One advantage the bubble method is that biofilms and other dirt is disturbed sufficiently so it is very reliable even with dirty fluids. It is used in sewage processing plants. Using inert gases like nitrogen it is even suitable for highly flammable or corrosive fluids. To get more repeatable results it is beneficial to cut the air tube at an angle on the bottom. It makes the bubbles more uniform.
Thanks for this contribution! Using readymade blood pressure units is a good idea! Also the HX711 module. I am still using a BME280 to measure the inventory of my oil tank since 2016. Error is less than 10 l at an 10 000 l tank, better 0.1%. Not bad for a crazy "proof of concept" without optimization. I have recorded daily consumption data since 2017 which help me now to decide about future energy saving measures.
Cool. My results were not as accurate as yours. Good to know that the BME280 is a good sensor for this purpose. Probably not as easy to connect, though.
@@AndreasSpiess The BM280 has a limited pressure range. So, it is not as versatile as dedicated differential pressure sensors. But it fits for my specific application. You can find some posts incliding pictures on the German homematic-forum with the unique search word "Tomatensaucenglas" - which indeed is my pressure vessel, housing the BME280. It looks as it is: a quick weekend hack using "off the shelf material". Just a temporary solution.... ;-)
Thank you Andreas! Our remote house in Scotland has an oil tank and I had a go using 40khz Ultrasound to measure the level. But the reinforcement of the tank walls makes lots of unwanted echoes. This is a much better solution and I can house everything in fuel store shed next door.
If you create a wave-guide pipe for the ultrasonic you can solve this problem. Make sure you have a small vent slot within the dead-zone on the ultrasonic sensor and a 45° cut on the bottom of the pipe. The wave-guide pipe diameter can be close to diameter of the sensor, it should still work.
A low value screw in pressure sensor will do what you want, but see my comment above on fitting direct to the tank. The connector is weatherproof, all you need add is environmental pressure sensor to compensate.
My DIY water-tank meter uses an ultrasonic parking sensor (basically the same as those other ultrasonic sensors, but with a single part acting as transducer and receiver in a watertight package, typically sold with the electronics needed to drive it) to measure the distance of the water-level from the mounting point of the sensor (which is above the overflow opening of the tank). I use an Attiny85 to do the measurement and have it fully turn off VCC to this distance sensor module when it goes into deep-sleep to reduce unwanted current consumption. The Attiny85 also has one of those simple 433MHz transmitters attached that it uses to send a packet of data to my main 433MHz receiver node (which I also use to receive a signal from an Attiny85 triggered by the 10VAC doorbell, some 433MHz remote controls and in the future hopefully a letterbox monitor). The Attiny85 is directly powered by one 18650 battery and then a second 18650 is put in series with that first one to create a higher voltage for the 433MHz transmitter to increase the range (distance needed is around 20m and works fine this way - I haven't tried whether a single 18650 could have been enough). The electronics are in a 3d-printed case, with only the sensor being in the tank - it is delivered with a waterproof cable attached). The setup works great for months now on a single battery charge, taking measurements every 10 minutes, so if I want longer battery-life, I can tune down that frequency quite a lot and have it go for years. Before this setup I used something similar with the same parking sensor, but it drew too much current while in deep-sleep due to the sensor getting VCC and using a large 12V battery which required voltage conversion. Every couple of months I had to recharge that large battery and as the Attiny85 also monitored the voltage of the battery, you could see it drop day by day. With the new setup the 18650 has only dropped from 4.4v to 4.3v in 2 months, so that should leave quite some time to go. I doubt a system with a pump and pressure sensor can become this power-efficient, right?
A good project! Some people mentioned that their ultrasonic sensors corroded over time because they have to be mounted above the fluid. So I am glad it did not happen to you. And 433 MHz transmitters are a very cheap and good solution for message transfer!
@@AndreasSpiess The risk for corrosion is exactly why I went with one of those car distance-sensors as their use in automobiles requires them to be much more water- and dirt-resistant. The sensor has been in use for around 3 years now and is still working fine.
Andreas, I may know why this pressure buildup happens: When you turn on the Air Pump (in fact a small piston/diaphragm compressor), there is a gas (air) flow rate. When tube is filled with water, the counter-pressure (measured in the sensor) is zero only when compressor is turned off. As soon as you turn it on, it measures 2 pressures: (1) Static pressure (“static head”) due to hydrostatic column displacement: internal meniscus levels versus your “reservoir” level. (2) Dynamic pressure composed or two parcels related to viscous friction: (2A) pressure drop caused by airflow in the tiny-Tube and (2B) pressure drop due to “head-loss” in the length of water flowing out of the tube/hose. At a given point along the pumping, the friction-losses + static head reaches a maximum, then equalizes NEAR the static pressure, when the tube is 100% air-filled. NEAR is not EQUAL, because even Air being much less viscous than Water, it will still create a counter-pressure added to the hydrostatic pressure (the one we would be interested). As this values can be minimized if using larger diameter hoses, or the error is mostly the same (“systematic”), a relatively accurate measurement of the VARIATION of the liquid level can be measured. Another point of great importance is to know the actual fluid density. For water, this is easy. For oil or other fluids, they will vary more. And if the fluid is not a simple substance, but a solution/mixture, things can become more complicated, if the concentrations and temperatures are not monitored. For instance, water with 35/1000 parts of salt (for marine aquariums, etc), density is 1030 Kg/m3 or 3% more than fresh water. Same goes for sugar or alcohol concentrations in beverages. Furthermore, for carbonated liquids (beers, sodas) the apparent density can vary if gas has reached the saturation (bubbling) limit or not. Bottomline is: for measurements of known liquids where 1~5% of measuring ACCURACY is acceptable, and even better reading RESOLUTION, this is a good level metering method. I hope it helps.
This is a wonderful idea to measure the level of a fluid. This question has been asked a few times before by my users (ESPEasy) and I never thought of using the same principle as a blood pressure sensor. Really great video!
One of my upcoming summer projects is to measure the depth of my rain barrels. This solution is very interesting as it doesn't require the sensor to be inside the barrel. Very cool! I too will be using LoRaWAN to transmit the data back to a collection node. Great video!
What a coincidence to find this one. I was about to setup the water tank in the garden and had prepared a TOF, though worrying about getting it water tight. Now, I recently had to replace my old blood pressure Hestia (20 years old) with a newer one and had not yet dumped it. Seems it just found a new purpose :-) I'm double happy!
13:03 just to clarify: 3D printed parts off an FDM machine (i.e. Ender 3 style printer) are not air/water tight. However, 3D printed parts off a Resin based machine can absolutely be air/water tight!
Covering an FDM print with a few layers of resin can also work, but that's a mess that isn't really worth the time. I did it once and don't want to do it again :D These days companies like jlcpcb offer also cheap 3d printing in resin. Even large parts cost just a few dollars so it could be a good option. I'll definitely give it a try the next time I place an order. Although I also like the idea of re-using old blood pressure sensors. Often you can find them on flea markets or people throw them away....so it's a good opportunity to reduce the e-waste.
@@AndreasSpiess You might want to investigate the process used for "smoothing 3d prints", where the FDM print is exposed to an enclosed atmosphere of vapour that softens the plastic surfaces so top layer flows and fills in gaps. It would reduce surface imperfections that enable air leaks. Less messy that painting or dipping parts in resin to seal surfaces. Still have to consider added time and investment (ie: may not meet cheap and convenient requirements)
Good video! I made an almost identical gadget a few years ago to measure the depth of water in the cistern that supplies water to my house. I didn't use a solenoid though, I simply switch the pump on for a few seconds to purge the water from the air line then take the measurement. I have the same issue of the the pressure "reading high" whilst the pump is active, this I presume is due to the back pressure of the air because the line is so thin. I bought the pressure sensor from Digikey that outputs a voltage proportional to the pressure but, like you, used a pump from a blood pressure monitor.
Love the idea and solutions shown here. The tinkering involved and testing is also great fun. When the problem was posed my mind went immediately to "bobber on string" kind of solutions though lol Simple weights, a float, strings, and trip switches.... but everything has it's own set of problems to overcome.
I built from separate parts. I use an aquarium membran air pump powered from mains, switched by a relay which went out of order after the last summer. I want to use an SSR in future. Your maximum pressure peak comes from a counter pressure wave in the system until the flow in the system has become steady (or is stationary the better word?). It also appears when you run the system with all the pipes outside the water and on the floor - albeit *much* shorter, by an order of magnitude. You should do such a unloaded measurement -I had to throttle down my air pump quite a bit, because its air flow was way too much for my aquarium hose diameter and length. Remember, in the blood pressure thingie you have half a meter, in your fluid level measurement setup you have 12 or 20 m of line, and the air inside will have quite some mass to it, as well as friction to the hose walls. I had quite some difficulties with fluctuation in my values. It improved much with building a silencer between my pump amd my measurement box, and the introduction of a unused cigarette filter right in front of the sensor. When standing still and valve open, the liquid will rise in the standing pipe, it's displacement at the start of the next cycle will even add to the peak. You can experience the counter pressure first hand (or better first lungs) if you blow your setup with the lungs. It will help you to shift your thinking from "static pressure" to "stationary flow". The measurement shall be done when the presure value changes have calmed down and become stationary, and the last N samples have been the same or nearly equal. You can avoid the rising fluid level in the stand pipe with an "lip valve" from silicone. Those are sold as spare parts for the rinse aid dispenser unit of dish washers. it has to be big enough to not introduce counter pressure on its own. I ordered some samples and mounted them and tried them with my mouth. Also, I plan to use a thin copper pipe as a stand pipe for this season, because I measure the level of an open pond, and I had various problems with biofilms, algae, etc. If you introduced air distribution valves, you can sequentially measure levels of multiple reservoirs, even on different heights, if you correct for the different ambient pressure at the different heights, in relation to the height of you pressure sensor.
Thank you for sharing your experience. I thought something similar as you describe with the "wave" but was not sure. A good point with the copper pipe against biofilms. My pump does not move a lot of volume but is able to develop quite some pressure (it has to be higher than blood pressure). But maybe slowing it down would reduce variations. This should be easy with a resistor.
Great video Andreas. I moved our washing machine to our basement and had to pump the dirty water to the ground floor. This pump had to work automatically off course. So I started with a variant based on ultrasonic distance measurement of the fluid hight. That flawed on the foam on top of the dirty liquid. Then placed level sensors, but those had to be in the water and got really dirty and stopped sensing. Then a colleague said that measuring the weight of the tank solves a lot of these problems. So I added a hx711 based scale under the tank and that worked like charm since a year now.
Great! I've implemented this idea with absolute sensor and atmospheric pressure sensor to measure water level in my well. Make the tubing air tight was a challenge, but finally conical thread and polypropylene tubes solved the problem. Another issue was to find an equation between pressure sensor readings and the actual water level. I've declined to recall hydrostatic lessons and acted as an engineer: just pushed the tube with markings each 50cm and thus got a calibration table. The final accuracy is around 50cm, and it's fine for me. I'm re-filling the air and re-calibrating the system each spring.
the pressure in the plastic pipe decreasing after reaching a max (i.e. the moment when the clear tube is just about purged but no air has left yet) is most likely due to the pipe getting filled with a mixture of air and water. the average density of this mix is considerably lower as of pure water, hance your readings are decreasing
Two thoughts came to mind - kicking a car tyre to see how much pressure it has in it (higher pitch = more pressure), and blowing across the top of a bottle which produces a different note (higher pitch = less air in the bottle therefore higher liquid level). I wonder whether there is an acoustic solution to this? You could rig a speaker/microphone to the outside of the tank, pulse some sound (probably just a click) then measure the pitch of the returning sound? Would probably need a lot of tuning to the individual circumstance but perhaps some machine learning could do that, or failing that just test the tank at various levels to obtain the values of the return signal at each level. I'm sure it's not easy and may not be very accurate, but it is at least external, no moving parts and quite low power.
I bet braces, welds and other artifacts in the tank beeing above or below liquid surface distort the sound, making the calibration a huge job in itself. but for a small homogenous tank it's a really interesting idea that might be genious in that rare instance. Thanks for the Input.
It exists. They measure the resonance frequency of the remaining volume above the liquid. This principle was discussed in the comments of my last video about the topic.
Great videos Andeas, you put in a lot of effort to create content, thank you! As you are aware, that design may be a little over-engineered, especially if you have a limited power supply source to drive the pump and valve coil. A bit more info about the common pressure sensors - you get absolute and relative pressure. The one with the holes are relative pressure sensors and they take your altitude into consideration. An error in measurement can occur if you do not calibrate your sensor at full scale at your location. Not a big error, but a Swiss in the mountains will have a different barometer reading to someone at the coast. Additionally, just the sensor with a sealed membrane close to the bottom of your tank is sufficient. Tap and thread a fitting (t-piece in the drain socket perhaps) , connect your sensor orifice, calibrate your reading with a full tank and you are done. Low power and easy to setup. Error percentages you need to be careful with and it depends on final accuracy required. 1% error does not seem much until you have a 100kl tank (or bigger). Once again, if you need more info I will be happy to share.
This is an industrial type sensor, sometimes it's easier, quicker, and probably cheaper in the long run. I have been doing this for a while in the automation world. You could use the 0-10V output version, that would ensure long battery life as you only need to sample once per hour, depending on fill/drain rates and size of tank.
Thank you, great video. Really interesting that you explain the principles, the shortcomings of what we can source cheaply and then go on to build your own sensor
I've had a throw-in type pressure sensor at the bottom of my rain tank for about a year now which had been working great but lately the values have been going up and down rather randomly. Quite frustrating. I have the same sensor in my pond and it doesn't have any problem so far :-\I've had a throw-in type pressure sensor at the bottom of my rain tank for about a year now which had been working great but lately the values have been going up and down rather randomly. Quite frustrating. I have the same sensor in my pond and it doesn't have any problem so far :-\
I am working in something like that, but for a pretty simple pump system. I started with the same idea, but dealing with the fail of air tighteness... just a fish tank air pump or something similar...thanks for your work, that was a nice idea.
Love your channel. I'd like to add to the discussion. Your readings are fluctuating because 1- The pressure pulses of the pump is not dampened, 2- Large bubbles are created because the air flow is too high. 3- The shape of the outlet of the tubing is not optimal. You could dampen the pump pressure pulses by using a small air camber reservoir to dampened it. Adding an adjustable flow meter on the outlet of that reservoir to limit the air flow or install a fixed orifice to reduce the flow would also help stabilize your readings. You only need a very minute amount of flow for bubbles to flow out the tubing but enough flow to empty the tube of water in a reasonable time at startup. I suggest cutting the end of the tube in a 45 degree angle. It makes a huge difference as it allows the bubbles to flow out the tube easier with less resistance and more uniform bubble size and consistency. Cutting it flat causes that larger concave bubble shape to form at the end of the tubing just before the bubble breaks free. Also Using a 45 cut on the outlet tubing you can butt it on the bottom of the tank while a flat cut tubing outlet cannot. Also another concern I would have using a flexible tubing is it rigid enough to stay in place at the bottom of the tank while in use? When the tubing is full of air, the tubing will have a tendency to lift off the bottom of the tank. Its better to use a small rigid pipe like PVC cut on a 45 degree angle. The tubing outlet must not be allowed to move and will not be reliable if not secured. Be aware for high precision the tubing outlet distance above the bottom of the tank must be subtracted from your overall calculated height range of the tank.
Thank you for your explanations. The fluctuation is not a big issue because I measure when the pump stopped. The other values were measured for learning only. And I agree: You have to attach a weight to the tube to keep it at the ground level (was not needed in my narrow "tank")
Thanks for this! We have an old underground rainwater cistern that I've never been able to work out measuring electronically. This is a really neat method that might actually work! Now I just need to figure out how wide the cistern is...
Perfect! I build a big water storage last weekend. And we discussed how to measure the water level. Problem solved. :-) In the most complicated way....
Thanks for the info and references on these cheap pressure sensors. Have been toying with idea of making a solid state anemometer (without moving parts) to measure wind. Using 3-4 such pressure sensors placed in circular pattern to measure wind pressure, and thus its speed and direction.
Great project, I enjoy your thorough investigative thought processes and low cost design approach! Well done, thanks for sharing your knowledge ,Andreas.
Great hack. I had one of those pressure sensor which I bought from Aliexpress to replace the one in my washing machine, and has always been wondering what to do with them. Now I know. Thanks very much
You are testing at zerro airspeed, which is only good for hovering. Maybe good to just go fly, try different configurations, see what works best.You are testing at zerro airspeed, which is only good for hovering. Maybe good to just go fly, try different configurations, see what works best.
In the industrial environment, one usually does not interrupt the pump. Instead, the flow is used through a volume with a downstream throttle to minimize the influence of back pressure with a minimum amount of air into the fluid. In German it is called Einperlmethode. Greetings from Hamburg
I use MP3v5010DP for my 150l tank for now. It is basically the same as MPX5010DP, but rated at 3.3V. ESP32 with lut-table or ADS1115 with it works somewhat good. I really like to try HX711 Wheatstone Bridge setup with four 4-wire cell type strain gates, because Tasmota HX711 driver has just got major fixes in the Tasmota dev build.
Love the idea and solutions shown here. The tinkering involved and testing is also great fun. When the problem was posed my mind went immediately to "bobber on string" kind of solutions though lol Simple weights, a float, strings, and trip switches.... but everything has it's own set of problems to overcome.Love the idea and solutions shown here. The tinkering involved and testing is also great fun. When the problem was posed my mind went immediately to "bobber on string" kind of solutions though lol Simple weights, a float, strings, and trip switches.... but everything has it's own set of problems to overcome.
As a sugest, it was interesting to do a new Level Meter using ultrasonic, for this it can be used simples electret microphones and earphones as ultrasonic source.
The problem with this is frequently water thanks are going to be exceptionally high in humidity. I've killed more sensors than I can count trying to use them. I've tried coating then, potting everything with epoxy, you name it, and eventually the ultrasonic sensor itself rusts and the system fails. I've been watching this topic on the channel while also searching for my own solutions... This once seems like it might FINALLY be a workable one paired with a solar panel and LoRa.
Einperlmessung 😊 twenty years ago a E+H PIR could be configured to be used as a local display for the level (can‘t recall if volume too) and present the level to a DCS in the form of an intrinsically safe current loop. Happy to see this can be done at home for a small investment minus the great instruments and electromechanics of those transmitters. Andreas, I am not sure if this is mentioned in the „Dubbel“ (I‘m not joking! it‘s one of the books I cherish the most!!) too but it is a source of inspiration for anyone interested in machines in general
These instruments aren’t used for precise level (and, if calibrated: volume) measurements but for rough estimation and then charging a process via a mass flow FQIC but often thereafter weighed in a buffer tank on a load scale. Temperature ranges in tank farms are quite broad (50C nothing out of the usual). Level indicators like this can be in the +/- 5% range due to the effects of temperature differences on the liquid in the tank. Hope it works out for you - maybe use a very fine needle and a steady stream of gas for more precise measurements (we used N2 or Ar)
Nice,. From my expirience, temperature coff is much higher then in spacifications. In order to realy use this simple sensor, you have to add temperature measurment and software for compansation.
Bubbler tube output is very sensitive to air flow, the commercial units usually have a air flow controller of some sort. Either a variable-area flowmeter (rotameter) or something similar. As level varies, pressure varies, so flow varies. Lots of variables to think about, but they are still very popular in special applications. Corrosive, flammable, or explosive fluids, and also where accumulation of solids are a problem, like sewage and sugar wort.
Andreas, thanks a lot for your educational videos, that helps to create a truly engineering mindset. Regarding this problem, I would have used a sumergible pressure sensors, they are cheap, with for example 4-20mA output
I'm showing my age, but I've worked on a foxboro model 40 pneumatic PID controller with a bubbler level measurement converting the 2m tank level via a pneumatic scaler to the 3..15psi signal the foxboro needs. As I'm writing, the details are coming back to me. It was two model 40s in cascade, with level controller output being setpoint of a flow controller connected to a Fisher proportional control valve. All pneumatic, no electrics whatsoever.
Very interesting project! Could be a good solution for a rain water tank under the ground that is hard to reach and only accessible from the top. So a range sensor or this could be used. One thing I wanted to say I have printed air and water tight parts before. Of course I do not know to what pressures they remain water tight. I printed many flower pots over the years. The design is very important to be watertight. You can also line the walls with super glue to fill holes if there are any.
It seems that airtight is much harder than watertight. At least I was not successful with airtight :-( The range sensors were part of my former video. They tend to corrode.
Thanks Andreas, I have two 5000 litre rain water tanks that I’d like to monitor. I do think though that there is a simpler and cheaper solution than the blood pressure monitor. I don’t know about Switzerland but here in Australia garden shops sell parts for small level irrigation. One item is a “T” piece made of ABS and capable of withstanding mains water pressure. Each leg is barbed and fits into appropriate size tube, with the addition of a small cable tie, it should be air tight. Add a small cheap battery air pump and you have an equivalent system.
If you want to go 100% analog and 0% electronics, you could even do it with a small air compressor (aquarium ones can “pump” up to 1m depth, medical nebulizers can do 5-10m of water column (1atm = 10m w.c. = 14.7psi) and those cheap tire inflators go to 20~40m c.w.! So, if you “pump” air into the tube, then monitor the air pressure (pressure gauge, or U-Tube manometer), you have your level reading, electronic-less - or use the electronic sensors as Andreas introduced.
Oh! I think I know why there is some drift in the pressure reading. The temperature of the water is typically lower than the ambient temperature. So when you pump the air into the tubing, it becomes colder and loses pressure. So it would be best to read the pressure immediately after the air pump stops.
Great content as always Andreas! A quick off-topic question: What labels are you using to label your project box? My labels don't stick on the boxes... E Gruess us Muttenz
I am working on something similar with a better sensor (differential pressure sensor from All Sensors, ELVH series). The nice thing is the sensor includes the ADC, but it also costs an order of magnitude more than the cheap sensors. In terms of the connections, bulkhead fittings with small hose barbs or Luer fittings are relatively inexpensive.
nice. A peristaltic pump and part for a small scale irrigation system from the gardening senter should work. There you have Tee's and valves. On offshore oil reservoir, there are sometimes huge concrete tanks at the bottom for local storage of crude. These tanks are open at the bottom so water flows in and lift oil up to the inside ceiling. The lower density of oil gives an internal pressure in this tank and that pressure is proportional with oil content. There is on some, no need for a pump to lift the oil up to the ship, due to this pressure. Just open valve and the oil comes gushing out. These tanks are near maintenance free and all instrumentation is at the outlet of the offloading hose. So the principle is similar.
The pressure should go up as you add air to force water out of the measuring tube. It has to be greater than the pressure of the water in the tank. Also have to take into account that air is compressible. So, probably not 100% accurate. Perhaps Nitrogen. But who's going to do that!!! Once you get air "bubbling" then you can cut off the motor and measure the pressureThe pressure should go up as you add air to force water out of the measuring tube. It has to be greater than the pressure of the water in the tank. Also have to take into account that air is compressible. So, probably not 100% accurate. Perhaps Nitrogen. But who's going to do that!!! Once you get air "bubbling" then you can cut off the motor and measure the pressure
Hi Andreas, in fact some years ago I did have something similar in mind but did not build it for several reasons. One was the non linearity as we have a uncompressable fluid and a compressable gas that do have non linear beahviour. However this could be corrected in software. in fact I did choose to use a capacitive sensor instead. Basically two isolated wires that do nowhere have contact to the media. These two wires act as an tunable capacitor for a NE555 based frequency generator that does change value linear with the coverage of media. So in the end I simply measure the frequency which is a proportional value to the coverage with the measured media. Actually I created an dedicate PCB with an Atmel ATMEGA 328 ( yes, it's an specialized Arduino clone in the end :) ) for that to be placed inside the cistern. Ä gieslie vum Südweschd ecke vu Dütschland.
Your setup with the tube looks exactly like an experiment I do with my students, but I went only for one meter long pipe and we vary the depth by lifting out the tube instead of changing the water level. OK, and we don't use the pump-solution but only the pressure sensor.
I also changed the inserted length of the tube (I have no possibility to check the water level inside the "tank". I just made sure that it was always completely filled.
Actually there is a way. You can have an external water column connecting the top and the bottom with some watertight fittings and a see through hose or pipe. It’s not easy to achieve in such a small diameter as the one you used, but in a normal tank it’s a very common solution and one that can at least get you in the right ballpark. If all that’s needed is to trigger the pump on and off, then it’s pretty straight forward. Most applications don’t require the volume of liquid contained in the tank thus making it sufficient for the majority of cases.
There is a way to make 3d printed parts pretty airtight and that is to paint them with plastidip.. that will do the trick as long as it is only between the outside and the insides... if it needs to separate 2 chambers in it... well that is more difficult but for infusion and fishtank supplies you can buy these kinds of joints and T-joints and taps.
Very good Andreas, but may I suggest a longer delay before taking the reading to allow the membrane to achieve a stable state. I installed a cheap chinese screw-in pressure sensor on the suction line between my underground water tank and adjacent underground pump, approximately 1.2m static hydraulic head. No atmospheric pressure compensation as yet. This was run in parallel with an ultrasonic sensor which was prone to forming a condensation drip on the transducer face in sub-zeroes. Result - The US failed at -10, the pressure sensor continued reading. Although I suppressed readings when the pump was running, what I had not considered was membrane distortion under vacuum (when the pump was running), and observed 20-30 minutes before the reading stabilised. Readings have drifted with such repeated distortions over time, plus of course atmospheric pressure effects. I will reconfigure the arrangement in summer during the annual tank clean out. You will not be abusing the membrane to that degree by blowing bubbles, but it will alter the SG until the air bubbles disperse from your column. If you check how soon the sensor takes a stable and consistent reading, you can set your delay to suit.
@@AndreasSpiess If the sensor pipe is airtight you cannot lose pressure, it will stabilise as air and water reach equilibrium. For a 3m pipe, 5 or 10 secs should be more than enough for the bubbles to clear - You will lose a small hydraulic head as the air displaces water but then disperses but it will probably be a constant. OT - In the event your water tank is groundwater fed with a high soluble iron level - I made a venturi aerator with a small submersible 100w water pump and installed it in my tank last year. No more soluble iron.
Page 1:08 demonstrate an intermediate remote air pressure sensing next to the local air pressure scheme. Both exhibits an offset error value in proportion to temperature at calibration. To remove the -ve offset ambient air is injected into the detection hose by the input port until reading settle to a steady state. That should be repeated regularly no longer than temperature cycles, or continuously as long as valid reading is desired. Air tight? Unnecessary. Just keep the top of liquid at ambient value and remember to keep refresh the air pressure in the air portion of sensing tube.
Great video.I jave the same requirement to measure the level in an underground rain water tank. Is it possible that the pressure rise during pumping is due to friction as the air and water move in the pipe, once the flow stops, there will be no differential as there is no flow. For the saturation problem, could a series resistor with the bridge drive voltage reduce the output (or parallel if current exitation).
Of course you could try to add a voltage divider. But you have to be cautious because it influences the resistors built into the sensor. Changing the HX710 seems to be a better solution for me.
You mention maintaining air tightness is difficult but what about all that pneumatic equipment that exists? That has to be air tight. Relying on a pump isn't ideal either. It can break and then you have to replace it, which is difficult in remote areas. What if water enters the pump and destroys it?
I agree. Both possibilities have advantages and disadvantages. Others mentioned that the pump approach removes clogging while the "passive" approach tends to clog with algae etc.
Dear Andreas, very interesting video. If it can help, an « easy » way to have a airtight container for measuring the pressure is to use a glass container (a bottle for example) and turn it upside down. You can then put the sensor inside the container and plunge it into water with the wires passing through the top of the container to the surface of the tank. In this case, sensor is in a water bubble and is able to measure the pressure inside the bubble with a linear correspondance to the water level of the tank (hope it is clear ;)). Other point, i also used a BME280 which is very accurate (detection of a difference of 5mm in the level ! But unfortunately with a limited range of measure…). Thank you for your great work
@@BenMitro Thank you! I’ve forgotten to mention that the sensor should be waterproof because there will a lot of humidity. I used a bme280 that I have "tropicalised" with some coating in spray. This has obviously no influence on the pressure measures as the coating is flexible. I have built this for measuring the water level of my pool. It worked with an unexpected accuracy (5mm) but unfortunately, after a few months, the sensor broke (I have tried 3 times…). I have installed this assembling on the circuitry of the pool in the technical room. But, unfortunately this leads to pressures around 1450mbar when the pump is on due to the pressure linked to the water flow. Even if the bme seems to be able to measure this level of pressure, it is outside of its working range. My analysis is that this is the reason why the sensor goes out of order. For the moment, I haven’t found a sensor with the same accuracy, ease of use but working at a higher level of pressure and I’m stuck. Does somebody has any idea ?
@@christopherisse6857 Christophe, do you use that spray on circuit sealing compound of just clear spray? I need to do this to some gear I am putting in my camper that is not intended for such environments. The specialist spray is expensive and I have a lot to do.
@@BenMitro Hi Benny, not sure to rightly understand your question. The main circuit with an ESP32 is outside the glass container and not in contact with humidity and doesn't need any specific coating, just to be in a waterproof box if you want to use it outdoors. The only part I coated is the sensor (BME280) itself. I linked it to a connector for the I2C bus and I coated the sensor + the connector + the base of the 4 wires. Hope it answers your question
After many abortive attempts at measuring my well depth over the years, I finally explored the 'air line' method, which is well known and as you discuss. You need not have upset your wife in using the blood pressure monitor! It'll do the job without modification. The first phase of blood pressure measurement is inflating the sleeve. The monitors I have show the inflation pressure as it progresses. In this application it'll fail to achieve target and report an error. But the pressure figure it give before failure is accurate in mmHg. So there you have it - a ready-made and cheap liquid level meter. Blood pressure monitor Inflation stops at about 0.3bar so the method is limited to 3m depth of water.
You may find some of the variation in readings is caused by the flexible tube moving upwards in your pipe once you pump it full of air - maybe a rigid tube would provide more consistent readings.
10:57 how do you know the loose end is not rising inside the pipe? The hose is light enough... In the industry, the vessel has a tap at the side, near bottom, if the tank is atmospheric, or at the side near bottom and top if it's pressurized. That takes care of holding the measurement points at the same height. Everything bellow the lower point can't be measured. You don't normally install at the bottom because that is prone to blocking.
This video reminded me very much of my washing machine. About 6 mos ago I removed the control board to replace some bad capacitors and it used a sensor almost exactly like this one only twice the size and DIP-8. It was on the control board. The sensor had a built in analog amplifier 5V in and 0-4V signal out, rated for 6kPa (ADP51B62A01) price is about $30. It has the same tube to the bottom of the tub for water level sensing, but since it is drained for every load it doesn't need the air pump. The water level is measured from the bottom of the tube, not the tub. I did some math at the time, changing the air volume in the tube changes the sensor calibration. Adding more air volume should cause the tub to add more water to reach the same pressure setting. Just theory though, I haven't tried it.
@@AndreasSpiess I started thinking about my washing machine again and of course it is a little different from your project. It occurred to me though that with the air pump, your goal is to fill the tube from the sensor completely full of air (not allowing any liquid to rise in the tube). If you used a small second tube from the pump that went to the bottom of the sensor tube you could just release air bubbles into the opening at the bottom to fill it. By doing that, the valve would not be required (but not as much fun). Also, the sensor tube should be large enough where the bubbles couldn't lift any water up to the sensor as pumping the air in from the top helps to keep the sensor dry. Just thinking.
For the variability in readings I am thinking either the water level during the tare phase is bouncing up and down in the measurement tube or the tube itself is collapsing after the purge and thus not a constant volume.
on cargo ships, the ballast tanks are usually monitored with pressure sensors but in my experience they were often unreliable and required constant maintenance.
I've had good luck printing airtight parts with a resin printer. Most resin printed parts are very brittle so I use non-brittle engineering resin from Siraya Tech. BTW I also use 100% infill.
This reminds me of a discussion of how to make sure a radon eliminator was active. The indicator can tell if the fan is working using a pressure differential. Sounds familiar.
@@AndreasSpiess A better word might be exhaust. It's a continuous running fan that moves air from a basement to the outside. It creates a vacuum and you can see it's working because they use a tube of liquid with marks on it. The vacuum that sucks the radon out also moves the liquid. We had a conversation on Reddit how to monitor and alert users of failure. I've given up posting links because RUclips deletes them but you can Google "Looking for a way to get a notification if my radon mitigation system loses pressure"
Very interesting .... and original solution..... I built my own DIY sensor by using a air absolute pressure sensor jailed into a completly sealed non circular plastic bottle (dishes bottle for example) filled with 1/3 of concrete (to let the bottle layed on bottom of my well) and 2/3 of comestible oil (I used oliver oil). Of course in order to report the air pressure, you need a small tube attached to the sensor, filled with air and dived into the oil. The sensor is jailed into the bottle capsule everything is waterproof with epoxy. On top of my well I got another air pressure in order to have differential pressure measurements. It works greats since 2 years from now..... My well is 20m depth..., reading analog value with an ESP32 w/ESPhome.
Have you done a video of measuring rain fall and the amount per hour, per day etc ? Looking at using a sensor that does this and solves the cleaning of the sensor after a certain time, to stop it getting dirty or "Clogged up". Thanks for such great videos, which have really helped me with my IOT weather projects.
I wonder if a rigid bubbler tube would reduce the variation that you see when you're pushing air into the tube? Or, alternately, if a smaller diameter air tube would show a different (reduced) variation at mid reading - if I understand what your statement was. Also, with a flexible tube, unconstrained in a large tank, will boyancy be a problem? Pushing air in may make the tube light enough that the end floats toward the surface.
I’ve found 3d printed parts with more than three walls to be airtight. More than airtight, in fact. They hold up well to 3 bar pressure(29PSI differential)
It was interesting :-). During my working life - quite a while ago now - I leak tested plastic containers but the test pressure involved was very low, it is easier like that. So anything involving pressure testing is interesting to me. I believe btw that Washing Machines use the pipe filling principle to control the water level in the machine. I have no idea of the details.
@@AndreasSpiess in fact they are simpler pressure switches - a diaphragm pushing electromechanical contacts - just like a bimetal thermostat, but here using diaphragm as force source.
10:50 i guess the hose is to wide so the pump does not pushes all the water out, because the airflow is to low and water passes up while pumping... the fluctuating values might be a result of the hose beeing flexible the the airbubbles get everything ocsillating. i would recomend to use a very little diameter for the measuring stick (maybe 1mm brass tube)
I agree that the diameter matters. However, I thought the pump moved the air/water quite fast. It only took a few seconds to get the water to the top of the tank. Others suggested adding a reduction at the end. I will do some experiments...
@@AndreasSpiess I answered few moments ago, on the main post. Here, and based on my past explanation there, I would summarize: it will WORSEN the accuracy and intermediate pressure buildup problem is you create a localized element to create a pressure drop at the end of the tube.
Another way of measuring fluid level would be to use an ultrasonic distance sensor somewhere above the top level of the fluid. There has to be an air gap between the maximum level of the tank and the sensor though (this is because whether you use a ready-made module or roll your own sensor with ultrasonic transducers and MCU, the pulse you emit has to finish before you begin to listen for the return echo). Pro's: No need to find airtight tubing and suitable pumps. Transducers which are air- and watertight are available. Instantaneous measurements (you could technically be able to follow the sensor in real time). Con's: Constraints as to the placing of the transducers relative to the walls of the container whose fluid level you wish to measure. Minimum gap between maximum fluid level and sensor. Having to delve into the dark arts of digital signal processing. Driving the transmitting transducer. Amplifying the signal from the receiving transducer. Temperature, air pressure and humidity compensation for accurate measurements.
The biggest problem in a water tank is high humidity. After a short time, water droplets form on the sensor and massively falsify the result. After numerous disappointing attempts I also switched to a pneumatic sensor and am very happy with this solution. What would make me even happier would be a transmission of the measured values via LoRa. So I will follow this project closely.
@@HansPirngruber Depends on the sensor shape and what if any signal processing one is doing. What you choose in the end depends on what applications a specific sensor is most suitable for. Was just posting it for completeness of the methods available for fluid level measurement :)
Next It would be great if you could build a capacitive level sensor, the prop itself is easy to build, I did it using two sheets of aluminum foil 2cm wide next to each other and laminated, its resolution was about 1mm and seems to give accurate results, the hard part is building the capacitive to voltage converter, I still couldn't do it with simple available parts with no arduino.
For my problem I will go with the pipe because the tanks are quite deep. But your proposal for sure is also good. The capacitive humidity sensors use a NE555 to measure capacitance. I did a video about it.
Def Great! I've implemented this idea with absolute sensor and atmospheric pressure sensor to measure water level in my well. Make the tubing air tight was a challenge, but finally conical thread and polypropylene tubes solved the problem. Another issue was to find an equation between pressure sensor readings and the actual water level. I've declined to recall hydrostatic lessons and acted as an engineer: just pushed the tube with markings each 50cm and thus got a calibration table. The final accuracy is around 50cm, and it's fine for me. I'm re-filling the air and re-calibrating the system each spring.
Very interesting setup, Andreas! A useful warning about the HX710 module which - if I understood your explanation right - means it's been engineered close to uselessness. Aside of the fun building this I wonder if a VL53L1X ToF sensor would be a simpler approach? I found them to be very precise. That with a LoRa module?
A ToF sensor had to be inside the tank which would sooner or later lead to corrosion. The advantage of this approach is that no electrical part is inside the tank.
At least here in Germany many older oil tanks are equipped with the tubing inside allowing this air bubble method. Decades ago (but also up to now), mechanical pressure meters were available to guess the inventory. But at least according to my experience it was far more or a guess than a measurement. Using modern electronics instead of the mechanical meters, I achieved measuring errors < 0.1% Even the restrictive German TUV had nothing to complain. All electronis is located outside of the tank. No cables inside the tank needed.
Can you try an Archimedic method? JustUst measure the change of weight of a floating cylinder, again with a 711. Another way i was considering was with a float, that pops out of the tank and has a lidar on top. The lidar reads the dostance to the roof. In both cases, I end up worrying about friction, bacteria, slime etc... and not doing anything :)
It looks like there was moisture in the small plastic tube that goes back to the pressure sensor. If that moisture drips into the actual sensor it will distort the sensors readings.
I can imagine that moister is not good for the sensor. In my case the tube was long enough that it did not go back. And I would hope that regular measurements will dry the tube at the top.
And you probably need to add a weight to the bottom of the measuring tube. the test tube probably was too narrow to have the measurement tube float to the surface.
Many people dispose of portable BP Monitors because the cuffs start degrading/leaking. I must remember the good salvage parts inside! Would the pressure sensor from the Blood Pressure PCB be more accurate, or have a more suitable scale/resolution/output?
Great idea. Do you think an HC-SR04 Ultrasonic Module or JSN-SR0T4-2.0 ultrasonic ranging module could be used for fluid measurement? The cheap level sensors used in travel trailers go bad almost immediately and I was wondering if such a module could be used. It would be great if they could be used attached to the tank externally but that would only work if the signal can pass through the plastic tank.
10:44 for the next Experiments in this direction I'd say you could add a funnel to the top of the pipe for the bubbles to get out of the water easier, with the increased surface the water would rise less. At 10:56 my guess is at the surface tension of the water, or because the air in the water directly above pushes the water up and therefore slightly lightens the weight? Do you think that pressure change could be useable to run the pump as short as possible? Especially for remote stuff every mA may make a difference depending on the amount of measurements. (Ok, that amount may be adjustable depending on the use of the fluid and in case of water the weather. If it's rainy and you don't Pump water for the plants you probably also don't need to update the tank to often to check if it's empty.
Significant pressure drop occurs when water is displaced and bubbles start to exit. You can feel it if you blow through a hose into a bucket. I would guess that it might be related to the fluid not being compressible but the gas is???? Similar to the way break-through pressure works on a saturated filter, once break through occurs, lower pressure is required to maintain flow.
One advantage the bubble method is that biofilms and other dirt is disturbed sufficiently so it is very reliable even with dirty fluids. It is used in sewage processing plants. Using inert gases like nitrogen it is even suitable for highly flammable or corrosive fluids. To get more repeatable results it is beneficial to cut the air tube at an angle on the bottom. It makes the bubbles more uniform.
Good information. Thank you!
Thanks for this contribution!
Using readymade blood pressure units is a good idea! Also the HX711 module.
I am still using a BME280 to measure the inventory of my oil tank since 2016. Error is less than 10 l at an 10 000 l tank, better 0.1%. Not bad for a crazy "proof of concept" without optimization. I have recorded daily consumption data since 2017 which help me now to decide about future energy saving measures.
Cool. My results were not as accurate as yours. Good to know that the BME280 is a good sensor for this purpose. Probably not as easy to connect, though.
@@AndreasSpiess The BM280 has a limited pressure range. So, it is not as versatile as dedicated differential pressure sensors. But it fits for my specific application. You can find some posts incliding pictures on the German homematic-forum with the unique search word "Tomatensaucenglas" - which indeed is my pressure vessel, housing the BME280. It looks as it is: a quick weekend hack using "off the shelf material". Just a temporary solution.... ;-)
@@klassichd10 link pls. I dont find it.
Thank you Andreas! Our remote house in Scotland has an oil tank and I had a go using 40khz Ultrasound to measure the level. But the reinforcement of the tank walls makes lots of unwanted echoes. This is a much better solution and I can house everything in fuel store shed next door.
My situation is similar and I also hope this solution avoids the problems you described.
If you create a wave-guide pipe for the ultrasonic you can solve this problem. Make sure you have a small vent slot within the dead-zone on the ultrasonic sensor and a 45° cut on the bottom of the pipe. The wave-guide pipe diameter can be close to diameter of the sensor, it should still work.
Why don't you simply use a reliable throw-in level sensor?
Any other solution is really not worth it!
A low value screw in pressure sensor will do what you want, but see my comment above on fitting direct to the tank. The connector is weatherproof, all you need add is environmental pressure sensor to compensate.
@@robertlamont9455 can you provide a link to one of these types of sensors please?
My DIY water-tank meter uses an ultrasonic parking sensor (basically the same as those other ultrasonic sensors, but with a single part acting as transducer and receiver in a watertight package, typically sold with the electronics needed to drive it) to measure the distance of the water-level from the mounting point of the sensor (which is above the overflow opening of the tank). I use an Attiny85 to do the measurement and have it fully turn off VCC to this distance sensor module when it goes into deep-sleep to reduce unwanted current consumption. The Attiny85 also has one of those simple 433MHz transmitters attached that it uses to send a packet of data to my main 433MHz receiver node (which I also use to receive a signal from an Attiny85 triggered by the 10VAC doorbell, some 433MHz remote controls and in the future hopefully a letterbox monitor). The Attiny85 is directly powered by one 18650 battery and then a second 18650 is put in series with that first one to create a higher voltage for the 433MHz transmitter to increase the range (distance needed is around 20m and works fine this way - I haven't tried whether a single 18650 could have been enough). The electronics are in a 3d-printed case, with only the sensor being in the tank - it is delivered with a waterproof cable attached). The setup works great for months now on a single battery charge, taking measurements every 10 minutes, so if I want longer battery-life, I can tune down that frequency quite a lot and have it go for years. Before this setup I used something similar with the same parking sensor, but it drew too much current while in deep-sleep due to the sensor getting VCC and using a large 12V battery which required voltage conversion. Every couple of months I had to recharge that large battery and as the Attiny85 also monitored the voltage of the battery, you could see it drop day by day. With the new setup the 18650 has only dropped from 4.4v to 4.3v in 2 months, so that should leave quite some time to go. I doubt a system with a pump and pressure sensor can become this power-efficient, right?
A good project! Some people mentioned that their ultrasonic sensors corroded over time because they have to be mounted above the fluid. So I am glad it did not happen to you. And 433 MHz transmitters are a very cheap and good solution for message transfer!
@@AndreasSpiess The risk for corrosion is exactly why I went with one of those car distance-sensors as their use in automobiles requires them to be much more water- and dirt-resistant. The sensor has been in use for around 3 years now and is still working fine.
Andreas, I may know why this pressure buildup happens: When you turn on the Air Pump (in fact a small piston/diaphragm compressor), there is a gas (air) flow rate. When tube is filled with water, the counter-pressure (measured in the sensor) is zero only when compressor is turned off. As soon as you turn it on, it measures 2 pressures:
(1) Static pressure (“static head”) due to hydrostatic column displacement: internal meniscus levels versus your “reservoir” level.
(2) Dynamic pressure composed or two parcels related to viscous friction: (2A) pressure drop caused by airflow in the tiny-Tube and (2B) pressure drop due to “head-loss” in the length of water flowing out of the tube/hose.
At a given point along the pumping, the friction-losses + static head reaches a maximum, then equalizes NEAR the static pressure, when the tube is 100% air-filled.
NEAR is not EQUAL, because even Air being much less viscous than Water, it will still create a counter-pressure added to the hydrostatic pressure (the one we would be interested). As this values can be minimized if using larger diameter hoses, or the error is mostly the same (“systematic”), a relatively accurate measurement of the VARIATION of the liquid level can be measured.
Another point of great importance is to know the actual fluid density. For water, this is easy. For oil or other fluids, they will vary more. And if the fluid is not a simple substance, but a solution/mixture, things can become more complicated, if the concentrations and temperatures are not monitored. For instance, water with 35/1000 parts of salt (for marine aquariums, etc), density is 1030 Kg/m3 or 3% more than fresh water. Same goes for sugar or alcohol concentrations in beverages.
Furthermore, for carbonated liquids (beers, sodas) the apparent density can vary if gas has reached the saturation (bubbling) limit or not.
Bottomline is: for measurements of known liquids where 1~5% of measuring ACCURACY is acceptable, and even better reading RESOLUTION, this is a good level metering method.
I hope it helps.
Thank you for the explanation!
This is a wonderful idea to measure the level of a fluid.
This question has been asked a few times before by my users (ESPEasy) and I never thought of using the same principle as a blood pressure sensor.
Really great video!
Thank you! It was not my idea, though. I also was not aware when I did my first video.
Pet stores usually have a good supply of tubing, connectors, splits as well as basic air pumps (for use in aquariums)
Good advice. Thank you!
Remember that the air pump must be air tight as well when turned off for this to work.
One of my upcoming summer projects is to measure the depth of my rain barrels. This solution is very interesting as it doesn't require the sensor to be inside the barrel. Very cool! I too will be using LoRaWAN to transmit the data back to a collection node. Great video!
Glad the content is useful for your project.
What a coincidence to find this one. I was about to setup the water tank in the garden and had prepared a TOF, though worrying about getting it water tight. Now, I recently had to replace my old blood pressure Hestia (20 years old) with a newer one and had not yet dumped it. Seems it just found a new purpose :-) I'm double happy!
Go for it! A second life for tech is a good thing.
13:03 just to clarify: 3D printed parts off an FDM machine (i.e. Ender 3 style printer) are not air/water tight. However, 3D printed parts off a Resin based machine can absolutely be air/water tight!
Covering an FDM print with a few layers of resin can also work, but that's a mess that isn't really worth the time. I did it once and don't want to do it again :D These days companies like jlcpcb offer also cheap 3d printing in resin. Even large parts cost just a few dollars so it could be a good option. I'll definitely give it a try the next time I place an order.
Although I also like the idea of re-using old blood pressure sensors. Often you can find them on flea markets or people throw them away....so it's a good opportunity to reduce the e-waste.
Thank you for the info. Unfortunately, I only have a Prusa FDM printer
@@AndreasSpiess You might want to investigate the process used for "smoothing 3d prints", where the FDM print is exposed to an enclosed atmosphere of vapour that softens the plastic surfaces so top layer flows and fills in gaps.
It would reduce surface imperfections that enable air leaks.
Less messy that painting or dipping parts in resin to seal surfaces. Still have to consider added time and investment (ie: may not meet cheap and convenient requirements)
You can definitely get air tight parts off an fdm printer with some tweaks.
FDM can be watertight, but it takes a lot of work and a lot of fiddling around with settings. It is very difficult though
Good video! I made an almost identical gadget a few years ago to measure the depth of water in the cistern that supplies water to my house. I didn't use a solenoid though, I simply switch the pump on for a few seconds to purge the water from the air line then take the measurement. I have the same issue of the the pressure "reading high" whilst the pump is active, this I presume is due to the back pressure of the air because the line is so thin. I bought the pressure sensor from Digikey that outputs a voltage proportional to the pressure but, like you, used a pump from a blood pressure monitor.
Good to know that the valve is really optional. Thanks!
Love the idea and solutions shown here. The tinkering involved and testing is also great fun. When the problem was posed my mind went immediately to "bobber on string" kind of solutions though lol
Simple weights, a float, strings, and trip switches.... but everything has it's own set of problems to overcome.
Other viewers suggested this solution in my first video and it seems to be used for lakes.
I built from separate parts. I use an aquarium membran air pump powered from mains, switched by a relay which went out of order after the last summer. I want to use an SSR in future.
Your maximum pressure peak comes from a counter pressure wave in the system until the flow in the system has become steady (or is stationary the better word?).
It also appears when you run the system with all the pipes outside the water and on the floor - albeit *much* shorter, by an order of magnitude.
You should do such a unloaded measurement -I had to throttle down my air pump quite a bit, because its air flow was way too much for my aquarium hose diameter and length. Remember, in the blood pressure thingie you have half a meter, in your fluid level measurement setup you have 12 or 20 m of line, and the air inside will have quite some mass to it, as well as friction to the hose walls.
I had quite some difficulties with fluctuation in my values. It improved much with building a silencer between my pump amd my measurement box, and the introduction of a unused cigarette filter right in front of the sensor.
When standing still and valve open, the liquid will rise in the standing pipe, it's displacement at the start of the next cycle will even add to the peak. You can experience the counter pressure first hand (or better first lungs) if you blow your setup with the lungs. It will help you to shift your thinking from "static pressure" to "stationary flow".
The measurement shall be done when the presure value changes have calmed down and become stationary, and the last N samples have been the same or nearly equal.
You can avoid the rising fluid level in the stand pipe with an "lip valve" from silicone. Those are sold as spare parts for the rinse aid dispenser unit of dish washers. it has to be big enough to not introduce counter pressure on its own. I ordered some samples and mounted them and tried them with my mouth.
Also, I plan to use a thin copper pipe as a stand pipe for this season, because I measure the level of an open pond, and I had various problems with biofilms, algae, etc.
If you introduced air distribution valves, you can sequentially measure levels of multiple reservoirs, even on different heights, if you correct for the different ambient pressure at the different heights, in relation to the height of you pressure sensor.
Thank you for sharing your experience. I thought something similar as you describe with the "wave" but was not sure. A good point with the copper pipe against biofilms.
My pump does not move a lot of volume but is able to develop quite some pressure (it has to be higher than blood pressure). But maybe slowing it down would reduce variations. This should be easy with a resistor.
Great video Andreas. I moved our washing machine to our basement and had to pump the dirty water to the ground floor. This pump had to work automatically off course. So I started with a variant based on ultrasonic distance measurement of the fluid hight. That flawed on the foam on top of the dirty liquid. Then placed level sensors, but those had to be in the water and got really dirty and stopped sensing. Then a colleague said that measuring the weight of the tank solves a lot of these problems. So I added a hx711 based scale under the tank and that worked like charm since a year now.
Cool solution with the scale. Beekeepers mentioned that most scales have some creep over time. So pay attention to calibrate it from time to time...
Great! I've implemented this idea with absolute sensor and atmospheric pressure sensor to measure water level in my well. Make the tubing air tight was a challenge, but finally conical thread and polypropylene tubes solved the problem. Another issue was to find an equation between pressure sensor readings and the actual water level. I've declined to recall hydrostatic lessons and acted as an engineer: just pushed the tube with markings each 50cm and thus got a calibration table. The final accuracy is around 50cm, and it's fine for me. I'm re-filling the air and re-calibrating the system each spring.
Good information. Thank you!
the pressure in the plastic pipe decreasing after reaching a max (i.e. the moment when the clear tube is just about purged but no air has left yet) is most likely due to the pipe getting filled with a mixture of air and water. the average density of this mix is considerably lower as of pure water, hance your readings are decreasing
Aha. Thanks for the explanation!
Two thoughts came to mind - kicking a car tyre to see how much pressure it has in it (higher pitch = more pressure), and blowing across the top of a bottle which produces a different note (higher pitch = less air in the bottle therefore higher liquid level). I wonder whether there is an acoustic solution to this? You could rig a speaker/microphone to the outside of the tank, pulse some sound (probably just a click) then measure the pitch of the returning sound? Would probably need a lot of tuning to the individual circumstance but perhaps some machine learning could do that, or failing that just test the tank at various levels to obtain the values of the return signal at each level. I'm sure it's not easy and may not be very accurate, but it is at least external, no moving parts and quite low power.
I bet braces, welds and other artifacts in the tank beeing above or below liquid surface distort the sound, making the calibration a huge job in itself. but for a small homogenous tank it's a really interesting idea that might be genious in that rare instance. Thanks for the Input.
Ian, acoustic measurement , an interesting concept.
It exists. They measure the resonance frequency of the remaining volume above the liquid. This principle was discussed in the comments of my last video about the topic.
Great videos Andeas, you put in a lot of effort to create content, thank you! As you are aware, that design may be a little over-engineered, especially if you have a limited power supply source to drive the pump and valve coil. A bit more info about the common pressure sensors - you get absolute and relative pressure. The one with the holes are relative pressure sensors and they take your altitude into consideration. An error in measurement can occur if you do not calibrate your sensor at full scale at your location. Not a big error, but a Swiss in the mountains will have a different barometer reading to someone at the coast. Additionally, just the sensor with a sealed membrane close to the bottom of your tank is sufficient. Tap and thread a fitting (t-piece in the drain socket perhaps) , connect your sensor orifice, calibrate your reading with a full tank and you are done. Low power and easy to setup. Error percentages you need to be careful with and it depends on final accuracy required. 1% error does not seem much until you have a 100kl tank (or bigger). Once again, if you need more info I will be happy to share.
I agree. We discussed these "throw-in" sensors in my last video about the topic. The seem to have other disadvantages if I believe the comments.
This is an industrial type sensor, sometimes it's easier, quicker, and probably cheaper in the long run. I have been doing this for a while in the automation world. You could use the 0-10V output version, that would ensure long battery life as you only need to sample once per hour, depending on fill/drain rates and size of tank.
Thank you, great video. Really interesting that you explain the principles, the shortcomings of what we can source cheaply and then go on to build your own sensor
You are welcome!
I've had a throw-in type pressure sensor at the bottom of my rain tank for about a year now which had been working great but lately the values have been going up and down rather randomly. Quite frustrating. I have the same sensor in my pond and it doesn't have any problem so far :-\I've had a throw-in type pressure sensor at the bottom of my rain tank for about a year now which had been working great but lately the values have been going up and down rather randomly. Quite frustrating. I have the same sensor in my pond and it doesn't have any problem so far :-\
I am working in something like that, but for a pretty simple pump system. I started with the same idea, but dealing with the fail of air tighteness... just a fish tank air pump or something similar...thanks for your work, that was a nice idea.
Glad it was helpful!
Tnx for a nice video and for making a follow up on previous video based on experiences and feedback from the community
My pleasure!
You amazed us all again my Swiss Genius!
It was the idea of my viewers!
I would love to see a follow-up with the implementation and some testing :)
We will see if it will be worthwhile.
Love your channel. I'd like to add to the discussion. Your readings are fluctuating because
1- The pressure pulses of the pump is not dampened,
2- Large bubbles are created because the air flow is too high.
3- The shape of the outlet of the tubing is not optimal.
You could dampen the pump pressure pulses by using a small air camber reservoir to dampened it. Adding an adjustable flow meter on the outlet of that reservoir to limit the air flow or install a fixed orifice to reduce the flow would also help stabilize your readings. You only need a very minute amount of flow for bubbles to flow out the tubing but enough flow to empty the tube of water in a reasonable time at startup.
I suggest cutting the end of the tube in a 45 degree angle. It makes a huge difference as it allows the bubbles to flow out the tube easier with less resistance and more uniform bubble size and consistency. Cutting it flat causes that larger concave bubble shape to form at the end of the tubing just before the bubble breaks free.
Also Using a 45 cut on the outlet tubing you can butt it on the bottom of the tank while a flat cut tubing outlet cannot.
Also another concern I would have using a flexible tubing is it rigid enough to stay in place at the bottom of the tank while in use? When the tubing is full of air, the tubing will have a tendency to lift off the bottom of the tank. Its better to use a small rigid pipe like PVC cut on a 45 degree angle. The tubing outlet must not be allowed to move and will not be reliable if not secured. Be aware for high precision the tubing outlet distance above the bottom of the tank must be subtracted from your overall calculated height range of the tank.
Thank you for your explanations. The fluctuation is not a big issue because I measure when the pump stopped. The other values were measured for learning only. And I agree: You have to attach a weight to the tube to keep it at the ground level (was not needed in my narrow "tank")
Thanks for this! We have an old underground rainwater cistern that I've never been able to work out measuring electronically. This is a really neat method that might actually work! Now I just need to figure out how wide the cistern is...
Or you display percentages...
@@AndreasSpiess Great point! You are the master engineer, for sure.
Perfect! I build a big water storage last weekend. And we discussed how to measure the water level. Problem solved. :-) In the most complicated way....
So enjoy your new project!
Thanks for the info and references on these cheap pressure sensors.
Have been toying with idea of making a solid state anemometer (without moving parts) to measure wind. Using 3-4 such pressure sensors placed in circular pattern to measure wind pressure, and thus its speed and direction.
I read about this principle, but never tried it...
Great project, I enjoy your thorough investigative thought processes and low cost design approach! Well done, thanks for sharing your knowledge ,Andreas.
Glad you enjoyed it!
Great hack. I had one of those pressure sensor which I bought from Aliexpress to replace the one in my washing machine, and has always been wondering what to do with them. Now I know. Thanks very much
You are welcome!
You are testing at zerro airspeed, which is only good for hovering. Maybe good to just go fly, try different configurations, see what works best.You are testing at zerro airspeed, which is only good for hovering. Maybe good to just go fly, try different configurations, see what works best.
@@Podemosllegaralossubs-ty7bq ???
In the industrial environment, one usually does not interrupt the pump. Instead, the flow is used through a volume with a downstream throttle to minimize the influence of back pressure with a minimum amount of air into the fluid. In German it is called Einperlmethode. Greetings from Hamburg
Interesting. Maybe I try to reduce the size of the tube at the end.
Very cool project Hope you make More Lora Sensor
Thank you!
Another amaizing video! 😍
Thank you!
You always have such fascinating projects.
Thank you!
I use MP3v5010DP for my 150l tank for now. It is basically the same as MPX5010DP, but rated at 3.3V. ESP32 with lut-table or ADS1115 with it works somewhat good. I really like to try HX711 Wheatstone Bridge setup with four 4-wire cell type strain gates, because Tasmota HX711 driver has just got major fixes in the Tasmota dev build.
The HX711 is a marvellous chip. And extremely cheap.
Very interesting. Always great to see alternative uses of devices and giving them a different life, well done
Glad you enjoyed it!
This is almost exactly what I intend to do with my friend's well holding tank. Great Video Thanks.
I hope it will work for you, too!
@@AndreasSpiess
I'm pretty confident. I made one years ago for a wastewater treatment operation. Very corrosive liquid.
Love the idea and solutions shown here. The tinkering involved and testing is also great fun. When the problem was posed my mind went immediately to "bobber on string" kind of solutions though lol
Simple weights, a float, strings, and trip switches.... but everything has it's own set of problems to overcome.Love the idea and solutions shown here. The tinkering involved and testing is also great fun. When the problem was posed my mind went immediately to "bobber on string" kind of solutions though lol
Simple weights, a float, strings, and trip switches.... but everything has it's own set of problems to overcome.
As a sugest, it was interesting to do a new Level Meter using ultrasonic, for this it can be used simples electret microphones and earphones as ultrasonic source.
You are right. These kind of sensors were mentioned in my last video about the topic.
The problem with this is frequently water thanks are going to be exceptionally high in humidity. I've killed more sensors than I can count trying to use them. I've tried coating then, potting everything with epoxy, you name it, and eventually the ultrasonic sensor itself rusts and the system fails. I've been watching this topic on the channel while also searching for my own solutions... This once seems like it might FINALLY be a workable one paired with a solar panel and LoRa.
@@DasWookie We are proposing a video camera and AI to measure levels. But the problem with humidity will exist.
Einperlmessung 😊
twenty years ago a E+H PIR could be configured to be used as a local display for the level (can‘t recall if volume too) and present the level to a DCS in the form of an intrinsically safe current loop.
Happy to see this can be done at home for a small investment minus the great instruments and electromechanics of those transmitters.
Andreas, I am not sure if this is mentioned in the „Dubbel“ (I‘m not joking! it‘s one of the books I cherish the most!!) too but it is a source of inspiration for anyone interested in machines in general
These instruments aren’t used for precise level (and, if calibrated: volume) measurements but for rough estimation and then charging a process via a mass flow FQIC but often thereafter weighed in a buffer tank on a load scale.
Temperature ranges in tank farms are quite broad (50C nothing out of the usual). Level indicators like this can be in the +/- 5% range due to the effects of temperature differences on the liquid in the tank.
Hope it works out for you - maybe use a very fine needle and a steady stream of gas for more precise measurements (we used N2 or Ar)
Since I am an electronic engineer, I do not use the "Dubbel". So I do not know ;-)
Blood Pressure - great idea
:-)
Nice,. From my expirience, temperature coff is much higher then in spacifications. In order to realy use this simple sensor, you have to add temperature measurment and software for compansation.
Thank you for the info!
Great project 👍
Great walkthrough of the process 👍
Thanks for sharing your experience with all of us 👍😀
This was a truly viewer-induced video. I did not know it before...
VERY cool!! Thank you for all your hard work!
My pleasure!
Bubbler tube output is very sensitive to air flow, the commercial units usually have a air flow controller of some sort. Either a variable-area flowmeter (rotameter) or something similar. As level varies, pressure varies, so flow varies. Lots of variables to think about, but they are still very popular in special applications. Corrosive, flammable, or explosive fluids, and also where accumulation of solids are a problem, like sewage and sugar wort.
Thank you for the additional info.
Andreas, thanks a lot for your educational videos, that helps to create a truly engineering mindset. Regarding this problem, I would have used a sumergible pressure sensors, they are cheap, with for example 4-20mA output
This one was covered in the last video…
I'm showing my age, but I've worked on a foxboro model 40 pneumatic PID controller with a bubbler level measurement converting the 2m tank level via a pneumatic scaler to the 3..15psi signal the foxboro needs. As I'm writing, the details are coming back to me. It was two model 40s in cascade, with level controller output being setpoint of a flow controller connected to a Fisher proportional control valve. All pneumatic, no electrics whatsoever.
We all have our ages. But with age also comes experience. Not bad, I think (I am 65)
Thanks. Great idea for my camper van.
:-)
Thank you!
You're welcome!
Very interesting project! Could be a good solution for a rain water tank under the ground that is hard to reach and only accessible from the top. So a range sensor or this could be used. One thing I wanted to say I have printed air and water tight parts before. Of course I do not know to what pressures they remain water tight. I printed many flower pots over the years. The design is very important to be watertight. You can also line the walls with super glue to fill holes if there are any.
It seems that airtight is much harder than watertight. At least I was not successful with airtight :-(
The range sensors were part of my former video. They tend to corrode.
Thanks Andreas, I have two 5000 litre rain water tanks that I’d like to monitor. I do think though that there is a simpler and cheaper solution than the blood pressure monitor. I don’t know about Switzerland but here in Australia garden shops sell parts for small level irrigation. One item is a “T” piece made of ABS and capable of withstanding mains water pressure. Each leg is barbed and fits into appropriate size tube, with the addition of a small cable tie, it should be air tight. Add a small cheap battery air pump and you have an equivalent system.
Others suggested to visit a pet shop because they use these pumps, etc for aquariums. So we already have three sources...
If you want to go 100% analog and 0% electronics, you could even do it with a small air compressor (aquarium ones can “pump” up to 1m depth, medical nebulizers can do 5-10m of water column (1atm = 10m w.c. = 14.7psi) and those cheap tire inflators go to 20~40m c.w.! So, if you “pump” air into the tube, then monitor the air pressure (pressure gauge, or U-Tube manometer), you have your level reading, electronic-less - or use the electronic sensors as Andreas introduced.
Oh! I think I know why there is some drift in the pressure reading. The temperature of the water is typically lower than the ambient temperature. So when you pump the air into the tubing, it becomes colder and loses pressure. So it would be best to read the pressure immediately after the air pump stops.
Thank you for your clarification!
It you want to use cheap air-tight and small components, there are a lot of options in drop-water agriculture/gardening.
Thanks. Others also suggested pet shops (aquarium stuff). So we have choice!
Great content as always Andreas! A quick off-topic question: What labels are you using to label your project box? My labels don't stick on the boxes...
E Gruess us Muttenz
I use Brother knock-offs from Aliexpress. They stick very well.
I am working on something similar with a better sensor (differential pressure sensor from All Sensors, ELVH series). The nice thing is the sensor includes the ADC, but it also costs an order of magnitude more than the cheap sensors. In terms of the connections, bulkhead fittings with small hose barbs or Luer fittings are relatively inexpensive.
Thank you for sharing. Another viewer uses the cheap BME280 which is I2C. So many possibilities seem to exist.
nice. A peristaltic pump and part for a small scale irrigation system from the gardening senter should work. There you have Tee's and valves. On offshore oil reservoir, there are sometimes huge concrete tanks at the bottom for local storage of crude. These tanks are open at the bottom so water flows in and lift oil up to the inside ceiling. The lower density of oil gives an internal pressure in this tank and that pressure is proportional with oil content. There is on some, no need for a pump to lift the oil up to the ship, due to this pressure. Just open valve and the oil comes gushing out. These tanks are near maintenance free and all instrumentation is at the outlet of the offloading hose. So the principle is similar.
Cool! We are only oil users where I live. So I did not know of these devices.
The pressure should go up as you add air to force water out of the measuring tube. It has to be greater than the pressure of the water in the tank. Also have to take into account that air is compressible. So, probably not 100% accurate. Perhaps Nitrogen. But who's going to do that!!! Once you get air "bubbling" then you can cut off the motor and measure the pressureThe pressure should go up as you add air to force water out of the measuring tube. It has to be greater than the pressure of the water in the tank. Also have to take into account that air is compressible. So, probably not 100% accurate. Perhaps Nitrogen. But who's going to do that!!! Once you get air "bubbling" then you can cut off the motor and measure the pressure
Hi Andreas, in fact some years ago I did have something similar in mind but did not build it for several reasons. One was the non linearity as we have a uncompressable fluid and a compressable gas that do have non linear beahviour. However this could be corrected in software.
in fact I did choose to use a capacitive sensor instead. Basically two isolated wires that do nowhere have contact to the media. These two wires act as an tunable capacitor for a NE555 based frequency generator that does change value linear with the coverage of media. So in the end I simply measure the frequency which is a proportional value to the coverage with the measured media. Actually I created an dedicate PCB with an Atmel ATMEGA 328 ( yes, it's an specialized Arduino clone in the end :) ) for that to be placed inside the cistern.
Ä gieslie vum Südweschd ecke vu Dütschland.
Thank you for the feedback about the capacitive method (I mentioned it in my last video about the topic). Good to have an alternative, I think.
Nice engineering thinking
:-)
Your setup with the tube looks exactly like an experiment I do with my students, but I went only for one meter long pipe and we vary the depth by lifting out the tube instead of changing the water level. OK, and we don't use the pump-solution but only the pressure sensor.
I also changed the inserted length of the tube (I have no possibility to check the water level inside the "tank". I just made sure that it was always completely filled.
Actually there is a way. You can have an external water column connecting the top and the bottom with some watertight fittings and a see through hose or pipe. It’s not easy to achieve in such a small diameter as the one you used, but in a normal tank it’s a very common solution and one that can at least get you in the right ballpark. If all that’s needed is to trigger the pump on and off, then it’s pretty straight forward. Most applications don’t require the volume of liquid contained in the tank thus making it sufficient for the majority of cases.
There is a way to make 3d printed parts pretty airtight and that is to paint them with plastidip.. that will do the trick as long as it is only between the outside and the insides... if it needs to separate 2 chambers in it... well that is more difficult but for infusion and fishtank supplies you can buy these kinds of joints and T-joints and taps.
Good idea with plastidip!
Very good Andreas, but may I suggest a longer delay before taking the reading to allow the membrane to achieve a stable state.
I installed a cheap chinese screw-in pressure sensor on the suction line between my underground water tank and adjacent underground pump, approximately 1.2m static hydraulic head. No atmospheric pressure compensation as yet.
This was run in parallel with an ultrasonic sensor which was prone to forming a condensation drip on the transducer face in sub-zeroes. Result - The US failed at -10, the pressure sensor continued reading.
Although I suppressed readings when the pump was running, what I had not considered was membrane distortion under vacuum (when the pump was running), and observed 20-30 minutes before the reading stabilised. Readings have drifted with such repeated distortions over time, plus of course atmospheric pressure effects.
I will reconfigure the arrangement in summer during the annual tank clean out.
You will not be abusing the membrane to that degree by blowing bubbles, but it will alter the SG until the air bubbles disperse from your column.
If you check how soon the sensor takes a stable and consistent reading, you can set your delay to suit.
I feared to lose pressure if I waited too long. But I will experiment with the delay in the future. Thank you!
@@AndreasSpiess If the sensor pipe is airtight you cannot lose pressure, it will stabilise as air and water reach equilibrium.
For a 3m pipe, 5 or 10 secs should be more than enough for the bubbles to clear - You will lose a small hydraulic head as the air displaces water but then disperses but it will probably be a constant.
OT - In the event your water tank is groundwater fed with a high soluble iron level - I made a venturi aerator with a small submersible 100w water pump and installed it in my tank last year.
No more soluble iron.
making videos under (controllable) pressure😂 Thanks for this contribution!
Exactly! With relief every Sunday morning at 9AM ;-)
Page 1:08 demonstrate an intermediate remote air pressure sensing next to the local air pressure scheme. Both exhibits an offset error value in proportion to temperature at calibration. To remove the -ve offset ambient air is injected into the detection hose by the input port until reading settle to a steady state. That should be repeated regularly no longer than temperature cycles, or continuously as long as valid reading is desired. Air tight? Unnecessary. Just keep the top of liquid at ambient value and remember to keep refresh the air pressure in the air portion of sensing tube.
Thank you for the additional information!
Great video.I jave the same requirement to measure the level in an underground rain water tank.
Is it possible that the pressure rise during pumping is due to friction as the air and water move in the pipe, once the flow stops, there will be no differential as there is no flow. For the saturation problem, could a series resistor with the bridge drive voltage reduce the output (or parallel if current exitation).
Of course you could try to add a voltage divider. But you have to be cautious because it influences the resistors built into the sensor. Changing the HX710 seems to be a better solution for me.
You mention maintaining air tightness is difficult but what about all that pneumatic equipment that exists? That has to be air tight. Relying on a pump isn't ideal either. It can break and then you have to replace it, which is difficult in remote areas. What if water enters the pump and destroys it?
I agree. Both possibilities have advantages and disadvantages. Others mentioned that the pump approach removes clogging while the "passive" approach tends to clog with algae etc.
Dear Andreas, very interesting video. If it can help, an « easy » way to have a airtight container for measuring the pressure is to use a glass container (a bottle for example) and turn it upside down. You can then put the sensor inside the container and plunge it into water with the wires passing through the top of the container to the surface of the tank. In this case, sensor is in a water bubble and is able to measure the pressure inside the bubble with a linear correspondance to the water level of the tank (hope it is clear ;)). Other point, i also used a BME280 which is very accurate (detection of a difference of 5mm in the level ! But unfortunately with a limited range of measure…). Thank you for your great work
What an excellent idea Christophe!
@@BenMitro Thank you! I’ve forgotten to mention that the sensor should be waterproof because there will a lot of humidity. I used a bme280 that I have "tropicalised" with some coating in spray. This has obviously no influence on the pressure measures as the coating is flexible. I have built this for measuring the water level of my pool. It worked with an unexpected accuracy (5mm) but unfortunately, after a few months, the sensor broke (I have tried 3 times…). I have installed this assembling on the circuitry of the pool in the technical room. But, unfortunately this leads to pressures around 1450mbar when the pump is on due to the pressure linked to the water flow. Even if the bme seems to be able to measure this level of pressure, it is outside of its working range. My analysis is that this is the reason why the sensor goes out of order. For the moment, I haven’t found a sensor with the same accuracy, ease of use but working at a higher level of pressure and I’m stuck. Does somebody has any idea ?
Good idea! Another viewer also suggested to use the BME280
@@christopherisse6857 Christophe, do you use that spray on circuit sealing compound of just clear spray? I need to do this to some gear I am putting in my camper that is not intended for such environments. The specialist spray is expensive and I have a lot to do.
@@BenMitro Hi Benny, not sure to rightly understand your question. The main circuit with an ESP32 is outside the glass container and not in contact with humidity and doesn't need any specific coating, just to be in a waterproof box if you want to use it outdoors. The only part I coated is the sensor (BME280) itself. I linked it to a connector for the I2C bus and I coated the sensor + the connector + the base of the 4 wires. Hope it answers your question
Thank you for sharing and helping fot the improvement of humans!
My pleasure!
After many abortive attempts at measuring my well depth over the years, I finally explored the 'air line' method, which is well known and as you discuss. You need not have upset your wife in using the blood pressure monitor! It'll do the job without modification. The first phase of blood pressure measurement is inflating the sleeve. The monitors I have show the inflation pressure as it progresses. In this application it'll fail to achieve target and report an error. But the pressure figure it give before failure is accurate in mmHg. So there you have it - a ready-made and cheap liquid level meter. Blood pressure monitor Inflation stops at about 0.3bar so the method is limited to 3m depth of water.
Cool! I did not think of this method (probably because I always wanted to have the values in a computer-readable form).
You may find some of the variation in readings is caused by the flexible tube moving upwards in your pipe once you pump it full of air - maybe a rigid tube would provide more consistent readings.
A rigid tube is for sure a good solution. Most people add a weight to the bottom of the flexible tube.
10:57 how do you know the loose end is not rising inside the pipe? The hose is light enough...
In the industry, the vessel has a tap at the side, near bottom, if the tank is atmospheric, or at the side near bottom and top if it's pressurized. That takes care of holding the measurement points at the same height. Everything bellow the lower point can't be measured.
You don't normally install at the bottom because that is prone to blocking.
In a tank you would have to attach a weight to the tube. In my small experiment this was not needed.
This video reminded me very much of my washing machine. About 6 mos ago I removed the control board to replace some bad capacitors and it used a sensor almost exactly like this one only twice the size and DIP-8. It was on the control board. The sensor had a built in analog amplifier 5V in and 0-4V signal out, rated for 6kPa (ADP51B62A01) price is about $30. It has the same tube to the bottom of the tub for water level sensing, but since it is drained for every load it doesn't need the air pump. The water level is measured from the bottom of the tube, not the tub. I did some math at the time, changing the air volume in the tube changes the sensor calibration. Adding more air volume should cause the tub to add more water to reach the same pressure setting. Just theory though, I haven't tried it.
I did not know that some washing machines use this method. Other viewers mentioned that they use simple on-off "switches"
@@AndreasSpiess I started thinking about my washing machine again and of course it is a little different from your project. It occurred to me though that with the air pump, your goal is to fill the tube from the sensor completely full of air (not allowing any liquid to rise in the tube). If you used a small second tube from the pump that went to the bottom of the sensor tube you could just release air bubbles into the opening at the bottom to fill it. By doing that, the valve would not be required (but not as much fun). Also, the sensor tube should be large enough where the bubbles couldn't lift any water up to the sensor as pumping the air in from the top helps to keep the sensor dry. Just thinking.
Thanks, very interesting
Glad you enjoyed it!
When you release the pressure of the pump, be careful momentum doesn't cause the upward rush of water to enter the air pump.
Thanks for the hint!
For the variability in readings I am thinking either the water level during the tare phase is bouncing up and down in the measurement tube or the tube itself is collapsing after the purge and thus not a constant volume.
Well possible. I do not see the movement of the water inside the tube because my "tank" is so narrow.
Now I have sold our acreage this would have been handy for the water tank I had placed up in the forest behind.
Maybe the buyer is also a subscriber ;-)
on cargo ships, the ballast tanks are usually monitored with pressure sensors but in my experience they were often unreliable and required constant maintenance.
Thank you for sharing your experience.
I have a new episode suggestion: lets hack a water tower and turn it into a blood pressure monitor. Sounds like a fun project!
:-))
I've had good luck printing airtight parts with a resin printer. Most resin printed parts are very brittle so I use non-brittle engineering resin from Siraya Tech. BTW I also use 100% infill.
Thank you for the info. Resin seems to be better in this respect. Others also mentioned it. Unfortunately, I have no such printer :-(
Love this hack. Hope maybe a future stable version :)
:-)
This reminds me of a discussion of how to make sure a radon eliminator was active. The indicator can tell if the fan is working using a pressure differential. Sounds familiar.
I never heard of "radon eliminators" ;-)
@@AndreasSpiess A better word might be exhaust. It's a continuous running fan that moves air from a basement to the outside. It creates a vacuum and you can see it's working because they use a tube of liquid with marks on it. The vacuum that sucks the radon out also moves the liquid. We had a conversation on Reddit how to monitor and alert users of failure. I've given up posting links because RUclips deletes them but you can Google
"Looking for a way to get a notification if my radon mitigation system loses pressure"
Very interesting .... and original solution..... I built my own DIY sensor by using a air absolute pressure sensor jailed into a completly sealed non circular plastic bottle (dishes bottle for example) filled with 1/3 of concrete (to let the bottle layed on bottom of my well) and 2/3 of comestible oil (I used oliver oil). Of course in order to report the air pressure, you need a small tube attached to the sensor, filled with air and dived into the oil. The sensor is jailed into the bottle capsule everything is waterproof with epoxy. On top of my well I got another air pressure in order to have differential pressure measurements. It works greats since 2 years from now..... My well is 20m depth..., reading analog value with an ESP32 w/ESPhome.
Cool! Good to know that you can get an airtight system for a long time. As I said, I feared it and was happy that the pump removes this need.
@@AndreasSpiess Yes I like your approach a lot for this point.
Have you done a video of measuring rain fall and the amount per hour, per day etc ? Looking at using a sensor that does this and solves the cleaning of the sensor after a certain time, to stop it getting dirty or "Clogged up". Thanks for such great videos, which have really helped me with my IOT weather projects.
No video. I purchased a Chinese weather station for this purpose...
Fun and educational - Thanks!
Glad you enjoyed it!
I wonder if a rigid bubbler tube would reduce the variation that you see when you're pushing air into the tube?
Or, alternately, if a smaller diameter air tube would show a different (reduced) variation at mid reading - if I understand what your statement was.
Also, with a flexible tube, unconstrained in a large tank, will boyancy be a problem? Pushing air in may make the tube light enough that the end floats toward the surface.
You are right. You either have to put weight on the end of the flexible tube or use a rigid one.
Thank You for your work, i watch Your content with pleasure. Is there a place were awiring diagram would be presented?
All links are in the description. So, if there is no link, I did not provide this info :-(
thank you
You are welcome!
I’ve found 3d printed parts with more than three walls to be airtight. More than airtight, in fact. They hold up well to 3 bar pressure(29PSI differential)
Lol, what is more than airtight?!
@@TheDigitChannel airtight typically refers to holding a vacuum i.e. 14.5PSI differential pressure. Very few things are airtight at 1000PSI, lol.
So you had more luck than me.
It was interesting :-).
During my working life - quite a while ago now - I leak tested plastic containers but the test pressure involved was very low, it is easier like that. So anything involving pressure testing is interesting to me.
I believe btw that Washing Machines use the pipe filling principle to control the water level in the machine. I have no idea of the details.
In the last video about the topic these sensors were mentioned, too. I never used one.
@@AndreasSpiess in fact they are simpler pressure switches - a diaphragm pushing electromechanical contacts - just like a bimetal thermostat, but here using diaphragm as force source.
10:50 i guess the hose is to wide so the pump does not pushes all the water out, because the airflow is to low and water passes up while pumping... the fluctuating values might be a result of the hose beeing flexible the the airbubbles get everything ocsillating.
i would recomend to use a very little diameter for the measuring stick (maybe 1mm brass tube)
I agree that the diameter matters. However, I thought the pump moved the air/water quite fast. It only took a few seconds to get the water to the top of the tank. Others suggested adding a reduction at the end. I will do some experiments...
@@AndreasSpiess I answered few moments ago, on the main post. Here, and based on my past explanation there, I would summarize: it will WORSEN the accuracy and intermediate pressure buildup problem is you create a localized element to create a pressure drop at the end of the tube.
Another way of measuring fluid level would be to use an ultrasonic distance sensor somewhere above the top level of the fluid. There has to be an air gap between the maximum level of the tank and the sensor though (this is because whether you use a ready-made module or roll your own sensor with ultrasonic transducers and MCU,
the pulse you emit has to finish before you begin to listen for the return echo).
Pro's:
No need to find airtight tubing and suitable pumps.
Transducers which are air- and watertight are available.
Instantaneous measurements (you could technically be able to follow the sensor in real time).
Con's:
Constraints as to the placing of the transducers relative to the walls of the container whose fluid level you wish to measure.
Minimum gap between maximum fluid level and sensor.
Having to delve into the dark arts of digital signal processing.
Driving the transmitting transducer.
Amplifying the signal from the receiving transducer.
Temperature, air pressure and humidity compensation for accurate measurements.
You are right. These sensors were discussed in the video I mentioned.
The biggest problem in a water tank is high humidity. After a short time, water droplets form on the sensor and massively falsify the result. After numerous disappointing attempts I also switched to a pneumatic sensor and am very happy with this solution. What would make me even happier would be a transmission of the measured values via LoRa. So I will follow this project closely.
@@HansPirngruber Depends on the sensor shape and what if any signal processing one is doing.
What you choose in the end depends on what applications a specific sensor is most suitable for.
Was just posting it for completeness of the methods available for fluid level measurement :)
great video thanks 👍
just wondering now how they might be doing it in commercial heat exchanger systems. if it is the same method
I do not know which method the use.
Next It would be great if you could build a capacitive level sensor, the prop itself is easy to build, I did it using two sheets of aluminum foil 2cm wide next to each other and laminated, its resolution was about 1mm and seems to give accurate results, the hard part is building the capacitive to voltage converter, I still couldn't do it with simple available parts with no arduino.
For my problem I will go with the pipe because the tanks are quite deep. But your proposal for sure is also good. The capacitive humidity sensors use a NE555 to measure capacitance. I did a video about it.
Def Great! I've implemented this idea with absolute sensor and atmospheric pressure sensor to measure water level in my well. Make the tubing air tight was a challenge, but finally conical thread and polypropylene tubes solved the problem. Another issue was to find an equation between pressure sensor readings and the actual water level. I've declined to recall hydrostatic lessons and acted as an engineer: just pushed the tube with markings each 50cm and thus got a calibration table. The final accuracy is around 50cm, and it's fine for me. I'm re-filling the air and re-calibrating the system each spring.
Very interesting setup, Andreas! A useful warning about the HX710 module which - if I understood your explanation right - means it's been engineered close to uselessness. Aside of the fun building this I wonder if a VL53L1X ToF sensor would be a simpler approach? I found them to be very precise. That with a LoRa module?
A ToF sensor had to be inside the tank which would sooner or later lead to corrosion. The advantage of this approach is that no electrical part is inside the tank.
I agree with Klaus. ToF and also ultrasound sensors work fine, but can corrode or be clogged by water or dirt.
At least here in Germany many older oil tanks are equipped with the tubing inside allowing this air bubble method. Decades ago (but also up to now), mechanical pressure meters were available to guess the inventory. But at least according to my experience it was far more or a guess than a measurement. Using modern electronics instead of the mechanical meters, I achieved measuring errors < 0.1%
Even the restrictive German TUV had nothing to complain. All electronis is located outside of the tank. No cables inside the tank needed.
Can you try an Archimedic method?
JustUst measure the change of weight of a floating cylinder, again with a 711.
Another way i was considering was with a float, that pops out of the tank and has a lidar on top. The lidar reads the dostance to the roof.
In both cases, I end up worrying about friction, bacteria, slime etc... and not doing anything :)
Many other methods were discussed in the first video about this topic.
It looks like there was moisture in the small plastic tube that goes back to the pressure sensor. If that moisture drips into the actual sensor it will distort the sensors readings.
I can imagine that moister is not good for the sensor. In my case the tube was long enough that it did not go back. And I would hope that regular measurements will dry the tube at the top.
And you probably need to add a weight to the bottom of the measuring tube. the test tube probably was too narrow to have the measurement tube float to the surface.
Yes, you are right. In my small "tank" the tube had no other way to move.
Many people dispose of portable BP Monitors because the cuffs start degrading/leaking. I must remember the good salvage parts inside! Would the pressure sensor from the Blood Pressure PCB be more accurate, or have a more suitable scale/resolution/output?
I assume the sensor inside the device has a similar range. But I did not desolder it to find out which part they used...
Great video. Thanks
You are welcome!
Great idea. Do you think an HC-SR04 Ultrasonic Module or JSN-SR0T4-2.0 ultrasonic ranging module could be used for fluid measurement? The cheap level sensors used in travel trailers go bad almost immediately and I was wondering if such a module could be used. It would be great if they could be used attached to the tank externally but that would only work if the signal can pass through the plastic tank.
Maybe you watch my other video about the topic an read the comments. There we talked about all other sensors.
Fabulous!
:-)
10:44 for the next Experiments in this direction I'd say you could add a funnel to the top of the pipe for the bubbles to get out of the water easier, with the increased surface the water would rise less.
At 10:56 my guess is at the surface tension of the water, or because the air in the water directly above pushes the water up and therefore slightly lightens the weight?
Do you think that pressure change could be useable to run the pump as short as possible? Especially for remote stuff every mA may make a difference depending on the amount of measurements. (Ok, that amount may be adjustable depending on the use of the fluid and in case of water the weather. If it's rainy and you don't Pump water for the plants you probably also don't need to update the tank to often to check if it's empty.
Significant pressure drop occurs when water is displaced and bubbles start to exit. You can feel it if you blow through a hose into a bucket. I would guess that it might be related to the fluid not being compressible but the gas is???? Similar to the way break-through pressure works on a saturated filter, once break through occurs, lower pressure is required to maintain flow.
Good ideas here. I will try it out in our bath where I can see what happens...
👍
:-)