may even not work. I say this not to discourage you to try but just to give you a possible reason in the case you meet troubles.
That doesn't mean the current flows through the fuel though. Anyway common fuel gauges are made using a potentiometer that is submerged into the fuel inside the tank. Petrol doesn't catch fire as long as there is no oxygen or there is no enough current to make a spark.
Dear sir, the circuit diagram is too complex to understand and replicate. It would’ve been nice if you had explained the circuit diagram and how it works. Otherwise it’s impossible to follow what is going on. From an amateur diy hobbyist perspective.
Good question. I am not an expert in capacitors, so what I can infere is: 1. the non-submerged part is a simple air-polymer capacitor; 2. the submerged part is not an electrolytic capacitor since those caps form a layer made out of a metal oxide that represent the actual dielectric while the electrolyte is used as a mean for charges. In this case however we have just water that does not behave as electrolyte but as dielectric. This is my guess, but if someone can contribute with more information they are welcome.
thanks for your valuable reply. I tried this in my home but it doesn't work properly. I made capacitor perfectly and I measure that with the multimeter, it shows some capacitance values with the level of water respectively. please guide me with the circuit.
What does not work properly? Does it generates a square wave when dry? About the capacitor, what the multimeter read? (Sorry for the time taken to reply, I follow my channel as long as I have spare time, hope you understand).
The multimeter only shows 2.5v constantly on dry and wet condition. but my own made capacitor works with 4nf to 40nf respective with dry and wet conditions. 2.5v in the point of voltage to frequency converter. can you guide me for the further process?
If you have a constant 2.5V at the point of the V/f converter, then the output of 74HC132 is constantly low, this means that there are no pulses in output of the very same IC at the point of the oscillator, so you have to ckeck why the oscillator does not work. That was also why I asked if you see a square wave. Reasons why the oscillator does not oscillate can be: a short circuit somewhere in that part of the ciruit, or bad connected component, or wrong or broken components. Please keep in mind that the oscillator works because it is a fast Schmitt trigger IC, it will not work if you do not use an equivalent standard NAND gate or slow gate! An other possible reason for not having the pulses to the V/f converter is some wrong component (say a 82K instead of a 82 ohm) or the diode is reversed, or the capacitor is too large (say 100nF instead of 10nF). You have to check step by step at each stage that the signal is as intended, the video should give enough information about what kind of signal are expected.
EDIT: A portion of this video has been trimmed out because of an error in the shown formula of capacitance. I apologize. Thanks to "All in the Familiy" for having spotted the error. I have to say that I was not in comfort with that formula from start, at the point I regret not having cut out the whole part from the video in the first place. That happened because I had a mismatch with theoretical calculated values and the measured values and being in hurry to publish the video and being this a simple showcase video I didn't keep it on hold until that point was cleared. Sorry. More in the reply I've given to "All in the Familiy"'s comment below.
Set apart this, the remainder of the content in the video still hold and after almost four years of operation I have to say that this sensor worked flawlessy, so I think it remains an useful piece of information for DIY electronics.
I will try to do more about this topic. Thanks for sticking with me.
Remark. Because the dielectric constant of water changes with temperature this sensor provides a reliable measure as long as the water has a limited variation in temperature. In the case the water has variable temperature (say from 20C to 90C / 68-194F) a pressure sensor is better suited. (Pipe-pressure, the same that I'm gonna make for gauging the level of the sludge into a manhole.)
thank you, when you make the pipe-pressure sensor? Then I want to measure the diesel and petrol. can you help me for that!
You mentioned doing another sensor for your "sewage manhole". Did you ever get to this project? I am very interested in doing something similar but encasing the capacitor plates in molded plastic.
You also could add a temp sensor to the bottom of the capacitor sensor and then design a bias ckt to make the comparator circuit automatically adjust (compensate) for temperature variations....
I built similar tubes a few years ago using copper tape and pvp pipe. It was for an aquaponics setup and designed so that the exposed portion was the hollow center and could be easily cleaned with a pipe cleaner brush. I appreciate your coverage of the maths in this and your rapid pcb production technique is spot on!
Thank you.
A VERY INTERESTING FULL EXPLANATION THAT I HAVE NEVER FOUND ELSEWHERE, EVEN AT THE PROFESSIONAL AND COMMERCIAL SITES (AS THEY DO NO LIKE TO LET THE PUBLIC KNOW THE SECRETS OF THE INVOLVED PROCEDURE). THANK YOU VERY, VERY, MUCH.
Thank you sir.
10:00 probably the most underrated thing in this video is mad skill of turning dips into surface mount soic packages 👏👏👏👏
Thank you again
I designed a new type of level meter that fit my device.
It's totally different from yours, but I got a hint from your video.
Your new attempt has been good teacher for me.
Your advanced surface mounting technology was really very good. Thanks for this video. Greetings from Turkey.
Thanks.
I initially thought about making a sensor like this, but as a level transmitter for my arduino. I planned on using a 555 in a monostable configuration in order to read its actual capacitance value. Then after would have to calibrate it in order to correlate its level to its reading.
[Edit] I shared your video to GreatScott. He had a project recently to monitor a water tank level and used an ultrasound sensor in the end. But I wanted him to be aware of your creation :)
I am an Electronics Engineer I think it is very useful site can be useful for amature /beginners
You're the best, I understood all the concept, and I speak Spanish! Best regards from Colombia ☝🏻😎
Thanks bud. Greetings from Alpine Veneto.
This is exactly what I was looking for!!! Beautifully explained thank you! Subscribed!
Very knowledgable video. I learn so much from this.. Thank you so much for make this video.. you are awesome
I used a common electrical cable with two internal wires as a sensor, I melted the tip of the outer sheath to isolate the two cables, and as an electronics I used a common 555 whose oscillation frequency is read by arduino, works perfectly even as a sensor of soil moisture :)
Can you share circuit diagram and some pictures of your sensor. It will be helpful
@@syedumarjunaid5499 www.electroniq.net/555-timer-circuits/square-wave-generator-using-555-timer.html - the probe is a dual isolated cable connected instead of the 'frequency select' capacitors, consider that a cable about 80 cm long has about 80 pF.
Excellent video and thank you for the information. Did not know Inspector Jacques Clouseau was still around. You sound just like him.
Hahaha I didn't know about that, but I acted a little bit to make the video not too serious :D
Nice one.
Can you please answer some,
1. What is the maximum and minimum voltage at f/v converter ?
2. How this f/v converter constructed with just RRC passive components ?
3. What is the maximum output drive current of NAND gate ?
4. What is the Hysteresis voltage?
5. Does MCP3302 work in place of lm311 (from the data sheet which you have choosen to calculate input impedance of IC, “typical (or) maximum” values of offset voltage and bias current ) ?
Thanks again.
Awesome video. Thank You. That's all i need to build a pressure sensor. Cheers brom Brazil!
Seriously, this video made me to subscribe your channel ........ please keep up the good work.
Thank you.
Excellent video, i will be attempting this myself. Thanks!
Thank you.
9:30 great design sir 👍✨🎸
Great, practical, useful, comprehensive. Thank you.
Thank you mate.
great into. It is a typical anecdote for us experimenters....
Too good. I love your experiment and outcome.
Hi sir, can you explain me circuit how the capacitance is measured in the fuel tank the value need to measure in the microcontroller the capacitance value ranges from 50pf to 500pf.
hi , I liked the the capacitance sensor that you made. Would it be possible for you to share the electronic circuit diagram with specifications of components needed? I am looking for making similar type of application for fuel level gauging. If you have any suggestions you are welcome.
Hi, were you able to get the electronic circuit diagram with the specifications of the components? I am a bit stuck with a project similar to this and it would be much appreciated. Thanks.
Grande...
Ottimo lavoro complimenti
Grazie.
Very nice, very nice video with a lot of useful info and also the necessary theory behind. I wish you would have spent a little longer on the math and on the circuit components.
This said, why not make the circuit with a proper CAD and use the transfer method? Also, I think disconnecting AC whit a relay when the tank is full is a poor failover method. Why not create some sort of passive overflow system that safely discards the excess water? It could be a series of pipes that passively discharge extra water or simply positioning the tank on a drain.
Last question: will water collected from a condenser be slightly acidic? Is aluminum a good material for this or would something like a stainless steel be better?
Sorry for not covering deeper the math involved, but that would be beyond the scope of this channel, also I am not a matematician ;)
The passive method for discharging the excess of water would it were possible wouldn't had rendered this circuit necessary. Also, since the water is produced only if the AC is on, even in the case of power outage this would not cause any problems making the circuit intrinsecally safe.
I don't know if the water become slightly acidic, I did never tested it. The aluminum of the plates is covered with resin so there is no direct contact of the aluminum plates with the collected water.
Okay, it works, but why did you go to the trouble of building what is essentially a capacitive liquid-level sensor capable of measuring the full range of liquid levels in the tube, when all you needed was a top-level discrete switch to turn on a warning? My experience with these circuits was as an aircraft mechanic. As I remember, the capacitive liquid level sensor was developed in Britain late in World War II to fulfill the need for an aircraft fuel-level sensing system that was accurate in flight and had no moving parts. It is still widely used today, although ultrasonic level sensors are now used in new designs. One of the most reliable and trouble-free systems I encountered was used in Convair airliners equipped with twin turboprop engines, the Convair 580. Based on a vacuum-tube amplifier, it used a simple capacitance-resistive bridge circuit to drive a servo motor inside the instrument that positioned the indicating needle at the scale position for the sensed fuel quantity, and also drove a potentiometer wiper that rebalanced the bridge circuit for every indicator movement. The capacitance input was provided by the total of all the in-tank, tubular capacitors connected in parallel, plus a separate reference capacitance signal from a submerged compensator. It worked amazingly well, and better than some of the digital capacitance circuits found on later aircraft models.
That sensor with pot feedback based on a vacuum tube, you nicely described is interesting indeed. I chose the capacitive sensor for three reasons: it is easier to adjust the alarm level, it's robustness because it has no moving parts and it cannot be damaged when it is pulled off the tank (when it has to be emptied), and to have fun 😊
Nice video inspired me to do it. I tried to make one. And modified the probe.
1. Probe capacitance varying 160pf to 205pf, so i calibrated RC ckt to tune with same, modified R 150kohm to 42kohm.
2. From f/v output I wanted to get 4 different outputs, so placed 4 individual comparators to glow 4 LEDs for 4 stages of water level 10%, 35%, 65% and 95%. And the comparators output given to RF encoder (HT12E) then 433mhz transmitter.
Please do let me know if my considerations would go wrong. Thanks!.
Good job. Thank you for letting me know. I'm on mobile now, I'll integrate this comment later.
I think your design should work correctly, in particular if the comparators have a decently high input impedence, which usually it is the case in particular if you use the LM311.
Of course you can change with other comparators, maybe a package that contains four.
Please note that the schematic shown in the video has an error: I forgot to draw a line from pin1 to pin4 of the LM311. The correct schematic is visible on my website.
Greetings from the Alps.
Very good presentation! I'll keep your method in mind for future projects. I'd also be interested in knowing about your pressure sensor project.
Look at the "scale with a can" video, the principle is pretty much the same. EDIT: The video is about weighing with a tin can, here it is the link: th-cam.com/video/meluEWZpsXY/w-d-xo.html
Very efficient but it's a too complex circuit for me. I will use an ultrasonic sensor
You can replace a lot of the circuitry with just a 555 timer, namedly the first 2 stages.
I'm thinking of making another one of these using a slightly different principle. The last one I made was really difficult to set up (was from EDN online magazine and cut down a little for a water level monitor for a boat, copper tubing around a plastic pipe with an insulated electrode down the centre, making a co-axial capacitor). I was thinking using a fixed frequency, and then having an output that tells you when the capacitor is charging or not 1=charging, 0=not charging to give a PWM output that is then filtered and can give a wider range output without too much amplification. Capacitor discharged quickly at rising edge of frequency generator, maximum frequency needed determined by capacitor and charging resistor time constant, at the time constant's highest value (variable charging resistor to adjust for maximum).
Just out of curiosity, may you provide the link to that article? The coaxial version of the sensor can be seen in my video at 4:21 .
I don't get how your solution works: since the capacitor is always charging from the raising edge to the falling one, and alwasy discharging from the falling edge to the raising one, if you look at the current (to see the charges' flow) then this current will lasts for the full positive cycle (and the full negative cycle on the other way). I am not saying it does not work but likely I missed something in your explanation, perhaps is your idea to cap the charge cycle at a given voltage by setting a threshold so that you can look at how long the capacitor takes to charge (up to that given voltage), switch to discharge, and hence generating a sort of PWM? That would be equivalent to the "method 1" I described in the video (at 2:29).
Anyway, good luck with your project, also let us know about the results. Thanks for having commented.
Cheers.
@@AccidentalScience Sure, the EDN article is this one: www.edn.com/design/test-and-measurement/4423158/Precision-capacitance-meter
I used an LM3914 bar graph to show water level. I had trouble with getting the right range with the EDN circuit - I think it is designed for much lower capacitance and smaller variation.
I guess I wasn't too clear, I was going to let the capacitor charge through a resistor and discharge it at regular intervals with a small pulse, like in your weight sensor video, but without a constant current and just a resistor (that you can vary too). The idea was when the capacitor is charging (below threshold) a 1 is output and 0 above threshold to make a PWM signal controlled by the discharging repetition frequency. I'm not sure if it's functionally equivalent to yours that uses PDM (pulse density modulation).
I've since found a circuit that would work in a similar manner to what I was thinking, although I've already designed my own version (with it's less than ideal behavior, but can compensate for that). I'm also trying a sensor with the elements wound in a spiral around a small diameter tube, hoping to get a measurable capacitance change between them as covered by the fluid (similar construction to an open wire line as used for RF). As the fluid is only on the outside (the tube can be blocked or a rod used) then it should be ok for measuring fluid with 'sludge'. PWM type sensor - here used for a capacitive humidity sensor: www.digikey.com/en/articles/techzone/2011/oct/humidity-sensors-and-signal-conditioning-choices
Sorry for the late reply, I have noticed today that your last comments were kept held by YT for review ...it's a plague common to other commentators, that I have to clear manually. About the circuit. I've seen the one on E D N, it is based on a fixed frequency and the capacitance is read by comparing a fixed time charge (the branch R14 and CA1D in their schematic) and the variable time charge (due to the test capacitance C?) on the branch R11 CA1C that is conditioned by the availability of the current that split at the node R8-R11 through the BAV99 "sucked" by the test capacitance.
Thus, at the input of the opamp would appear kind of a race between the fixed time charge and the variable one from the two branches. If I did not missed something, because CA1C is quite large (10nF) the test cap should be in a comparable range otherwise the difference would need a large amplification, they set to AV=10. Also, the two CA1x capacitors should be matched for the best results, but that should be secondary for the functioning of the circuit.
It seems a complex circuit to me, and frankly I don't like it.
One thing you have to keep in mind is the design of the layout or the PCB, as the devil hide in details. Beware of every stray capacitance, and leakage from the surface of the circuit because the common (cheap) FR4 adsorb non trivial amount of humidity from the air (and after the etching process), so after cleaning the PCB with (warm) isopropil or ethilic 99,9% alchool (which can grab up to about 30% of water in it), it should be backed at 60C for 30' before covering it (still warm) with a good quality conformal coating.
Also, the spiral idea do not convince me, it could become an inductor that may cause weird behaviours, pick up RF or LF EM ...and cause headakes :) To have a good capacitance simply reduce the distance between the plates.
I've been very busy the last three weeks, but I will upload an other video on capacitive sensing.
Cheers.
Awesome video, very clear, thanks, keep it up.
Thank you.
Great video. Thanks for the upload!
I want use this project for measure vegetable oil level on a bottle. I have a question: if the oil is dielectric, Can I use the electrodes without epoxy resin? I'm thinking make the electrodes with stainless steel.
Epoxy covering (actually I used polyester) is not aimed to protect against corrosion but to provide insulation. Look for a food grade covering or just use bare metal as oil is not conductive. However oil has a quite different dielectric coefficient and both the plates and the circuit would require a redesign for such an application.
sorry folks that is incorrect, the formula for capacitors with stacked dielectric materials is C=Eo*A / ((d1/Er) + (d2/Ew) + (d3/Er))), Eo=8.85E-12F/m, Ew=10, Er=3, A=area,(lxw), d1=thickness resin plate1=1mm, d2=thickness air/water=3mm, d3=thickness resin plate 2=1mm,. With these values the calculated capacitance with air is less then 1pF namely 0.8pF, with water the capacity is 3.1Pf.
First off, thanks for your comment. You somewhat put the finger where the spot hurts, since that is the formula I used at first when I tested this circuit and probe but then I used an alternative formula because it didn't fit the experimental results. Well, you got the wrong result because you put 1 mm thickness for the resin layer when it is more like .1 mm (not sure about the exact thickness), but that does detract nothing to your point: that should be the correct formula and it is basically where the formula for series capacitors comes. Because intuitively a stack of dielectrics sould behave as a series of capacitors. But as I've said the results didn't fit experimentally. My first theoretical results (using the formula you mentioned) were: 9.9pF dry and 219pF full of water (apart a small portion); while the actual measurements were 34pF dry (measured with a capacimeter) and 2.34nF in water (calculated with the circuit). They were too off. At first I thought to have some leakage, a resistor in parallel or maybe a series resistor caused by a bad junction. But after thoroughly checking with the multimeter I didn't find any significative resistance that could have changed the response of the circuit and the resulting capacitance of the probe *that much*. So I searched online and found a paper where they said something related to the density of charges, giving a formula where the relative dielectric constant (permittivity) of multiple layers was simply multiplied to the permittivity in vacuum. Unfortunately I didn't keep a reference. However that formula gave results that was close to the experimental data so I trusted that formula. Though I never felt in comfort with that and I should have skipped the theoretical part in the video. This is what I've done now: I trimmed off that part!
That said, still something do not add up properly because the experimental data were way too off from the theoretical data.
As soon as I'll have a bit of time I will try to go deeper into this thing.
I can't pin your comment, but definitely I will include it in some way into my already pinned comment.
Thank you again.
Hi , excellent video . I got two major issues when I made cylindrical capacitive diesel sensor. Please can some one answer following queries. :
1, it suddenly changes reading ,mostly to 100% or even more than that. It may be because of " charge accumulation on central rod ". Mind it I did not coat with resin/acrylic.
2, it changes reading with temperature . e.g. increases in afternoon and decrease in night.
3, wire b/w probe n circuit also causes reading drift.
Missing coating makes the probe resistive. Diesel is not a conductor but it could mess up the capacitive vs. Resistive ratio. Not sure though, it would require a thorough investigation. Wires cause parasite impedance so the probe must be directly connected to the circuit: the two parts must be seen as one.
You could calibrate an existing sending unit from a fuel tank. Attached to a gauge it would give an accurate reading of the fluid level. Or, just a float attached to a switch that triggers a warning light when the fluid level in the tank is too high or too low.
Very nicely explained.
What is the effect of the thickness of the Aluminum strips?
Would a thin copper foil work (instead of the Aluminum strips)?
Thank you.
The thickness of the aluminium strips is irrilevant, and as long as it is a good conductor the material is irrilevant too, soo even foil copper can be used.
My experience is that when titanium is used for the ungrounded probe... It's size, and the thickness of any titanium oxide/anodised layer has big effects on the capacitance.
My understanding of this while "tap water" is a poor conductor... Compared to titanium oxide it's not that bad a conductor... So in effect the capacitor is formed between inside of plate, and water. Ie across the titanium oxide layer which 1) is very thin 2)has a fairly high dielectric constant.
Hi,
I really like your video and appreciate your efforts. I have some questions and i hope you would like to answer these.
1. Can I make 4.5 feet long capacitance probe?
2. How long wire from probe terminal to circuit is recommended.
3. What is accuracy if water changes from clear water to dirty (contains mud or other particles) water, does it give same reading?
4. Is it mandatory to use Acrylic Color on top of Crystal resin coating?
5. Can we use any other material than Aluminum for probes?
6. If Arduino have ADS, will these capacitance reading be readable in Arduino?
Regards,
1. Yes, requires tuning.
2. As short as possible. This is one of the trickiest part when designing these kind of circuits.
3. I don't know, I had to make a test but never had the time to do it.
4. No. But insulation and the dielectric of the insulator is of paramount importance.
5. Yes, provided it is a good conductor of electricity.
6. What ADS stands for? If it was meant to refer to ADC, as long as the ADC input is decoupled from the primary stage of the circuit (after the F to V conversion) it does not pose any problem. However for a professional design I would not use an ADC, instead I would directly gauge the frequency.
Cheers.
Hi, I just saw your video and it is great. You are really an engineer with hands on. Is it possible to speak to you to discuss a problem that I am having at the moment? Thank you
Leave your question here, I'll see what I can do. Instead, if you need a consultancy drop an email with a detailed description of your problem, and I'll give you a quote. Email address can be found in the about section of this channel.
I love the detailed math!
Very informative video! I have a 20gal chlorine tank used for well water treatment and would like to use an ESP 8266 to send an MQTT message to my node-red/raspberry Pi system so as to flash one of the lights in my living room which is connected to a Sonoff Basic running Tasmota.
Flashing the ESP 8266-01 or some other variant to send a MQTT message at some level {approx 20% full} might tax my limited programming skills however.
What a wonderful video!
Awesome Video 👍
wow awesome! What dielectric used in between the two plates? Can it be just an air?
Air.
I want to replicate this project. Can you please answer few of my queries? It would also help other, those who visit this video!
1. Can I attach the probe separately with a wire and keep the circuit at a distant?
2. Which wire should I use? Twisted pair or Single core shielded?
3. Can I use this same circuit for moisture count in soil?
Answers:
1. No. The closer the probe is to the circuit the better the sensor will work. I attached the probe with two short leads for testing purposes only, but those leads was source of noise and troubles that I compensated in some way ...you know, experience helps here :)
2. see 1.
3. No, the circuit and the probe have to be redesigned for that specific purpose.
Please list the two coatings you used to seal the aluminium strips. Thanks!
Acrylic and polyesther resins.
very helpful thanks!
Very good! I would just like to ask which integrated circuit used 8 legs that appears on the left side of the scheme in the video, when mounted the board could not identify, is a 555 oscillator? grateful for the help
It is a 393 dual comparator (one of the two is used only). It is mounted in the etched board, not on the breadboard (where you can see an other opamp because I've been simultaneously testing an other circuit not discussed in the video).
great video
Hi,
Would you like to share the PCB layout for diy purpose? And the Arduino program too. Thanks in advance.
thanks very much..very good video
Glad you liked it!
In your circuit diagram, what part at the most front end called to generate the quare wave pulses ?
It's the first NAND gate. Looking at the configuration you can guess that it's used as an inverting gate. Being a Schmitt trigger is fundamental to operating properly.
Happy new year!
Can you change this circuit for 4ft and 5ft lenght tank and give me components list and PCB layout because I don't have oscilloscope that's why I can not do this calculation.
I will waiting for your prompt reply.
Best regards
Umar
Very gud
OK but I'm trying to do this outdoors for rainwater containers. Complex circuitry breaks down due to humidity. I tried to enclose my circuits in jars and cover the jars from the rain but humidity still gets in. Only simple conductive circuits seem to work reliably -- nothing with timers or oscillators has worked for me yet. Any chance you could make one of these with components that don't care about moisture? Thanks!
This is already meant to be moisture proof. You have to spray the circuit with (insulating) lacquer. Or put it in a box and fill the box with epoxy.
Conductive circuits suffer of corrosion and moisture find a way to get in easier than with capacitive that can be completely insulated.
@@AccidentalScience Thank you for your reply. My question really was whether some components have better moisture resistance than others and whether they can be used preferentially to build a circuit such as this. I think the most effective way to keep moisture out of the circuit is to submerge it in a non-conducting liquid that's heavier than water. That way the circuit stays dry even if there are small cracks in the container. But these are just brute force approaches. I was wondering if there was something more elegant available.
@@SashaXXY all electronic components have more or less the same problem with moisture. The only standard technique used to prevent this is lacquer or epoxy coating aka conformal coating.
@@AccidentalScience Ok, thank you for the information.
thank you for sharing! Is it possibile to develop an homemade sensor to measure the capacity of the soil?
...the capacity of the soil? Do you mean the dielectric or actual capacitance?
What would be the purpose of such a sensor?
Very neat!
Great
Excellent!
Thank you.
If the level sensor will connected to the pcb using about 1/2 meter thin coaxial wiring could this cause problems to the high impedance input of oscillator's schmitt trigger gate?
Sure!
Congratulations, could you build another one to measure the fuel level?
Sure but fuels are trickier and a different type of circuit is required. At the moment I have no time to build one.
Hi...
¿How can I find information about that oscillator and pulse generator?
Now I know I can simulate it in a computer program or built mine in house but in both cases if I want to model it mathematically I would have to measure...
In resume: could somebody tell me where can I see theory about oscillator and pulse generator?
While I don't have a specific answer right now, a good starting point is to understand the theory of operation. Some more information is available on the relative page on my website, but there is literature (I think it should be available online) that covers oscillators and other signal manipulation through logic gates. I have an old book titled "Elettronica Digitale" (don't remember the publisher though and AFAIK it's only in Italian) that provides good coverage on the topic. Maybe searching for a similar title you may find something.
@@AccidentalScience
Yeah... Thnx for your answer...
Another question I have is:
Did you review your book for design that circuit?
I can understand how the oscillator part works by seeing connections in NAND, resistors and capacitor because it is the first part but I can understand pulse generator and the other parts by simple inspection 😓
thanks for your video.. but can i get your manual book or reference ?
Soon I'll have my website updated, there you'll find several information.
@@AccidentalScience ya.. i'm waiting for it
You mentioned doing another sensor for your "sewage manhole". Did you ever get to this project? I am very interested in doing something similar but encasing the capacitor plates in molded plastic.
Not yet!
Adoro este video :D. Saludos!.
Good Job. Congrats.
Thank you.
exelente video!
Thank you
Wanted to ask if the volume of the water will have any effect on the capacitance? What I mean is if you use a bigger tank with the same height, will it needs to be calibrated
No.
what if want to make signal when water leve is below some point, like almost empty
Just adjust the voltage at the comparator stage.
Can anyone please explain why you can't just measure the change in resistance between two closely separated metal plates? It seems that the circuitry (and maths) involved would be much simpler!
Obviously, you wouldn't want it running all the time because you'd corrode the electrodes, but using something like an Arduino you only need to power it for a few milliseconds when you're actually measuring.
Obviously you can. Provided you have tap water, or salt water. With water coming from HVAC which is far less conductive things become a little bit complex. More, water changes its conductivity because variations in dissolved minerals way more than its dielectric constant.
That said I made a water level sensor using stainless steel rods of various length and a simple row of LEDs that worked nicely for years. In a non-linear fashion like that stainless steel rods are way more convenient.
Maybe I will make a video about this point. Thanks for commenting.
Did you ever make the next video about measuring unpleasant stuff? I can't seem to find it in your list. Thanks.
Sorry, never finished. Too many things to do :)
@@AccidentalScience that's life! Thanks
Hi, great project do you think it can be used for an House oil heating tank to measure the oil level? It’s a 2500 litre tank.
As it is no, it requires the redesign of the probe as the oil has quite a different dielectric.
British Humor 👏🏻👏🏻👏🏻 🇨🇱
like the buzzer :D
Hi,
Great video🎥
Can you tell me in immersed condition how did you measure in MM or CM the relative dielectric const water -80
Sorry, I'm not sure about your question: what do you mean for "water -80" ?
@@AccidentalScience
In that video you were told like immersed condition the relative dielectric const. Water : 80 .... What is that 80
@@balajiramesh6814 oh gotcha. Water have a typical dielectric constant (AKA relative permittivity) of 80. This is a dimensionless number being a ratio. See here for more: en.wikipedia.org/wiki/Relative_permittivity
The relative permittivity (or dielectric constant) does not change by varying the immersion of the sensor's probe, what it changes is the ratio between air and water between the plates, which makes the capacitance to change because the dielectric constant of air is ~1.
The relation between depth of immersion and capacitance can be calculated by the ratio of the segment that is not immersed using the dielectric cons. of air, and the segment that is immersed using the dielectric const. of water. Through this means you can make a relationship between depth in mm and capacitance, or frequency (after the oscillator stage) or voltage (after the VF converter stage).
Could you clarify how the liquid levels can be output to a smart phone. Did you say ESB 2286?
ESP 8266. You obviously need to make a program in order to connect and see the output on a smart phone. The program has to read the analog signal (at the point indicated in the schematic) and generate some kind of user interface to display the information, i.e. a web page, and operate as a wifi access point.
What about mineral buildup on the 2 strips and the eventual blocking of the pipe?
Your question is a good point and to try to give an answer we should wonder what are the reasons limescale/minerals build up:
1) it could be caused by sedimentation, in this case there is no problem because the gravity will attract the particles down, outside the pipe;
2) it could be caused by chemical concentration, at the point the particles stick to the surfaces, this typically happens when high temperatures are involved, and it is usual in resistors used to heat up boilers, however the stripes are not heated (and with no high temperature spots the particles stay diffused). However after long time some particles will stick anyway but far from the point to cause the pipe be chocked;
3) it could be caused by electrical attraction: if the plates (the stripes) were connected to a continuous electric field that could attract polar minerals eventually. However here not only the electric field is very small but it is also alternating so there is no risk to attract polar particles.
About the thin layer of minerals/limescale that may build up in long times, mentioned at the point 2 above, one thing that we may wonder is if that layer may modify the dielectric properties of the sensor, thus shifting the returned measured value. I have no answer for this last point because I think the only way to give an answer is to try experimentally. However since the sensor as shown in the video is quite simple I think that for the purpose of its use it should be fine anyway, it is however possible to design a different kind of sensor that have intrinsic compensations for this kind of "interferences". Finally it is also possible to make the sensor to work in a completely harsh environment (say with plenty of minerals/limescale or other particles suspended in the fluid) by changing the design of the stripes.
Thank you for the question. May be I could make a video about this topic.
Thank you for the detailed answer.
A reliable, maintenance free liquid sensor for a sump pump is what I am looking for. LevelGuard seems to make what I need but it is too expensive.
They use a proprietary sensor which I would like to replicate.
If you are interested in collaborating in this project please let me know how I can contact you.
Thank you.
Well, let's talk, my email is in the about section here in my channel. When you have done please leave a comment here because once in a while the mail server lacks to send me notifications.
Contact made :)
Loving your humor.... smart guy👍😉
Hi, did you manage to build the pressure capacitive sensor? I can't seem to find that upload.
I did not made it yet. I had some troubles in making the pressure discs. I will post a video when it will be ready.
There are pressure sensors for water for around 10€ I have one but will build a capacitive one because I like them better
@@AccidentalScience Is it possible if the pressure sensor uses a barometric pressure sensor. Then the resulting pressure will be proportional to the water level(?)
@@rainbowforest09 no. Barometric sensors are made for air pressure, water pressure is way higher.
Can i use this sensor for more than one meter and can u explain the way and the code for using this sensor with arduino thank u so mach for this great information
The sensor as shown in the video is "monolitic" that means that the probe (the two plates) and the circuit form a single sensor, so you can use more than one to connect to a single unit meant to work as a metering station.
Since the sensor measure the capacitance between the plates there is almost no risk of interference if more than one sensor is plunged into the same tank, provided a minimum distance (say some cm) between the sensors.
I do not love much Arduino because I prefer to code directly or to use my framework that makes multiple event programming much easier, so I am not able to direct you for specific code, however essentially you have to setup the ADC and by polling the ADC in the main loop to read out the analog signal coming from the sensor, and then use this data to do whatever you want in the program.
With the NDC framework the things go similarly: you select the extension of the main loop and in the device loop file enter the code to poll the ADC, in this case it is easier as the ADC is already read out and you simply need to pick up the value using getIn(ANALOG0). For more information about the NSC framework look among my videos.
Accidental Science thank u a gain
Hello, can you talk about how you calculated the *K value?
Hello, sorry for the delayed reply. As mentioned in the video it depends by the circuit. For a NAND oscillator the constant value typically is approx. 2.1 to 2.3 times R C or f = 0.47 * 1/(RC). I cannot provide more details because as said above it depends up how the circuit is designed, not to mention the fact that I am not a guru at math :)
Considering capillarity, would it be inaccurate ?
Good question. The distance between the two plates is large enough to avoid capillarity, this was a thing that I taken into consideration when I designed this sensor. Accuracy however was not the primary goal, the instrument has shown an accuracy of circa 3mm over 250mm or ~1% (measured against the water column); the final stage of the circuit, the trigger that excite the output alarm, is far less accurate tough.
Thanks for answering. By the way, I was trying to take a similar approach. But my water container doesn't support insertion, so I am trying to use two aluminum plates to measure the water between。
what does NC meaning on the lower left side of the circuit @2:13 time frame,and is this 2k7 = 2k ohms? thanks
NC : not connected. Closest value for 2k7 is 3k not 2k, not sure it would be a negligible difference if you change that resistors, I should check the circuit but I have not time to do it.
Can this capacitor system work to test level of vegetable oil in a tank or funnel?
Yes it can. The tuning resistor at the input has to be recalibrated to the permittivity of vegetable oil (I don't know how much it is so unless you know just do some trial and error), and keep the plates apart enough to avoid the oil to clog the gap. Good luck.
@@AccidentalScience Wow! Thank you for your very quick reply!
I am not an electronics buff, but I will attempt to build the equipment. It would help me if the circuit part was available as an add-on to an Arduino. I am building a computerised log stove that will also burn used veg oil using a special burner shpe and a drip feed. The aim is to make it so safe that it can be left running without attendance.
@@biopoweruk cool. Send me a picture of the finished product, I will post it on my website.
You are a genious :)
Haha :) thank you mate.
hi,great work
i have one question
i want to make a sensor like this for my own lab
first i wanted to use arduino to read the data(by means of charge and discharge time)
but i think its better to use the frequency method
but i cant make the exact circuit
can you send a technical drawing of the circuit?
i want to connect the point you mentioned in video to arduino
thank you very much
The point indicated in video is the analog output, the connection indicated is symbolic as this output have to be buffered with an opamp to preserve the high impedance. To read the frequency you need to connect at the output of the last gate (74HC132). It is mandatory to use exactly THAT component otherwise the circuit need to be retuned (and likely wouldn't work properly). You will read a pulse that vary in frequency, as you can see in the video.
@@AccidentalScience so i cant connect any point of this board directly to arduino?
i have to find a way to read this frequency by arduino?
@@David-xr4pd Why? I have indicated where you can connect an input of the Arduino to read the signal as frequency instead of analog: the output of the last gate or (for a reversed logic) the output of the previous gate (the pulse generator, see the schematic in the video at 7:20). Also looking at the waveform read from the scope you can infer the frequency and the pulse width.
Also you have to choose an appropriate input in the Arduino board as it has to cope the frequency and width of the pulse.
@@AccidentalScience thank you very much sir
@@AccidentalScience Hi, I have connected exactly to the analog read of the arduino but it only shows a constant value of 513 without any change in its value with variation in the capacitance. While, I have matching the circuit for a week, even the oscillator output frequency varies with change in the sensor water level but the analog read value doesnt changes. I made the spice model of the circuit till the output of the last 74hc132 but there also it is not a constant voltage while it is a frequency pulse.
Hi, Can we use Not gate instead of Nand Gate ST ?
Why Resistance value is kept as specific 560 Ohm in Pulse generator section. what would be the impact of if we increase its value ?
Yes you can use a NOT gate 74HC14 even though it is a little bit slower than the NAND, reading from NXP datasheet. This however shouldn't undermine the circuit. Changing the resistor in the Pulse generator section would change the charging pulse width, you can of course change the resistor *and* change the capacitor so that the RC result is the same (approx. 6us). However this must follow the series resistor (82ohm in my design) to keep the discharing pulse width accordingly (approx. 1us). Because this would also be influenced by the input capacitance of the gate, and stray capacitances of the PCB, sticking with relatively small resistance values helps to avoid problems.
@@AccidentalScience Thank you for explaining.
I am testing the circuit under different temperature conditions (from 25 Degree Cel. to 65 degree cel.). The value of 10 K pot at LM311 needs to change according to change in temperature. Please suggest how to compensate / modify the circuit for wide range of temp.
@@shiveshsainisavitsaini507 it would require another circuit because water's dielectric changes with temperature. Too complex to give an answer here, sorry.
@@AccidentalScience No Issue, I will share if i got some solution. Thanks again.
@@shiveshsainisavitsaini507 you simply need another sensor that picks a fixed capacitive segment that must be always immersed. Then use the deviation from a known capacitance-temperature to compensate the deviation from the measured capacitance of the water sensor. This is the most accurate method of compensation. An alternative solution is a thermistor immersed into the water connected to an opamp with adjustable gain and floor to compensate the voltage from the sensor using a second opamp as an adder and then use its output to feed the comparator already in the sensor (the lm311 or for this application even a lm193 or half 393).
Why where theoretical and actual capacitance different?
Water is often slightly electrically conductive... If water has considerably better conduction than insulation... the water can be approximated as a conductor... And arrangement modeled as two capacitors in series... With inner most plates being formed out water. And the dialectic being the insulation
This can be shown by making one plate out of say copper, and the other out of aluminium. Applying a fixed DC current will anodise the aluminium plate, showing water will conduct until the anodising layer builds up. The anodised layer is a dielectric of thickness proportional to anodising voltage. Once built the anodised layer stays there.
.
The plates were insulated, look at the beginning of the video, they have been covered with polyester. This layer could explain the difference.
can your sensor detect the level of fluids with varying dielectric constants , from high to low.. this is for detection of human blood levels
No, it can't.
@@AccidentalScience is it possible for you to tell me, how much current would be detected if the probe reaches the surface
@@GeogyStephen Current? The probe is insulated. There is no current flowing between the plates, save instantaneous flow due to the moving charges back and forward because the oscillator, and that *alternating* current is negligible, I don't remember the exact values but in the order of microamps I guess.
Happy birthday to you.
can you pls send the circuit connection for using with arduino for fuel sensing
I don't know the pinout of the Arduino board (also I think it depends by the board model) but essentially you have to follow the diagram shown in the video to connect to the analog input of the ESP8266, in place of the ESP you'll connect the analog input of your Arduino board (for those that have an ATmega328, any pin that goes to any PA0 to PA7 Mega's pins).
i followed the same circuit but the output is 3.5to 3.9 V. I used DM74LS132N IC.
do i have to insulate the plates ?
Yes you have to. Plates must be covered with a resin with low water adsorption, or the circuit won't work.
Also, DM74LS132N is not suitable for this application, you need the 74HC132 (not the HCT version because have different output voltage).
Can we use a capacitor to determine two different fluids with different polarities?
Hmmm, you can detect two different fluids with capacitive sensing, but I am not sure what do you mean for "a" capacitor and "different polarities", may you elaborate?
@@AccidentalScience I want to design a capacitor which can detect water from fuel . (Water- Polar, Fuel- Non polar)
@@DinuBoy gotcha. Cool problem. Polar or non polar fluids have different behavior as dielectric so it can be certainly done.
which chips did you use for the circuit? can you also share the circuit with us? id love to build the same sensor
The schematic with chips should be visible in the video. I don't remember right now. Anyway in a month or so my website should be up and I will put there the schematics.
@@AccidentalScience i am doing it with a 555 astable multivibrator to convert the capacitance to frequency, then am using the LM331 to convert frequency to voltage which is sent to an arduino analog pin
@@kevin-kw8bv 555 is slow and have low impedance, which is far from ideal for this application. I don't say it won't work but it may even not work. I say this not to discourage you to try but just to give you a possible reason in the case you meet troubles.
@@AccidentalScience yes you are right, the 555 multivibrator is not working correctly, what other chip would you advice me to use?
@@kevin-kw8bv look at the schematic in the video.
does that sensor also apply to oil??
Retuning the circuit, yes.
because in the rod is flowing electricity this method is not suitable for fuel gas, for a moto for example
That doesn't mean the current flows through the fuel though. Anyway common fuel gauges are made using a potentiometer that is submerged into the fuel inside the tank. Petrol doesn't catch fire as long as there is no oxygen or there is no enough current to make a spark.
10/10
Reference link for schematic diagram pls
Go to the website: accidentalscience.com
Dear sir, the circuit diagram is too complex to understand and replicate. It would’ve been nice if you had explained the circuit diagram and how it works. Otherwise it’s impossible to follow what is going on. From an amateur diy hobbyist perspective.
Please the clear circuit
what type of capacitor is this..? like electrolytic capacitor
Good question. I am not an expert in capacitors, so what I can infere is: 1. the non-submerged part is a simple air-polymer capacitor; 2. the submerged part is not an electrolytic capacitor since those caps form a layer made out of a metal oxide that represent the actual dielectric while the electrolyte is used as a mean for charges. In this case however we have just water that does not behave as electrolyte but as dielectric.
This is my guess, but if someone can contribute with more information they are welcome.
thanks for your valuable reply. I tried this in my home but it doesn't work properly. I made capacitor perfectly and I measure that with the multimeter, it shows some capacitance values with the level of water respectively. please guide me with the circuit.
What does not work properly? Does it generates a square wave when dry? About the capacitor, what the multimeter read?
(Sorry for the time taken to reply, I follow my channel as long as I have spare time, hope you understand).
The multimeter only shows 2.5v constantly on dry and wet condition. but my own made capacitor works with 4nf to 40nf respective with dry and wet conditions. 2.5v in the point of voltage to frequency converter. can you guide me for the further process?
If you have a constant 2.5V at the point of the V/f converter, then the output of 74HC132 is constantly low, this means that there are no pulses in output of the very same IC at the point of the oscillator, so you have to ckeck why the oscillator does not work. That was also why I asked if you see a square wave.
Reasons why the oscillator does not oscillate can be: a short circuit somewhere in that part of the ciruit, or bad connected component, or wrong or broken components. Please keep in mind that the oscillator works because it is a fast Schmitt trigger IC, it will not work if you do not use an equivalent standard NAND gate or slow gate!
An other possible reason for not having the pulses to the V/f converter is some wrong component (say a 82K instead of a 82 ohm) or the diode is reversed, or the capacitor is too large (say 100nF instead of 10nF).
You have to check step by step at each stage that the signal is as intended, the video should give enough information about what kind of signal are expected.
👍👍👍👍👍👍