Before the company changed my meter to a digital one I had a small reed contact mounted in my meter. Every half liter a pulse. Used the Electric imp(001) to count the pulses. Pushed the data in the mysql db and used AMcharts to visualise everything. Brings back memories. 🙂
@@greatscottlab haha indeed. You can recognise patterns after a while. When my wife took a shower after her night shift at the hospital, or our diswasher, ..
Version two: You set a baseline temperature and it automatically dials up the cold water mix percentage as you get closer to the self-defined cutoff time/usage amount.
Just use cold water, shower time will decrease automatically and therefore monetary spending will decrease as well. I can’t applaud enough for your ingenuity, great work 👏 A budding UA-camr.
@@ErebosGR Just make sure you don't breathe all the water in from the involuntary gasping that will occur from cold shock. I've had too many falls and close calls from that due to being dizzy and light headed, it's just a bad idea.
If you're relying on constant water pressure and flow from your shower for an accurate reading, then you may as well skip the flow meter and have just flow switch and use a timer on the mcu to time how long the water is on for. In fact, just put a waterproof clock in your shower.
I was thinking the same. This is an over engineered solution for something that doesn’t actually save you money. A clock or timer would achieve the same (without the data,) cost you less time and money and have value outside of showering.
You should really mark `pulseCount` as volatile since you are updating that value in interrupt so that compiler does special optimizations for it and you avoid some of the really hard to debug issues. I'd also implement sleep mode which Atmel supports. In that mode its power consumption is really really low around 0.1 micro-amp and interrupt can wake it up. It's simple to implement and makes rest of the code even easier to implement since you'd just have to upload stuff, call sleep and be done with it.
Good idea, gets rid of the bulky float switch. Also I would add price calculation for hot water, since you probably don't use cold water for showers, and the bulk of the cost comes from heating the water with electricity or natural gas, not the water itself
@@ffoska This. Though the heat may come off as an estimate since it takes time to run the hot water through or heat it up. You could probably do an estimate based on the temperature.
Trouble is, it's ESP8266 in the end which has a different idea about deep sleep and interrupts, but yeah, it is possible to solve this problem without an extra water switch.
@@Gurux13 I’m not so sure that ESPHome - which integrates nicely with Home Assistant - implements deep sleep very well. Maybe I’ve just missed it so far.
In defence of the cheap flow meter, I suspect that higher end devices do exactly the same thing but have a calibration in their datasheets. The problem is that you aren't directly measuring flow rate, you're measuring rpm as a proxy for flow rate, and since it's an impeller with gaps around the blades a higher flow rate will change the relationship between flow and speed (guessing here but presumably higher flow rates didn't increase speed by as much, would expect to see bypassing to an extent/less efficient operation). Intuitively it makes sense that pulses should match up with fixed volumes but once you lose the assumed linear relationship that goes out the window.
There are other types of more exact flow meters on the market. For example, one working principle is measuring the difference in the reflection of ultrasound through flowing media. Or you can measure the different electromagnetic features of flowing (electrically conductive) media vs non-flowing. These two principles are much more reliable than the little repeller doing its funny thing. They are a bit pricey though.
That should be the manufacturer's job though, to provide the user with a curve/ function or lookup table for the sensor they are selling. It's not that hard to build an experimental setup for getting halfway decent curve, so shame on them for not doing it
The flow meter needs a consistent stream to work - most flow meters use some form of restrictor or other method to remove turbulent flow. It wouldn’t matter if it was ultrasound or conductive or mechanical - the sensor won’t read consistently if the application is correct.
Using a flow meter is still a widely used way to measure volumes in industry. Usually you don't have a curve for the flow/pressure/etc as it depends on your liquid (water vs oil for example) they give the relationship for pulse per turn. Some flow meters do have a relationship for pulses, but those sensors measure an actual volume (ex 3.14ml/pulse). What is done to make accurate and repeatable measures is called a K factore. If you have a constant flow for example coming from a pump you can simply do the calculation at the rated flow, but if the flow or pressure or liquid change, you need a multi K factor compensation. It is done by taking multiple measurements across the flow variation. Then you do a linear interpolation between the K factors. An example would be for example a variable speed pump in an oil pipe and you need the result in kg. You will start by taking the number of pulses and the frequency at 10% speed, and for each 10% up to 120%. You can do it for a couple of different temperature points as well if you have a temp sensor. Then in production if you read a frequency at 63% you interpolate from you 60 and 70%. I don't know if it make sense the way I wrote it, but I do those kind of things regularly where I work. We are able to get below 0.5% error with cheap sensors and this method. Our process requires less than 2% error.
@@philippegauthier2212 you are correct on a number of levels, as are some of the other comments to some degree. High quality flow meters often use a four wheel paddle and hall effect pick up. The issue we have always found is turbulence in the fluid flow. There must be some distance between the entering and leaving ports of the sensor from any kind of bend, valve etc that creates any kind of turbulence
I agree but it's still quite interesting from an outsiders perspective, and he seems a really nice bloke. (First vid I've clicked on from his video list)
I can't belive he is at 1.7M subscribers. I ve been following his channel for years and never been disappointed. He is working soo hard for these videos, and explains things pretty well.
Great project! Two thoughts: 1) it would have been yet more complication in the code, but you could also measure the pulse frequency and use that to adjust the conversion factor based on flow rate. 2) They probably don’t exist as commercial parts, but you could use the water flow to run a small generator to charge the battery whenever the water is running. I think it could easily produce enough power that you’d never need to charge the battery externally. The project was already more than complex enough and works plenty well, so these would be needless additional complications, but I don’t think it would have been *too* difficult to implement a first-order correction for flow rate.
I think #2 is actually nice, although my noob idea was to hook up some alternator to a spinning wheel and charge automatically 😂 wouldn't be very waterproof probably
Apparently you can use MEMS to 'listen' to your water flow. I've seen sensors sold that work non intrusively and I'm guessing this is how they work. One should be able to connect this to their main to detect all water usage even in apartments. I'd love to see a project using MEMS (or even a cheap analog microphone I suppose) in place of a flow meter.
Have you considered using a micro turbine to help with powering the circuit/charging battery? It would probably make this a lot bigger, but it may be interesting how long it would last.
I thought that I was over complicating stuff when it comes to home automation, but you sir, have beaten me. I also measure water usage, but I added a proximity sensor to the outside of my water meter. My water meter has a small metal piece that rotates and passes by when one liter has been used. The proximity sensor can sense that disc. Now, my boiler has an API, which allows me to check if it is heating water or not, so I know when hot or cold water is being used. As far as I can see, you have Techem water meters. I believe they can be read from the outside using a 866mhz signal.
@@cokeacolasucks I am pretty sure that the whole thing does not need a lot of power to operate, so a capacitor might be enough and easier to charge than a battery. It's same idea as RFID cards.
@@cokeacolasucks Battery need a lot of power to charge. extraction this power from a fossed flow is not possible. i you even found a generator to do so, the generator will convert the kinetic energy of the water flow to electricity, reducing the pressure at shower head, ruining the shower experience
Good job Great Scott! I did a similar project with a YF-B5 flow sensor. At the end of each shower, I programmed a Service Call to Google Home Mini speakers to shout out how many litres have been used, how long the shower lasted, and what % the shower duration was compared to national average. For the ON/OFF flow sensor, I didn't use a mechanical one, I programmed a boolean to tell if flow was On or Off.
As someone with no "running water" during about 6 months in winter in, north of Sweden, I'm using a reservoir with 1200 liter (used to be 600). That is about 6.5 liter per day for drinking, cooking, showers, dishwasher and some limited washing machine (I try to wash at other places than home). It absolutely astonish me to see others water consumption 😃 I'm planning a mobile tiny home and will have water consumption messure like this and ultrasonic water level messure but also just now building a shower that recirculating water with inline filters for a usage of about 5 liter per shower no matter how long it is used (with inline instant water heater to keep temperature). But the upside is that my water/waste bill is zero since I live in the forrest not connected to the water grid 😀
I love these flow switches and flow metres. When my boiler needed replacing I salvaged them out of the old boiler. I also got some out of an old coffee machine from work. I used them in various projects to detect water and biodiesel flow.😉👍
Really what are you a weirdo with nothing better to do.....Want to show how much water you use in a shower....Put the plug in and when you've finished, then measure it! No need whatsoever to moan about the 'Environmment' whilst making over complicated parts, that in total actually cost what to produce and ship and package, in plastic no doubt, plus using Lithium Batteries on projects.....The hypocrisy is laughable. Plus if you can't get in a shower, clean yourself in about 2 mins, then get out, there's something missing....
Did you include the cost of heating the water to a comfortable shower temperature? I don't know how it's done in other places in Germany, but did you also include the waste water fee?
You can read the blinks on your watermeter... Most electronic meters generate power by the water sensor and has a indicator light that registers a certain amount of liters.
Off-Topic fun fact: I closed a meeting this morning with the phrase "aaannnddd see you next time" thinking the whole day where this phrase and pronounciation came from. Now I remembered. #addicted ;)
Really nice project! But I think the price/recourse calculation are quite off, since heating up the water is way more expensive than the water itself and heated tap water has a much higher climate impact than cold one. Tap water has a footprint of something like 0.35 gCo2/liter while heating the same amount water for 20° (for example with natural gas) costs around 13 gCo2/liter. So much more than the water itself!! So adding a temperature sensor would be really useful I think.
Oh please stop peddling "save the planet" claptrap. We are on the cusp of extremely cheap electricity thanks to extremely cheap solar and extremely cheap wind energy. Steel and aluminum used to be extremely expensive too; nowadays they are extremely inexpensive. Burning fossil fuels at scale is very rapidly being phased out. In another 20 years fossil fuels will probably hardly be a significant source of energy. Once that happens all of this greenhouse effect, er, uh, I mean global warming, er, uh, I mean climate change hysteria will no longer "be a thing." Learn to think clearly and scientifically. Don't succumb to fearmongering.
Awesome build, I’d be interested if the variability in the water sensor is consistent enough to be calculated… I’m thinking if you can measure rpm and adjust the output accordingly. E.g, if 1k rpm = 5 liters per minute and 2k rpm = 20 liters per minute and 3k = 25 lpm we should be able to create a curve to approximate the output based on flow rate… makes me want to build it just to find out :)
That sounds like a good idea. Definitely worth it to look into it. Only problem for my setup would be that integrating such a curve into the home assistant programming would be kind of hard.
@@greatscottlab You could implement the curve into the program on the ESP and let it report ’linearised’ pulses instead of actual pulses. That way you can still use a constant conversion factor from pulses to L/min in Home Assistent.
@@greatscottlab Let the ESP do the linearization and conversion to liters and liters/minute. You already know all you need. Just take the low flow and high flow pulse-per-second and make linear interpolation between the two from the input pulses in a time interval. Should be pretty much precise enough to do the math once every second.
With access to the water meter the "AI on the edge" project could also be interesting. It's using an ESP32 cam module to read & analyze an analog water meter. The live update rate is probably a lot slower, but if statistics by hour or day are enough it's a nice project.
Link to the video mentioned by @Leif: ua-cam.com/video/d_u8c3bu-zg/v-deo.html&ab_channel=AndreasSpiess Edit: Apologies, that is not the original video, but mentions it. Leaving the link as it's interesting and related anyway.
Рік тому
It is very clever solution! The first thing on my mind was - why just not to read plain numbers from the water meter - directly in the place where water pipes are. You can "simply" take a photo on regular interval, process the image for extracting numbers and then store it in Home Assistant even with better precission and not only for one shower pipe. Anyways, the flow sensor seems to be much more fun and challenging project. Everytime I watch your video you get me inspired!
The relevant time to know you’re in the shower for too long is when you in the shower. Toss in an RGB LED visible on the outside of the box that will indicate where you are for your current daily water use limit.
Yeah, I was thinking the same when he said that he can check on his water usage from anywhere in the world. That description just so happens to exclude the one place where I'd want to check it. I really don't get why people get excited every time they can connect something to the internet. It's similarly annoying to the trend of more and more things getting battery powered. Alarm clocks, smoke detectors, kitchen timers and similar really don't need to be mobile. You COULD easily just plug them in. For example, since (in Germany) electrical junction boxes are usually near the ceiling, they could just include a USB port for powering smoke detectors. A kitchen timer could be in a feed-through wall plug and wouldn't even take up an outlet spot. In fact, EVERY "you plug the transformer into the outlet" type plug could be feed-through. You wouldn't want to plug a transformer into a transformer into a transformer (that is, if that contraption is hanging from a wall outlet), but plugging a desk lamp into your phone charger shouldn't be an issue at all. (Unless the desk lamp is connected to the internet and will loose it's settings every time you want to take the phone charger with you.)
Another idea is to measure three temperatures: - cold inlet - hot inlet - common outlet. Where you already know the amount of water at the outlet, you can calculate the cold and hot inlet water flow with a "Dreisatz" ;-) . When you also have the temperature difference ∆T from hot inlet to cold inlet, you can calculate the amount of energy you used for heating the water with: w = c*m*∆T. Where c is the specific thermal capacity from water with 4190 [J/(kg*K)], m the amount of water from the hot inlet in litre or kilogram [kg] and ∆T the temperature difference from hot inlet to cold inlet in Kelvin [K].
Ive been watching for years, I didn't even realize I wasn't subscribed to the channel because every new upload was on my feed. Just subbed. Sorry I know it doesn't help the analytics but just know that I have learned a great deal watching this channel.
I've thought about that too. My idea was to set up a small camera that films the numbers on the water meter. The clue is that the arduino camera module is hooked up to a Raspberry Pi Pico which converts the pictures into numbers by a machine learning algorhitm. That way you won't interfere with the water meter and your readings would be quite acurate. Also there are tutorials for TensorFlow Lite on UA-cam which is designed to be used on micro controllers. After googling my idea I wasn't sure if this already exists, but find out yourself! Greetings!
Did you consider using home water metering valve for this same purpose? Possibly in combination with occupancy sensor in bathroom (that could also trigger ventilation if needed).
As much as I love my Electronics Hobby and Microcontroller Projects, using a low-flow shower head is a much better solution. I received one of these shower heads many years ago from my Gas Company. There is no guilt in enjoying a longer shower with one of these devices. And they don't require batteries.
Same, I can now shower three times a long using the same amount of water. You can get them super cheap and as long as you don’t have super hard water require no maintenance
For an SR latch, I've used 555 timers, which is surprisingly easy to wire. It's called "bistable mode" and if you search for "555 bistable" you will find it. Hence 555 timers can be used not only as flip flops (SR latches), but also accept a wide range of input voltages (4.5 to 16 volts), they cost very little, and you could always use them to make blinkenlights.
This is a flow sensor. Proper way to use would be making a pulse rate to flow rate calibration table (hopefully it's linear?) Then use a lut in the code to convert pulse/s to flow rate. This flow rate can be integrated to determine volume of water used
on utility meters, you can often add an esp in front, with a light sensor to capture the light pulses emitted. If they are purely mechanical, you often have a disk spinning with contrasting color, that should be readable too. Last option is more complex, but you can spin a solution with a pi, a camera module, and opencv to read the numerical values. However, this is a nice project, and i am building something similar to monitor garden watering. I wasn't aware of the water pressure variance however. Hopefully, for drip irrigation, you need a pressure regulator, thus the pressure of watering is limited to a specific value; thus i should be able to proceed as you did to calibrate it
Nice! What with droughts ongoing at the moment it's super good to track this. I have always just turned my shower on and off during my shower whilst using it, like the old showers that used pull handles to allow water to run.
There are clamp on flow sensors that can measure flow in a pipe without making any changes to the pipe. Also I had a showerhead that had a small hydro-generator to power some LEDs and would change color from blue to red when the water was warm. Something like that could be used to power your project, so no batteries required, maybe a super cap to buffer the power.
Great project as usual! If you search for "water turbine 5V", you'll find different kinds of micro hydro generators you can put in series to you measurement part. It could be used to recharge the battery again. Maybe you need to convert from 12 V to lower voltage, but as I know you are an expert in voltage conversion ;-)
Many of the moderm power/water meters have either a wireless or a ir/wired interface where you can "easily" read the data of them in a similar way that your utility company does. I know a few people that do exactly that. This gives you the most accurate graphs possible(literally the exact same that your utility company sees and charges you for) and does not require any modifications to the meters themselfs
Thank you very much Scott. I have been thinking of making something similar. I live in Scotland where water is charged at a flat rate (It rains a lot) and there isnt a significant environmental problem with using a lot of it. What is not free however is hot water. So I intend to measure the temperature as well as the flow rate and work out the Kwh consumption. Unfortunatly my wife has installed one of those fancy showers with multiple outlets so that will nicely complicate the system! I'll need a display in the shower too because thats where the infomation is going to do the most good.
Hey cool project! I installed one of those flow metres in my beer fridge so I would know how much beer I had drunk, or not for that matter. Interesting and useful to know the actual usage, compared to the perceived usage of any liquid that's costing you money.
All good lead has its place, the one I used was food safe plastic with copper fittings designed to be used with water filtration equipment I'm pretty sure, I'll be fine. Interesting project, however I found the impeller agitated the beer and produced more of the fluffy stuff than it would have at the same psi if served from a tap without a flow metre so maybe not really worth doing.
I like your idea.. This is great.. Our water company has an app which allows me to see how much water we are using. It has graphs and history plus other features.
I think the next big step is to make this device absolutelty passive and battery-less with just the power of water stream (maybe even heat) making the circuit work. Great work!
Well this is an interesting topic, since I was busy with the water flow sensors today as well. But my main idea is not for water usage measurement per se, but the measurement in flow between hot and cold water paths. The reason for it is to use the percentage in the flows between the two paths to automatically set a temperature point for a water heater within a showerloop system. In this way I can have a completely regular shower install, but have the temperature and the rest of a showerloop system be completely automated. If I finish this project then I will definetely post something about it on youtube and github about it :D
Just an idea but maybe for showers you could use the waste water. Measure it's temperature under the drain and adjust accordingly. Although that's definatelly a workaround and does not work for baths or filling stuff with water like buckets
@@ffoska Well it is a showerloop, so all the waste water will go through a pump, filters (sediment, uv) and the be reheated back to a desired temperature (complete water re-usage). In this case I am trying to get the temperature adjusted by using a standard shower value (so no thermostatic valve) and checking the water flow on the separate paths to set a temperature. So far I have all the individual components working as necesarry. But the code and design is rather primitive and experimental (breadboards, janky scripts). So the next step is to get it all together and create a more professional cohesive system :) Also looking at maybe automatic removal and refill of shower water after a certain amount of rotation. btw. water will always be purged after a shower end, so every shower will have fresh water.
Only yesterday I was thinking about making a similar thing for the shower where the first cold water is diverted to a large toilet's flush reservoir and then, once up to temp, automatically switches to shower. I lose about 5L until the water has heated up so might as well catch it for flushing.
I'm just starting to watch, but it made me remember of a small generator I've seen once from China... You basically cut the pipe, let's say, and put it in the middle. You know? Perhaps it could be a good way to measure the water as well... And produce some energy at the same time! 😃
Cool. You can also combine it with data from your boiler or water heater to know how much money you spend on hot water, that is really the cost of making a shower. On a modern boiler with the OpenTherm interface you can get a lot of data about what it's doing so it shouldn't be that difficult to get it (you can also get the flow of the water that passes trough it, thus you can avoid the sensor in the shower, if you estimate the percentage of hot water that you mix with cold one).
Nice Project, The missing precision of lowend waterflow sensors is also one thing i had my share of problems with. Right now i am trying to create a easy to build and accurate DIY flow sensor. Also you could have added a temp sensor so you can also calculate the energy usage from your shower, in times of raising energy prices it would have been a neat feature.
like the project Idea, I have a suggestion: could do a the same sort of project for button shushing toilets, (or may be toilet in general) a different angle, of the same problem, false, or phantom, flushing does not stop, and in only noticed next time, the toilet or bath/shower is used, could be happing for hours or even days, if happens and you not being in you home at the time? alter the project a little, to say a little more than say one load water, without an interrupt in the water flow, (a water on, off as fill water tank, for next use? a water solenoid (auto on/off) would be nice, but even just big loud sound (like remote door Bell) would also help?
When I first saw this title and the thumbnail with no wires, I was sure that it is using some kind of water generator to power itself. Well, it isn't but that's a pretty good suggestion! Edit: Maybe use the generator to charge the internal battery to get waaay more runtime and still preserve the stability of a battery.
I have an electronic valve to the hot water pipe leading to my shower. Before showering I need to ask my google assistant for hot water. It closes automatically after 10 minutes. That is a very good way to save some money.
To solve your charging port problem you can use magnetic charging port, it is really cheap and I've used this solution in few projects! For warm water price there's few solutions, just measuring water temperature and calculate the amount of energy necessary to heat the volume from 18 to 38 and assume that your heating system is 100% efficient (if it's electric). Or you can simply measure the power consumption during a day and average the consumption. Last solution is to transfer this device from your shower to your heating vessel and measure all your hot water consumption
Very nice idea, which can be also extended to measure how much washing machine or dishwasher use. Isn't it possible to read pulses from your water measure? By the way I achieved similar, but obviously less accurate shower water usage sensor with humidity sensor next to shower. I observed that there is more or less the same pressure and mix of hot and cold water used everytime, so I calculate just time based on humidity changes.
Loved the video! Good topic and good solution. I would have approached the home assistant part in another way: I wouldn't have used home assistant at all, but I would have used grafana+prometheus. It's really easy to code for the esp32 to communicate with prometheus, and it's really easy to host grafana and prometheus with docker. Grafana has awesome graphs, and you can easily configure it to do math and calculate cost among tons of other things. Great video though!
@@NicosLeben Good question, I've built various ESP32 projects with Prometheus and also tried InfluxDB, and they have a different working principle. InfluxDB works with a library and interacts with an API which means you need keys and various forms of configuration. Prometheus works pretty differently: It sends http get requests to a target IP periodically, and the target (in this case an esp32) replies to the request with the data. This makes the code pretty simple, since the esp32 only has to reply to a get request with whatever data you want. Also, Grafana and Prometheus are the classic dream team :P so integration is excellent It really comes down to preference, both do a really good job. I prefer Prometheus' simplicity, but it's completely fair to prefer InfluxDB's more featureful solution
I'd put this sensor on the cold input to the hot water service, where the pressure is much more constant, and it then also accounts for all hot water usage. And provided that the HWS is outside of the bathroom, much of the special safety issues are obviated. The alarm could be a simple radio activated standalone unit. My external HWS is controlled by an AVR with an RTC that follows a temperature schedule, so that overnight the water is kept lukewarm, but gets the water hotter at the times that I will need it. It also records inlet, tank, cupboard and external temperatures. I think I'll add one of these flowmeters when I upgrade it.
That reminded me about my plan to get water pressure meter with IoT. My house is small and have no space for water tank and pump, sometimes water flow is really slow so I need to know when the water pressure is highest and pick the best time to shower.
The problem with that sensor measures flow at a specific pressure. If the pressure is increased and one liter is taken as a sample, the time taken is less. The same amount of water will pass but in less time. The little impeller of the sensor has to turn faster. Flow=volume/time....in this case time depends on pressure.
Simple photodiode at old measurement device or pulse-wire at new one (without corrupting certified device). You already have measurement device on input pipe, use this.
Love the videos! I want a diy water temperature display and keeps the water at that temperature. My hot water doesn’t last long and need to adjust the temp every minute to keep a constant temp
Awesome project, that was my next project for my garden watering system 👍 Have you considered using ultrasonic water sensors for your main's water supply? This way you don't have to modify your water install ?
@@greatscottlab To not modify your main water supply, it only works with an clampon method. Its not that simple, but it can work if fitted well. For ultrasonic flow measurement there is a lot to consider instead of just counting pulses. Yet the principle is quite simple. Two transducers alternately sends a burst and this time is measured with a TDC. The ultrasound in flow direction is faster than the ultrasound against flow direction. This difference of time is directly proportional to the flow. So with a simple formula with offset, diameter of your pipe, length etc you get a decent result. You may need some correction factors, becauer there are different flow pofiles (laminar and turbulent). I just remembered that you can also buy ready systems. Then, of course, you don't need to know anything about ultrasound :D
This is a very cool project. If I build this thing, I will also add a temperature sensor to calculate the consumed energy for heating the water. (Maybe I need to measure the cold water temperature in addition for accurate measurement?). I still want to build a self made weather station that accurately measure how much rain falls. Maybe I utilize some of your ideas. Up to now I planned to use a weight cell combined with a Pythagoras cup syle container.
Nice project! I will add some more circuitry like a shower timer that turns of a solenoid valve so the water going to the hose gets blocked. Ofcourse i can turn of the timer aswell if i don't want the feature
Hi, I'm writing through the translator so sorry for the quality. Around 2004-2006, I was interested in cooling a computer with liquid. the problem was that as the water stopped flowing, the computer would overheat. One man figured out a great way to measure the velocity of a liquid. Well, he passed an optical fiber through the tube. Laser diode and receiving diode, measured the time it took for the light to travel over such a short distance. As the liquid flows, it creates pressure and deflects the optical fiber, so the path increases and the time taken for the light to travel that path is longer. Unfortunately, I do not have this study archived anywhere. But in 2006, the chips were slower than they are today, so you can make it. Additionally, you can use the turbine to charge the battery (by flowing water), which will increase the runtime.
It's a nice project. The water flow sensor that you got is cheap as it is not calibrated, it is just a flow sensor (yes/no) and not a calibrated sensor which can cost 10 or 100 times more than that and is restricted to flow ranges, pressure, temp, etc... If you get a calibrated one definitely the project can be moved into the main pipe, right after your internal water cut-off valve so you can track all house water expenditure. I have a similar project (it can be powered all the time as it is in a humid zone and not in a wet zone) but used a cheap meter and faced the same issues as you did with different flows.
You could probably use the pulse interval to calculate flowrate and then calculate volume from that, with consistent water pressure this would work. If you wanted to get fancy you could even add a water pressure sensor to add accuracy.
One thing you can google out from literature and other liquid flow mechanicl measuring projects is that there are two factors apart from mechanical friction, ballance and fluid dynamics affecting accuracy of a propeller sensor: first is the density, or volume weight of the propeller's material should ideally be equal to the dencity of the liquid measured; and the second factor is that flow resistance in the pipe before and after the sensor should be equal. Regarding the second factor some orientation is provided in the printed mounting manuals provided with regular water meters saying that before and after the meter a straight piece of pipe of the same diameter with lenght of minimum 5 pipe diameters shold be provided (in other words, it's more accurate to experiment finding out pulse count to volume relation with shower hose mounted after the sensor).
yeah I bought a shower head with a off valve and it saves me a lot of water. $9 and you can get a valve adapter that lets you turn any shower head into a shut-off head. the reason you may want this is so you can find the perfect temperature once and then you don't have to touch it ever again, you just walk in and turn the valve on and let the water through.
Could be really interesting to add a separate microcontroller measuring temperature into and out of the water heater, then factor in the cost of heating the water as well.
I NEED THIS. I always use way too much water and my water boiler runs out... Also, this makes it easier to calculate the flow rate when programmed properly.
nice project. same applies for driving with the car. a smart taximeter could also be something that makes it easier to "push" you into using public transport or your bicycle.
I've been thinking of a similar system, together with a digital temperature mixer... I.e. knowing the flow of hot and the flow of cold and the temperature of each and mixing them to maintain the perfect temperature while showering, a nice rgb indicator to know exactly when it it ready... and we can define the pressure and temperature we want per person using the shower... Just some shower thoughts for ya
Or you could y'know get a thermostatic mixer valve. These are purely mechanical devices. You can get high-end ones that are built into your shower fixtures.
I think it's possible to ditch the battery completely, and use the pressure of the water itself to spin a small dc motor to act as a generator, with a cap to smooth things up and to provide for power spikes(when wifi is involved). This way, it only runs while water is flowing, and no battery is required (nor charging). With an energy/gas meter on the water heater, a calculated cost per shower cand be sent to the phone automatically! :)
Neat project. Several years back we had a $1000+ USD water bill because of a broken water line after the meter. (In Hawaii, palm roots broke the line. The soil is very porous so there was little visual indication we had a huge leak.) We found a commercial product called Flume the goes against a USA style water meter and senses the flow. It requires batteries (special high capacity AAs) that last about a year. Still, it works great, gives you stats, warnings on usage, peace-of-mind, etc. I would have enjoyed making something like this instead, but for the whole house. You'd need to work out sensing that measured the flow without making any changes to the piping, though. A GreatScott version of that would be a great follow-up project.
9:53 You don't need to install HACS in order to get the additional cards and integrations, you can install it manually (most of the times there is a guide for that at the specific github - mini-graph-card has ti there as well). But good for you that you instaled HACS anyway, it's much easier with it for your future HA modifications.😜 Nice project though.
You can save water by collecting it from drain and use it again for shover, Since there is clean water in drain after sometime. Idea seems doesn't cool But good and can possible. Personally I don't use shover at all. Thank you so much
one nice trick for delayed actions like turn off logic is to use the throttle_average filter followed by a lambda filter like 'if(isnan(x)) { turn_off_code_here }'. Through throttle_average the sensor gets activated periodicaly regardless if it received data or not. If it did not receive data (case for turn off) it responds with NaN. If you wait for it with lambda, you cann then emit an output to poweroff the sensor... Hope this is what you wanted to achieve :)
Friend of mine did similar by stealing the wireless signals transmitted by the water meter - it transmits every minute or so the account number and usage. With an RTL sdr he can listen to all the neighbors - once he figured out which was his he pointed the sdr to an mqtt thingamajig. Granted it dilutes things down to much longer intervals, and misses multiple users at once - but he's a bachelor so it doesn't much matter.
We have a hole house meter that measures in cubic feet, I can take a picture with my phone daily to track usage. 1 cubic foot equals 7.481 gallons of water.
You can use Bernoulli's theorem to measure the exact mass of water used which states that total energy per unit mass is the sum of pressure engery , potential energy due to height , and the kinetic energy much be a constant value so technically more water molecules with this energy constant moves the turbine faster so you need to measure these energies and get the constant by normalisation with mass of water to get the exact amt of water usage
My first thought was that it would be better to place a measuring device on the output of the water heating tank, since that’s the most expensive water you can use but also powering device from mains wouldn’t be any problem since most water heaters located inside “dry” zones.. but then again, for MDUs it’s probably a bit hassle since you probably have some kind of central heating system supplying other apartments
Before the company changed my meter to a digital one I had a small reed contact mounted in my meter. Every half liter a pulse. Used the Electric imp(001) to count the pulses. Pushed the data in the mysql db and used AMcharts to visualise everything. Brings back memories. 🙂
Also sounds very useful :-)
The Digital ones usualy have a S0 / SML or Modbus interface, sometimes all tree. They can give out a much higher resulution and some statistics
What's nice about that setup is it lets you call out the water company's BS when they overcharge you!
We still have a analog meter with arrow's turning and a counter with numbers.
No problem to use a webcam and a raspi to read it :)
@@greatscottlab haha indeed. You can recognise patterns after a while. When my wife took a shower after her night shift at the hospital, or our diswasher, ..
I've never thought of showering as a hobby before, but that's EXACTLY what it is. I also forget time and space while immersed in the warm flow!
😂
Yeap, I know the feeling.
@@devalue7064 Germans are masochists. This thing will probably take off there.
You are lucky. Water makes me feel pain.
@@devalue7064 Crazy that you can't even enjoy a nice long shower now. Politicians need to hang.
Version two: You set a baseline temperature and it automatically dials up the cold water mix percentage as you get closer to the self-defined cutoff time/usage amount.
TBF, I thought that's what he did at first. Or constricting flow altogether.
thats called a water heater
it's also not really accurate given that warm water costs a lot more than cold water and the mixture depends on the desired temperature.
Just use cold water, shower time will decrease automatically and therefore monetary spending will decrease as well.
I can’t applaud enough for your ingenuity, great work 👏
A budding UA-camr.
Great tip!
Ending your hot shower with a quick cold rinse can also reduce anxiety and depression.
Also shower less
@@ErebosGR
Just make sure you don't breathe all the water in from the involuntary gasping that will occur from cold shock. I've had too many falls and close calls from that due to being dizzy and light headed, it's just a bad idea.
@@Lively_1185 Umm, you can make the switchover gradual if you want to...
If you're relying on constant water pressure and flow from your shower for an accurate reading, then you may as well skip the flow meter and have just flow switch and use a timer on the mcu to time how long the water is on for. In fact, just put a waterproof clock in your shower.
I was thinking the same. This is an over engineered solution for something that doesn’t actually save you money.
A clock or timer would achieve the same (without the data,) cost you less time and money and have value outside of showering.
My thoughts exactly!
Brilliant, and spot on. There’s gotta be a pretty linear relationship between shower time and water usage.
Water pressure isn't constant if you can turn a knob on the shower to adjust how powerful it is
Do you keep your water running all the time youre showering? Do you never change mode on your shower head?
You should really mark `pulseCount` as volatile since you are updating that value in interrupt so that compiler does special optimizations for it and you avoid some of the really hard to debug issues. I'd also implement sleep mode which Atmel supports. In that mode its power consumption is really really low around 0.1 micro-amp and interrupt can wake it up. It's simple to implement and makes rest of the code even easier to implement since you'd just have to upload stuff, call sleep and be done with it.
Good idea, gets rid of the bulky float switch. Also I would add price calculation for hot water, since you probably don't use cold water for showers, and the bulk of the cost comes from heating the water with electricity or natural gas, not the water itself
@@ffoska This. Though the heat may come off as an estimate since it takes time to run the hot water through or heat it up. You could probably do an estimate based on the temperature.
Trouble is, it's ESP8266 in the end which has a different idea about deep sleep and interrupts, but yeah, it is possible to solve this problem without an extra water switch.
ESP is not atmel.
@@Gurux13 I’m not so sure that ESPHome - which integrates nicely with Home Assistant - implements deep sleep very well. Maybe I’ve just missed it so far.
In defence of the cheap flow meter, I suspect that higher end devices do exactly the same thing but have a calibration in their datasheets. The problem is that you aren't directly measuring flow rate, you're measuring rpm as a proxy for flow rate, and since it's an impeller with gaps around the blades a higher flow rate will change the relationship between flow and speed (guessing here but presumably higher flow rates didn't increase speed by as much, would expect to see bypassing to an extent/less efficient operation). Intuitively it makes sense that pulses should match up with fixed volumes but once you lose the assumed linear relationship that goes out the window.
There are other types of more exact flow meters on the market. For example, one working principle is measuring the difference in the reflection of ultrasound through flowing media. Or you can measure the different electromagnetic features of flowing (electrically conductive) media vs non-flowing. These two principles are much more reliable than the little repeller doing its funny thing. They are a bit pricey though.
That should be the manufacturer's job though, to provide the user with a curve/ function or lookup table for the sensor they are selling. It's not that hard to build an experimental setup for getting halfway decent curve, so shame on them for not doing it
The flow meter needs a consistent stream to work - most flow meters use some form of restrictor or other method to remove turbulent flow.
It wouldn’t matter if it was ultrasound or conductive or mechanical - the sensor won’t read consistently if the application is correct.
Using a flow meter is still a widely used way to measure volumes in industry. Usually you don't have a curve for the flow/pressure/etc as it depends on your liquid (water vs oil for example) they give the relationship for pulse per turn. Some flow meters do have a relationship for pulses, but those sensors measure an actual volume (ex 3.14ml/pulse).
What is done to make accurate and repeatable measures is called a K factore. If you have a constant flow for example coming from a pump you can simply do the calculation at the rated flow, but if the flow or pressure or liquid change, you need a multi K factor compensation. It is done by taking multiple measurements across the flow variation. Then you do a linear interpolation between the K factors.
An example would be for example a variable speed pump in an oil pipe and you need the result in kg. You will start by taking the number of pulses and the frequency at 10% speed, and for each 10% up to 120%. You can do it for a couple of different temperature points as well if you have a temp sensor. Then in production if you read a frequency at 63% you interpolate from you 60 and 70%.
I don't know if it make sense the way I wrote it, but I do those kind of things regularly where I work. We are able to get below 0.5% error with cheap sensors and this method. Our process requires less than 2% error.
@@philippegauthier2212 you are correct on a number of levels, as are some of the other comments to some degree. High quality flow meters often use a four wheel paddle and hall effect pick up. The issue we have always found is turbulence in the fluid flow. There must be some distance between the entering and leaving ports of the sensor from any kind of bend, valve etc that creates any kind of turbulence
This is how you take the joy out of everything. Not every aspect of your existence needs to be logged and generated as a report in your inbox.
I agree but it's still quite interesting from an outsiders perspective, and he seems a really nice bloke. (First vid I've clicked on from his video list)
I can't belive he is at 1.7M subscribers. I ve been following his channel for years and never been disappointed. He is working soo hard for these videos, and explains things pretty well.
Great project! Two thoughts:
1) it would have been yet more complication in the code, but you could also measure the pulse frequency and use that to adjust the conversion factor based on flow rate.
2) They probably don’t exist as commercial parts, but you could use the water flow to run a small generator to charge the battery whenever the water is running. I think it could easily produce enough power that you’d never need to charge the battery externally.
The project was already more than complex enough and works plenty well, so these would be needless additional complications, but I don’t think it would have been *too* difficult to implement a first-order correction for flow rate.
I think #2 is actually nice, although my noob idea was to hook up some alternator to a spinning wheel and charge automatically 😂 wouldn't be very waterproof probably
Apparently you can use MEMS to 'listen' to your water flow. I've seen sensors sold that work non intrusively and I'm guessing this is how they work. One should be able to connect this to their main to detect all water usage even in apartments. I'd love to see a project using MEMS (or even a cheap analog microphone I suppose) in place of a flow meter.
Now that's an idea. Thanks!
Ultrasonic sensing is far more expensive than turbine.
Have you considered using a micro turbine to help with powering the circuit/charging battery? It would probably make this a lot bigger, but it may be interesting how long it would last.
I thought that I was over complicating stuff when it comes to home automation, but you sir, have beaten me. I also measure water usage, but I added a proximity sensor to the outside of my water meter. My water meter has a small metal piece that rotates and passes by when one liter has been used. The proximity sensor can sense that disc. Now, my boiler has an API, which allows me to check if it is heating water or not, so I know when hot or cold water is being used.
As far as I can see, you have Techem water meters. I believe they can be read from the outside using a 866mhz signal.
MHz, mHz - whats nine orders of magnitude between friends
What about using the waterflow to generate a small amount of power and charge a capacitor ? (so there is no need of battery)
Or, could the running water charge the battery, instead of a capacitor? I am intrigued about the capacitor idea, though...
Quint BUILDs tried to power some electronics with his shower once.
@@cokeacolasucks I am pretty sure that the whole thing does not need a lot of power to operate, so a capacitor might be enough and easier to charge than a battery. It's same idea as RFID cards.
@@cokeacolasucks Battery need a lot of power to charge. extraction this power from a fossed flow is not possible.
i you even found a generator to do so, the generator will convert the kinetic energy of the water flow to electricity, reducing the pressure at shower head, ruining the shower experience
Tried it. All commercial water turbines I found could not nearly produce enough energy for the system.
Good job Great Scott! I did a similar project with a YF-B5 flow sensor. At the end of each shower, I programmed a Service Call to Google Home Mini speakers to shout out how many litres have been used, how long the shower lasted, and what % the shower duration was compared to national average. For the ON/OFF flow sensor, I didn't use a mechanical one, I programmed a boolean to tell if flow was On or Off.
thing with sensor used there is that those are sold as FLOW METER sensor, not meant to measure volume of water flowing through it
As someone with no "running water" during about 6 months in winter in, north of Sweden, I'm using a reservoir with 1200 liter (used to be 600). That is about 6.5 liter per day for drinking, cooking, showers, dishwasher and some limited washing machine (I try to wash at other places than home). It absolutely astonish me to see others water consumption 😃
I'm planning a mobile tiny home and will have water consumption messure like this and ultrasonic water level messure but also just now building a shower that recirculating water with inline filters for a usage of about 5 liter per shower no matter how long it is used (with inline instant water heater to keep temperature).
But the upside is that my water/waste bill is zero since I live in the forrest not connected to the water grid 😀
I love these flow switches and flow metres. When my boiler needed replacing I salvaged them out of the old boiler. I also got some out of an old coffee machine from work. I used them in various projects to detect water and biodiesel flow.😉👍
Fantastic idea! I need this to show my family how much water they use in the shower. :)
Yes :-) It is really a revelation to see how much water gets used there.
Really what are you a weirdo with nothing better to do.....Want to show how much water you use in a shower....Put the plug in and when you've finished, then measure it! No need whatsoever to moan about the 'Environmment' whilst making over complicated parts, that in total actually cost what to produce and ship and package, in plastic no doubt, plus using Lithium Batteries on projects.....The hypocrisy is laughable. Plus if you can't get in a shower, clean yourself in about 2 mins, then get out, there's something missing....
Did you include the cost of heating the water to a comfortable shower temperature? I don't know how it's done in other places in Germany, but did you also include the waste water fee?
Waste water is simply the water that you use.
He is speaking of sewer fees
He didn't include the costs of warm water heating, as seen on his improvements sheet in the end.
You can read the blinks on your watermeter...
Most electronic meters generate power by the water sensor and has a indicator light that registers a certain amount of liters.
Off-Topic fun fact: I closed a meeting this morning with the phrase "aaannnddd see you next time" thinking the whole day where this phrase and pronounciation came from. Now I remembered. #addicted ;)
😉
Really nice project! But I think the price/recourse calculation are quite off, since heating up the water is way more expensive than the water itself and heated tap water has a much higher climate impact than cold one. Tap water has a footprint of something like 0.35 gCo2/liter while heating the same amount water for 20° (for example with natural gas) costs around 13 gCo2/liter. So much more than the water itself!! So adding a temperature sensor would be really useful I think.
Oh please stop peddling "save the planet" claptrap. We are on the cusp of extremely cheap electricity thanks to extremely cheap solar and extremely cheap wind energy. Steel and aluminum used to be extremely expensive too; nowadays they are extremely inexpensive.
Burning fossil fuels at scale is very rapidly being phased out. In another 20 years fossil fuels will probably hardly be a significant source of energy. Once that happens all of this greenhouse effect, er, uh, I mean global warming, er, uh, I mean climate change hysteria will no longer "be a thing." Learn to think clearly and scientifically. Don't succumb to fearmongering.
Awesome build, I’d be interested if the variability in the water sensor is consistent enough to be calculated… I’m thinking if you can measure rpm and adjust the output accordingly. E.g, if 1k rpm = 5 liters per minute and 2k rpm = 20 liters per minute and 3k = 25 lpm we should be able to create a curve to approximate the output based on flow rate… makes me want to build it just to find out :)
That sounds like a good idea. Definitely worth it to look into it. Only problem for my setup would be that integrating such a curve into the home assistant programming would be kind of hard.
@@greatscottlab Ahhh I'm just starting to dabble with home assistant. Definitely didn't consider that element. Thanks for sharing!
Endress and Hauser has some interesting in deph explanation Videos about All kinds of flow meters
@@greatscottlab You could implement the curve into the program on the ESP and let it report ’linearised’ pulses instead of actual pulses. That way you can still use a constant conversion factor from pulses to L/min in Home Assistent.
@@greatscottlab Let the ESP do the linearization and conversion to liters and liters/minute. You already know all you need. Just take the low flow and high flow pulse-per-second and make linear interpolation between the two from the input pulses in a time interval. Should be pretty much precise enough to do the math once every second.
With access to the water meter the "AI on the edge" project could also be interesting. It's using an ESP32 cam module to read & analyze an analog water meter. The live update rate is probably a lot slower, but if statistics by hour or day are enough it's a nice project.
An esp32 for edge processing? It can barely emulate Atari 2600
@@hicknopunk It uses Tensorflow Lite which is optimized for the use on Micro Controllers.
@@Leif_YT that is bloody impressive
Link to the video mentioned by @Leif: ua-cam.com/video/d_u8c3bu-zg/v-deo.html&ab_channel=AndreasSpiess
Edit: Apologies, that is not the original video, but mentions it. Leaving the link as it's interesting and related anyway.
It is very clever solution! The first thing on my mind was - why just not to read plain numbers from the water meter - directly in the place where water pipes are. You can "simply" take a photo on regular interval, process the image for extracting numbers and then store it in Home Assistant even with better precission and not only for one shower pipe. Anyways, the flow sensor seems to be much more fun and challenging project. Everytime I watch your video you get me inspired!
The relevant time to know you’re in the shower for too long is when you in the shower. Toss in an RGB LED visible on the outside of the box that will indicate where you are for your current daily water use limit.
Yeah, I was thinking the same when he said that he can check on his water usage from anywhere in the world. That description just so happens to exclude the one place where I'd want to check it.
I really don't get why people get excited every time they can connect something to the internet. It's similarly annoying to the trend of more and more things getting battery powered. Alarm clocks, smoke detectors, kitchen timers and similar really don't need to be mobile. You COULD easily just plug them in. For example, since (in Germany) electrical junction boxes are usually near the ceiling, they could just include a USB port for powering smoke detectors. A kitchen timer could be in a feed-through wall plug and wouldn't even take up an outlet spot. In fact, EVERY "you plug the transformer into the outlet" type plug could be feed-through. You wouldn't want to plug a transformer into a transformer into a transformer (that is, if that contraption is hanging from a wall outlet), but plugging a desk lamp into your phone charger shouldn't be an issue at all. (Unless the desk lamp is connected to the internet and will loose it's settings every time you want to take the phone charger with you.)
Another idea is to measure three temperatures:
- cold inlet
- hot inlet
- common outlet.
Where you already know the amount of water at the outlet, you can calculate the cold and hot inlet water flow with a "Dreisatz" ;-) .
When you also have the temperature difference ∆T from hot inlet to cold inlet, you can calculate the amount of energy you used for heating the water with:
w = c*m*∆T. Where c is the specific thermal capacity from water with 4190 [J/(kg*K)], m the amount of water from the hot inlet in litre or kilogram [kg] and ∆T the temperature difference from hot inlet to cold inlet in Kelvin [K].
Ive been watching for years, I didn't even realize I wasn't subscribed to the channel because every new upload was on my feed. Just subbed. Sorry I know it doesn't help the analytics but just know that I have learned a great deal watching this channel.
I've thought about that too.
My idea was to set up a small camera that films the numbers on the water meter.
The clue is that the arduino camera module is hooked up to a Raspberry Pi Pico which converts the pictures into numbers by a machine learning algorhitm.
That way you won't interfere with the water meter and your readings would be quite acurate. Also there are tutorials for TensorFlow Lite on UA-cam which is designed to be used on micro controllers.
After googling my idea I wasn't sure if this already exists, but find out yourself!
Greetings!
Half way through the vid had to run in and shower (before company arrives) and man it feels good on a hot summer day! :D
Did you consider using home water metering valve for this same purpose? Possibly in combination with occupancy sensor in bathroom (that could also trigger ventilation if needed).
Great project as always. If you ever plan on making a new and improved version it would be cool to make it charge from the water flow as well.
As much as I love my Electronics Hobby and Microcontroller Projects, using a low-flow shower head is a much better solution. I received one of these shower heads many years ago from my Gas Company. There is no guilt in enjoying a longer shower with one of these devices. And they don't require batteries.
Same, I can now shower three times a long using the same amount of water. You can get them super cheap and as long as you don’t have super hard water require no maintenance
For an SR latch, I've used 555 timers, which is surprisingly easy to wire. It's called "bistable mode" and if you search for "555 bistable" you will find it. Hence 555 timers can be used not only as flip flops (SR latches), but also accept a wide range of input voltages (4.5 to 16 volts), they cost very little, and you could always use them to make blinkenlights.
i'm pretty sure they accept even higher voltages
This is a flow sensor. Proper way to use would be making a pulse rate to flow rate calibration table (hopefully it's linear?) Then use a lut in the code to convert pulse/s to flow rate. This flow rate can be integrated to determine volume of water used
Great info for your German audience. Heating prices this winter are going to be insane.
Yep. Trying to help a bit here :-)
on utility meters, you can often add an esp in front, with a light sensor to capture the light pulses emitted. If they are purely mechanical, you often have a disk spinning with contrasting color, that should be readable too.
Last option is more complex, but you can spin a solution with a pi, a camera module, and opencv to read the numerical values.
However, this is a nice project, and i am building something similar to monitor garden watering. I wasn't aware of the water pressure variance however.
Hopefully, for drip irrigation, you need a pressure regulator, thus the pressure of watering is limited to a specific value; thus i should be able to proceed as you did to calibrate it
Nice! What with droughts ongoing at the moment it's super good to track this.
I have always just turned my shower on and off during my shower whilst using it, like the old showers that used pull handles to allow water to run.
There are clamp on flow sensors that can measure flow in a pipe without making any changes to the pipe. Also I had a showerhead that had a small hydro-generator to power some LEDs and would change color from blue to red when the water was warm. Something like that could be used to power your project, so no batteries required, maybe a super cap to buffer the power.
Great project as usual!
If you search for "water turbine 5V", you'll find different kinds of micro hydro generators you can put in series to you measurement part. It could be used to recharge the battery again. Maybe you need to convert from 12 V to lower voltage, but as I know you are an expert in voltage conversion ;-)
Many of the moderm power/water meters have either a wireless or a ir/wired interface where you can "easily" read the data of them in a similar way that your utility company does. I know a few people that do exactly that. This gives you the most accurate graphs possible(literally the exact same that your utility company sees and charges you for) and does not require any modifications to the meters themselfs
This was what I thought about doing
Thank you very much Scott. I have been thinking of making something similar. I live in Scotland where water is charged at a flat rate (It rains a lot) and there isnt a significant environmental problem with using a lot of it. What is not free however is hot water. So I intend to measure the temperature as well as the flow rate and work out the Kwh consumption. Unfortunatly my wife has installed one of those fancy showers with multiple outlets so that will nicely complicate the system! I'll need a display in the shower too because thats where the infomation is going to do the most good.
Hey cool project! I installed one of those flow metres in my beer fridge so I would know how much beer I had drunk, or not for that matter. Interesting and useful to know the actual usage, compared to the perceived usage of any liquid that's costing you money.
Haha that also sounds like a totally german application for such sensors.
idk what flow meter you have but most of them (at least the ones that made of brass) are not lead free...
All good lead has its place, the one I used was food safe plastic with copper fittings designed to be used with water filtration equipment I'm pretty sure, I'll be fine. Interesting project, however I found the impeller agitated the beer and produced more of the fluffy stuff than it would have at the same psi if served from a tap without a flow metre so maybe not really worth doing.
Old joke, fridges ordered the beer already 20 years ago over online.
I like your idea.. This is great.. Our water company has an app which allows me to see how much water we are using. It has graphs and history plus other features.
I think the next big step is to make this device absolutelty passive and battery-less with just the power of water stream (maybe even heat) making the circuit work. Great work!
I think a small battery acting a bit like a ups alongside a flow charging system would be great
add a coil to a second flow meter and power the esp entirely from that! Exactly the sort of project I would expect from GreatScott
Well this is an interesting topic, since I was busy with the water flow sensors today as well. But my main idea is not for water usage measurement per se, but the measurement in flow between hot and cold water paths. The reason for it is to use the percentage in the flows between the two paths to automatically set a temperature point for a water heater within a showerloop system. In this way I can have a completely regular shower install, but have the temperature and the rest of a showerloop system be completely automated.
If I finish this project then I will definetely post something about it on youtube and github about it :D
Just an idea but maybe for showers you could use the waste water. Measure it's temperature under the drain and adjust accordingly. Although that's definatelly a workaround and does not work for baths or filling stuff with water like buckets
@@ffoska Well it is a showerloop, so all the waste water will go through a pump, filters (sediment, uv) and the be reheated back to a desired temperature (complete water re-usage). In this case I am trying to get the temperature adjusted by using a standard shower value (so no thermostatic valve) and checking the water flow on the separate paths to set a temperature. So far I have all the individual components working as necesarry. But the code and design is rather primitive and experimental (breadboards, janky scripts).
So the next step is to get it all together and create a more professional cohesive system :)
Also looking at maybe automatic removal and refill of shower water after a certain amount of rotation.
btw. water will always be purged after a shower end, so every shower will have fresh water.
@@misthafalls That's a really nice concept you got going on. Nice work!
Only yesterday I was thinking about making a similar thing for the shower where the first cold water is diverted to a large toilet's flush reservoir and then, once up to temp, automatically switches to shower. I lose about 5L until the water has heated up so might as well catch it for flushing.
Why not divert all the waste water to your toilet flush?
@@andrewhussey2002 too much!
Nice build!
I'm just starting to watch, but it made me remember of a small generator I've seen once from China... You basically cut the pipe, let's say, and put it in the middle. You know?
Perhaps it could be a good way to measure the water as well... And produce some energy at the same time! 😃
Cool. You can also combine it with data from your boiler or water heater to know how much money you spend on hot water, that is really the cost of making a shower. On a modern boiler with the OpenTherm interface you can get a lot of data about what it's doing so it shouldn't be that difficult to get it (you can also get the flow of the water that passes trough it, thus you can avoid the sensor in the shower, if you estimate the percentage of hot water that you mix with cold one).
Nice Project,
The missing precision of lowend waterflow sensors is also one thing i had my share of problems with.
Right now i am trying to create a easy to build and accurate DIY flow sensor.
Also you could have added a temp sensor so you can also calculate the energy usage from your shower, in times of raising energy prices it would have been a neat feature.
like the project Idea, I have a suggestion: could do a the same sort of project for button shushing toilets, (or may be toilet in general) a different angle, of the same problem, false, or phantom, flushing does not stop, and in only noticed next time, the toilet or bath/shower is used, could be happing for hours or even days, if happens and you not being in you home at the time?
alter the project a little, to say a little more than say one load water, without an interrupt in the water flow, (a water on, off as fill water tank, for next use? a water solenoid (auto on/off) would be nice, but even just big loud sound (like remote door Bell) would also help?
What a nice project, greanpeace also love it. Thank you GreatScott.
You're welcome :-)
When I first saw this title and the thumbnail with no wires, I was sure that it is using some kind of water generator to power itself. Well, it isn't but that's a pretty good suggestion!
Edit: Maybe use the generator to charge the internal battery to get waaay more runtime and still preserve the stability of a battery.
I have an electronic valve to the hot water pipe leading to my shower. Before showering I need to ask my google assistant for hot water. It closes automatically after 10 minutes. That is a very good way to save some money.
To solve your charging port problem you can use magnetic charging port, it is really cheap and I've used this solution in few projects!
For warm water price there's few solutions, just measuring water temperature and calculate the amount of energy necessary to heat the volume from 18 to 38 and assume that your heating system is 100% efficient (if it's electric). Or you can simply measure the power consumption during a day and average the consumption.
Last solution is to transfer this device from your shower to your heating vessel and measure all your hot water consumption
Great project. More money/energy saving gadget ideas please. We are gonna need them this winter.
I AM DEFINITELY MAKING THIS! NO DOUBT!
Very nice idea, which can be also extended to measure how much washing machine or dishwasher use. Isn't it possible to read pulses from your water measure? By the way I achieved similar, but obviously less accurate shower water usage sensor with humidity sensor next to shower. I observed that there is more or less the same pressure and mix of hot and cold water used everytime, so I calculate just time based on humidity changes.
Loved the video! Good topic and good solution.
I would have approached the home assistant part in another way: I wouldn't have used home assistant at all, but I would have used grafana+prometheus.
It's really easy to code for the esp32 to communicate with prometheus, and it's really easy to host grafana and prometheus with docker.
Grafana has awesome graphs, and you can easily configure it to do math and calculate cost among tons of other things.
Great video though!
Is it really more easy to put Prometheus in front of it? I usually use InfluxDB for such temporal based data and show it in Grafana. 🤔
@@NicosLeben Good question, I've built various ESP32 projects with Prometheus and also tried InfluxDB, and they have a different working principle.
InfluxDB works with a library and interacts with an API which means you need keys and various forms of configuration.
Prometheus works pretty differently: It sends http get requests to a target IP periodically, and the target (in this case an esp32) replies to the request with the data.
This makes the code pretty simple, since the esp32 only has to reply to a get request with whatever data you want.
Also, Grafana and Prometheus are the classic dream team :P so integration is excellent
It really comes down to preference, both do a really good job. I prefer Prometheus' simplicity, but it's completely fair to prefer InfluxDB's more featureful solution
nothing like seeing a guy develop a contraption for saving a couple of bucks on water using a $20k+ Keysight scope 😂 made my day
@:35 showering is a hobby😱🎈thanks for the video. I had a similar idea but got stuck on the automation too. I’ll give your idea a try!
I'd put this sensor on the cold input to the hot water service, where the pressure is much more constant, and it then also accounts for all hot water usage.
And provided that the HWS is outside of the bathroom, much of the special safety issues are obviated. The alarm could be a simple radio activated standalone unit.
My external HWS is controlled by an AVR with an RTC that follows a temperature schedule, so that overnight the water is kept lukewarm, but gets the water hotter at the times that I will need it. It also records inlet, tank, cupboard and external temperatures. I think I'll add one of these flowmeters when I upgrade it.
You are a genuine teaching University.
Hope i can work under your guidance
Great Idea!
Now we want a metering of drawn current by the "Durchlauferhitzer" while showering :D
That reminded me about my plan to get water pressure meter with IoT.
My house is small and have no space for water tank and pump, sometimes water flow is really slow so I need to know when the water pressure is highest and pick the best time to shower.
The problem with that sensor measures flow at a specific pressure. If the pressure is increased and one liter is taken as a sample, the time taken is less. The same amount of water will pass but in less time. The little impeller of the sensor has to turn faster. Flow=volume/time....in this case time depends on pressure.
Simple photodiode at old measurement device or pulse-wire at new one (without corrupting certified device). You already have measurement device on input pipe, use this.
Love the videos! I want a diy water temperature display and keeps the water at that temperature. My hot water doesn’t last long and need to adjust the temp every minute to keep a constant temp
Awesome project, that was my next project for my garden watering system 👍
Have you considered using ultrasonic water sensors for your main's water supply? This way you don't have to modify your water install ?
Ultrasonic? Never heard of them. Got a link?
ultrasonic flow sensors are hideously expensive
The sensor itself is usually referred as an Ultrasonic Flow Meter Transducer on AliExpress paired with an Ultrasonic Water Flow Meter main unit
@@greatscottlab To not modify your main water supply, it only works with an clampon method. Its not that simple, but it can work if fitted well. For ultrasonic flow measurement there is a lot to consider instead of just counting pulses. Yet the principle is quite simple. Two transducers alternately sends a burst and this time is measured with a TDC. The ultrasound in flow direction is faster than the ultrasound against flow direction. This difference of time is directly proportional to the flow. So with a simple formula with offset, diameter of your pipe, length etc you get a decent result. You may need some correction factors, becauer there are different flow pofiles (laminar and turbulent).
I just remembered that you can also buy ready systems. Then, of course, you don't need to know anything about ultrasound :D
@@yannickziermann9731 Not sure you can use clamp on systems for that, since you would want to isolate the transducers from the brass.
This is a very cool project. If I build this thing, I will also add a temperature sensor to calculate the consumed energy for heating the water. (Maybe I need to measure the cold water temperature in addition for accurate measurement?).
I still want to build a self made weather station that accurately measure how much rain falls. Maybe I utilize some of your ideas. Up to now I planned to use a weight cell combined with a Pythagoras cup syle container.
Nice project! I will add some more circuitry like a shower timer that turns of a solenoid valve so the water going to the hose gets blocked. Ofcourse i can turn of the timer aswell if i don't want the feature
Hi, I'm writing through the translator so sorry for the quality. Around 2004-2006, I was interested in cooling a computer with liquid. the problem was that as the water stopped flowing, the computer would overheat. One man figured out a great way to measure the velocity of a liquid. Well, he passed an optical fiber through the tube. Laser diode and receiving diode, measured the time it took for the light to travel over such a short distance. As the liquid flows, it creates pressure and deflects the optical fiber, so the path increases and the time taken for the light to travel that path is longer. Unfortunately, I do not have this study archived anywhere. But in 2006, the chips were slower than they are today, so you can make it. Additionally, you can use the turbine to charge the battery (by flowing water), which will increase the runtime.
It's a nice project. The water flow sensor that you got is cheap as it is not calibrated, it is just a flow sensor (yes/no) and not a calibrated sensor which can cost 10 or 100 times more than that and is restricted to flow ranges, pressure, temp, etc... If you get a calibrated one definitely the project can be moved into the main pipe, right after your internal water cut-off valve so you can track all house water expenditure. I have a similar project (it can be powered all the time as it is in a humid zone and not in a wet zone) but used a cheap meter and faced the same issues as you did with different flows.
You could probably use the pulse interval to calculate flowrate and then calculate volume from that, with consistent water pressure this would work. If you wanted to get fancy you could even add a water pressure sensor to add accuracy.
Great project!
Super cool!
One thing you can google out from literature and other liquid flow mechanicl measuring projects is that there are two factors apart from mechanical friction, ballance and fluid dynamics affecting accuracy of a propeller sensor: first is the density, or volume weight of the propeller's material should ideally be equal to the dencity of the liquid measured; and the second factor is that flow resistance in the pipe before and after the sensor should be equal. Regarding the second factor some orientation is provided in the printed mounting manuals provided with regular water meters saying that before and after the meter a straight piece of pipe of the same diameter with lenght of minimum 5 pipe diameters shold be provided (in other words, it's more accurate to experiment finding out pulse count to volume relation with shower hose mounted after the sensor).
Missing this kind of video , finally got after long 😀😀
yeah I bought a shower head with a off valve and it saves me a lot of water. $9 and you can get a valve adapter that lets you turn any shower head into a shut-off head.
the reason you may want this is so you can find the perfect temperature once and then you don't have to touch it ever again, you just walk in and turn the valve on and let the water through.
Also sounds like a good idea :-)
Such a cool project
Also saw the Tweet that you Tweeted saying you got those Drone parts hopefully we’ll get the video soon
You are an amazing person. This stuff is awesome!
Could be really interesting to add a separate microcontroller measuring temperature into and out of the water heater, then factor in the cost of heating the water as well.
I NEED THIS. I always use way too much water and my water boiler runs out... Also, this makes it easier to calculate the flow rate when programmed properly.
regular water meter counter is driven by rotating magnet. you can modify it: remove front panel with digits, install hall sensor, send data
Thanks for the feedback :-)
Pretty sure messing with your meter in any way is illegal, or at least it is here in the US.
@@OtherDalfite i don't mean to touch working one. they are cheap and maybe old meter is laying around
@@yurcchello oh ok
nice project. same applies for driving with the car. a smart taximeter could also be something that makes it easier to "push" you into using public transport or your bicycle.
I've been thinking of a similar system, together with a digital temperature mixer...
I.e. knowing the flow of hot and the flow of cold and the temperature of each and mixing them to maintain the perfect temperature while showering, a nice rgb indicator to know exactly when it it ready... and we can define the pressure and temperature we want per person using the shower...
Just some shower thoughts for ya
Or you could y'know get a thermostatic mixer valve. These are purely mechanical devices. You can get high-end ones that are built into your shower fixtures.
Hi Great Scott. Every Sunday I am visiting for a new video
Awesome :-)
I think it's possible to ditch the battery completely, and use the pressure of the water itself to spin a small dc motor to act as a generator, with a cap to smooth things up and to provide for power spikes(when wifi is involved). This way, it only runs while water is flowing, and no battery is required (nor charging). With an energy/gas meter on the water heater, a calculated cost per shower cand be sent to the phone automatically! :)
Neat project. Several years back we had a $1000+ USD water bill because of a broken water line after the meter. (In Hawaii, palm roots broke the line. The soil is very porous so there was little visual indication we had a huge leak.) We found a commercial product called Flume the goes against a USA style water meter and senses the flow. It requires batteries (special high capacity AAs) that last about a year. Still, it works great, gives you stats, warnings on usage, peace-of-mind, etc. I would have enjoyed making something like this instead, but for the whole house. You'd need to work out sensing that measured the flow without making any changes to the piping, though. A GreatScott version of that would be a great follow-up project.
9:53 You don't need to install HACS in order to get the additional cards and integrations, you can install it manually (most of the times there is a guide for that at the specific github - mini-graph-card has ti there as well).
But good for you that you instaled HACS anyway, it's much easier with it for your future HA modifications.😜
Nice project though.
You can save water by collecting it from drain and use it again for shover,
Since there is clean water in drain after sometime.
Idea seems doesn't cool
But good and can possible.
Personally I don't use shover at all.
Thank you so much
one nice trick for delayed actions like turn off logic is to use the throttle_average filter followed by a lambda filter like 'if(isnan(x)) { turn_off_code_here }'. Through throttle_average the sensor gets activated periodicaly regardless if it received data or not. If it did not receive data (case for turn off) it responds with NaN. If you wait for it with lambda, you cann then emit an output to poweroff the sensor... Hope this is what you wanted to achieve :)
Friend of mine did similar by stealing the wireless signals transmitted by the water meter - it transmits every minute or so the account number and usage. With an RTL sdr he can listen to all the neighbors - once he figured out which was his he pointed the sdr to an mqtt thingamajig. Granted it dilutes things down to much longer intervals, and misses multiple users at once - but he's a bachelor so it doesn't much matter.
We have a hole house meter that measures in cubic feet, I can take a picture with my phone daily to track usage. 1 cubic foot equals 7.481 gallons of water.
You can use Bernoulli's theorem to measure the exact mass of water used which states that total energy per unit mass is the sum of pressure engery , potential energy due to height , and the kinetic energy much be a constant value so technically more water molecules with this energy constant moves the turbine faster so you need to measure these energies and get the constant by normalisation with mass of water to get the exact amt of water usage
Lol
My first thought was that it would be better to place a measuring device on the output of the water heating tank, since that’s the most expensive water you can use but also powering device from mains wouldn’t be any problem since most water heaters located inside “dry” zones.. but then again, for MDUs it’s probably a bit hassle since you probably have some kind of central heating system supplying other apartments
Awesome, thank you
Gut gemacht