Water level sensor using load cell

[mqbus767]

Wanted to float an idea here to see if anyone might have experience with using load cells (strain gauges) to measure water/holding tanks. The tank level systems that I've seen so far use an internal (probes/float) or external (sonar/capcitive) sensors. I'm wondering about using a load cell to measure the weight of the tank.

I was thinking about a sensors similar to these:

https://www.amazon.com/Pressure-Force-S-type-Sensor-Cable/dp/B01F6IOWB8
https://www.amazon.com/DYHW-116-Compression-Force-Sensor-Applicable/dp/B07H25NM5T?th=1

My thinking goes something like this; my tanks all sit flat on either bay floor or on a framed platform. If I can build a platform for the tank to lay on and embed one or multiple of these sensors in the center or corners, I can measure the relative change in weight between full and empty. The sensor doesn't even have to be all that precise, just consistent. I also think that measuring just one corner of the tank with a single sensor might be sufficient. My water tanks are 100 gallons each, so I would need sensors that, when combined, could handle about 850lbs.

What do ya'll think?

[chessie4905]

i would think that would be too complicated. also 850 lbs will distort pressure in different locations unless botton is pretty robust. they make sensors installed on side of reservoir that is easier serviceable. In your idea, if you have problems later ( or sooner), will you need to drain and remove tank to diagnose. also consider condensation corrosion to sensor over time. the tanks will sweat at certain times. ex. cold water on hot day, like a glass with icecubes.

[mqbus767]

That's true that the placement of the sensor(s) would be important to not create a hot spot on the tank that might wear through. I'm thinking that a plate or plywood base sufficiently thick would prevent that from happening. If I place a single sensor in a corner it might be really challenging to calibrate. Two sensors on opposite sides of the tank might work. It would still be tricky to calibrate though because I have a static downward force exerted by the straps holding the tanks in place. This actually could vary over time as they lossen and then are tightened during mainteance checks. I would imagine this could be solved by having a zero-out button or a trigger in the software that tells the system when the tanks are empty.

Existing systems that use through holes are just not appealing to me. They seem prone to leaks and I don't like the idea of sensors in the black tank. External sensors seem fragile and their resolution is limited to 4ths or 5ths of a tank. Weighing the tanks would be ideal because, in theory, you could know precisely how much is in them. Load cells are already used extensively in industrial applications so I'm not worried about the operating environment.

Part of this is also a desire to interface in with my already existing automation system. These sensors could be combined with an arduino to send messages back to the main computer that could then send alerts, display gauges, and log trends.

[luvrbus]

Check the marine world for good tank level sensors

[Jim Blackwood]

That's a pretty novel approach, but there's no reason it couldn't be made to work. It'll take some engineering. First, consider tank restraints, particularly against braking forces. You need restraints that prevent horizontal movement but allow some vertical freedom. Bear in mind a top restraint can be as effective as a bottom one and both in combination can be the best. Next, your load cell has to handle at least half of the tank weight plus 50%. You can allow it to pivot on the front edge and put the cell in the center at the rear for instance if you make a rigid frame both above and below the cell that will distribute the load evenly. This could be steel strap for instance, maybe something like 2" x 3/8" by the width of the tank. The bottom of the tank may have to be fully supported, such as with a sheet of 3/4" plywood.

For my water tank I plan to use a sight tube. It can be mounted remotely and does not need to be checked often. I am considering translucent plastic for the waste tanks.

Jim

[bobofthenorth]

Agricultural applications for load sensors are numerous. They work in a fertilizer environment and mobile farm installations. No reason they wouldn't work in the relatively benign environment of a bus cargo bay.  Nothing you do will be as corrosive as fertilizer or as abusive as a farmer.

[TomC]

I have the most reliable tank level gauges ever. I have a bright LED light shining into the opaque tank and you can easily see the level. Good Luck, TomC

[windtrader]

I like the idea of a light based sensor and I suspect on fresh water tank it could work pretty well, not so sure about the black tank. LOL

[thomasinnv]

You might look into pressure transducers that lay inside the tank. They are extremely accurate. Some folks in the victron community have used them and love them.

[mqbus767]

I didn't know about the Victron system. Looks like their GX 140 uses the pressure transducers that you mentioned. The system is pretty expensive when you add up all the components. Maybe I can find just the pressure transducers for sale and use them as the start to my own system.

[windtrader]

I would think there are plenty of pressure transducers that work. You most need to get one that has the spec for what weight range you are measuring. The spec should also indicate the voltage range. You just need something that measures that voltage, you create a scale to match the voltage output and varying load (i.e. empty to full).

[mqbus767]

Thanks Don,

I really like the simplicity of the drop in pressure transducers. That seems simpler than a load cell and more servicable. I wonder how reliable the Amazon cheap versions are. Industrial brand name sensors seem to be $150+ per sensor which is a little cost prohibitive. The ESP32 arduinos come with 15 or so A-to-D inputs, so monitoring 4 tanks should be pretty easy. One concern I would have is how they work in the waste tanks without getting plugged up. Also, if experience with Ebay/Amazon components has taught me anything its that there are multiple interpretations of "corrision resistant" and even "stainless". 

[RJ]

. . . so monitoring 4 tanks should be pretty easy.

Four tanks? Fresh water, Grey water, Black water, what am I missing?

TomC's idea of a bright LED light illuminating a translucent tank is actually pretty brilliant. Mount it low and aimed 90º across from the drains and even the murky black tank's content level will show up. Unless, of course, you hide the tanks behind some type of trim pieces. But even then, cutting a 1" wide vertical sight-strip will provide results - I have those on my tanks, and even without a light I can tell their level.

FYI, my coach is 30 years old, and has a tank level monitoring system above the windshield right next to the genset controls. Hasn't worked since the day I bought the bus four years ago - always reads "E" when I push the buttons. Have I missed it? Not in the least. It's just another fancy doodad that's failed and to me isn't worth the time, effort or dollars to fix. When I open the water bay, I can see how full my tanks are, but that means I have to get up off my tush and go outside - which I should do more often anyway - but it's part of my weekly routine walk-around, so no big deal.

Fancy electronics are cool, and your idea is really unique, but two things to consider: When they fail, they can be a nightmare to fix, and two, if you ever sell your rig, will the new owners be able to understand what you've designed and be able to fix it if/when it fails? Ask any Marathon Coach owner about his Crestron system. . .

Keep the KISS principle in mind during your build!

FWIW & HTH. . .

RJ

[richard5933]

I agree with RJ - simple is better any day of the week.

On our 4108 we had no tank gauges at all. Never once did we need them.

We had two 90-gallon tanks, one for fresh water and one for waste water. The tanks were translucent poly and the level of the contents was just a glance away. The reality is that like most people we were in our bays often to retrieve things. Checking the tank levels was as easy as looking at them, and when the waste tank was over 2/3 full we were on the lookout for a dump site.

We always operated from the fresh water pump and never from a city water source directly. City water sources were only used to fill the tank. Since we had two identically sized tanks, there was no way for us to overflow the waste tank (unless we drank a few too many while out for dinner, I guess.) When the waste tank was full, the fresh tank was empty.

In our Airstream we have internal tank sensors. They work okay, and they provide readings in 1/4-tank increments. We can't see the tanks to get an eyes on measurement. In reality, the system works just fine. When the tanks are at 3/4 full we need to look for a dump site and prepare to dump. Knowing anything more precise isn't going to help us any or improve our enjoyment of the trailer any.

Building a complex or overly involved system is great if doing it is part of your enjoyment of the bus. Some enjoy the process of doing it, but it's not necessary for functionality at all. My guess is that after some time of having a system like this you'll get most of your information about the tanks like many - by looking at the tanks when you're in the bays for something else.

[windtrader]

KISS where you can. My tank sensors were on the fritz day one and now done. I could troubleshoot it but it won't do want I want. That is to have all sensors integrate into a digital dashboard. So the arduino or better the raspberry pi would be an excellent board to have sensors connect to, log, analyze, compute the numbers, and most importantly connect to the display panel showing pages made with python. this is done all over the place, so nothing novel here and plenty of places to get up and running.
So, any standalone system is not the route I'd take but fits the bill for many others wanting the KISS solution