DIY Capacitive tank sensor

[FW28z]

I actually found a Eagle .brd file on the website for that FDC1004 breakout board and got 3 boards made from OSHPark. The first board I destroyed as I am still a novice at SMD soldering.

I was successful in populating the second board, but have not tested it yet.

By the way, check out OSHPark for any PCBs you want to make. They are fairly inexpensive, and run $5 per sq-in for 3 boards. That is, you have to buy 3 boards minimum, and the total cost is $5 per sq-in. I got three of the FDC1004 breakout boards made (4 layer boards) for $6 - including shipping. That's $2 per board. Small boards are cheap - but the price gets high if you have a larger board.

So over the summer, I plan on messing around with these ICs.

By the way, I have been kind of busy with other projects... one I idea I had from when I had the Class C that I finally made for my 5th wheel.

It is a Battery/Charger Monitor.

Here is the prototype:



I used this form-factor because I had an unused slide-out switch in my control panel for a 4th slide (my RV only has 3 slides), so it was a convenient place to put it.









You can see the board OSHPark made here. It is double-sided and includes silkscreening, gold plating of the thru-holes, and a solder mask. Pretty nice boards I think. If you have not used the Eagle PCB designer, it is one of the most popular, and free for non-commercial use (but you are limited in the size boards you can make, and are also limited to double sided boards).

So what does it do?

It monitors the voltage on the battery, either with the charger connected, or when the battery is discharging (i.e. boondocking). Sure, you can do this with a voltmeter, but who remembers what voltage a Bulk, Trickle, or Absorption charge occurs at?

So I provide a LED for all 4 different conditions; Bulk, Absorption (aka Normal), Trickle (aka Float) and battery discharge.

But wait... there's more.

If the battery is in discharge, but between 100~75%, the RED discharge LED will be on. If between 50~75%, the LED will blink slowly, and if under 50%, the LED will blink rapidly. When that occurs, you better get your battery recharged.

The Yellow LED will light when the charger is in Trickle/Float mode. It will blink for the first hour, then remain on steady thereafter.

Similarly, the Green LED will light when the charger is in Absorption/Normal mode. As well, it will blink for the first hour, then remain on steady thereafter.

Finally, the Blue LED will light when the charger is in Bulk mode. It will not blink like the Green or Yellow LEDs. However, it will blink for 24 hours upon exiting the Bulk mode.

So what does all this blinking mean.

If your charger cycles between Normal and Bulk too often, it could mean a battery is going bad. So the blinking of the Normal LED will show that.

And some chargers are classified as 4 state chargers. But they are in reality only 3 state chargers but with an "equalization charge". An equalization charge is just the charger going into bulk mode for 15 minutes every 24 hours, so the blinking blue LED can keep tabs on when the last equalization charge was applied.

And when the RV is in storage and hooked to the charger, the monitor should be in Float mode. But if it cycles between float and normal, both LEDs will blink. This could mean too much of a parasitic load exists.

Pretty neat, eh?

This turned out to be an incredibly simple project, and uses an ATTiny85 with a bit of Arduino code. It is one of the major success stories for my design skills... in other words, it is one of the projects that actually worked!

I got the circuit board made from OSHPark, for about $15 for three boards (so the boards are $5 each). And about $10 in parts. The most expensive part is the front panel, which was about $25. But still, for $40 or so, not bad for a prototype project.

If you are interested, I shared the PCB project on OSHPark so that you can order the board directly from them (I get no commission for anyone ordering a board), and can provide links to the PCB, front panel, and Arduino code.

[FW28z]

I have not yet given up on the tank monitor but I want to explore the FDC1004 a bit more. I see that the output is I2C, but I don't think those levels can be transmitted more than a couple of feet. So I also ordered a couple AD5171s, which are an I2C to variable resistor IC, so I can shoot an analog voltage to the monitor... probably via opto-isolator, etc.

I just received the AD5171s, and I am going to be RV'ing for the next 3 weeks, so I might take a project box of stuff with me.

Not sure if the ATTiny85 can do I2C, but it might. If so, probably could output a PWM signal to an optoisolator and send an analog voltage that way as well.

[JamieGeek]

Originally Posted by FW28z

Pretty neat, eh?

This turned out to be an incredibly simple project, and uses an ATTiny85 with a bit of Arduino code. It is one of the major success stories for my design skills... in other words, it is one of the projects that actually worked!

LOL! That is really cool and fits right in on the panel there. Good work.

I have not yet given up on the tank monitor but I want to explore the FDC1004 a bit more. I see that the output is I2C, but I don't think those levels can be transmitted more than a couple of feet. So I also ordered a couple AD5171s, which are an I2C to variable resistor IC, so I can shoot an analog voltage to the monitor... probably via opto-isolator, etc.

Yeah the V2 plan is to mount the new circuit in a box next to the tanks. Since I'm going to bluetooth it to the phone (got all the examples and test apps working already) I only need to wire power there. By being next to the tanks my I2C bus length will be really short (1" or less) and the sensor wires will only be 1-2ft long (I'll probably try to make use of the shielding pins on the FDC as well).

[JamieGeek]

Making some good progress on V2 of the DIY sensor. The idea to bluetooth the sensor to the phone worked out better than I thought:

https://www.youtube.com/watch?v=qxs47rM5SdQ

Next step is to solder the sensor breakout board to an Arduino prototype shield instead of the white proto board you see there (that should also remove a bunch of stray capacitance from that proto board).

Note that I even put in some notifications on the phone. One for "almost full" >75% and one for "Full" >95%.

[JamieGeek]

Well its installed! Yup finally got off my rear and put the thing in the camper.

Had everything all soldered up and ready to go sitting on my desk here waiting for the last thing. The last thing was to figure out where to get power from. I figured there is a really nice switched 12V power right there: The tank heater pads. We never go camping when its cold enough to use them anyway so they are never turned on and when they are they are thermostatically controlled (checked based on the part # on them). Even if the tank heater is on my little Arduino isn't going to draw enough on top of that to make a difference (50ma @ 5V). So now the monitor can be on or off at the flick of a switch.

Took some time to mount it in the back of the bay where the sewer drain is. I figure that bay is pretty dry:

(I'm using really short screws to hold it to the back wall and there is a good 1" or more gap between the back of the screw and anything else back there.)

That is the "sensor": two pieces of copper tape close to each other. The circuit measures the capacitance between the two (which changes with the fullness of the tank).

For now protecting the sensor with some black duct tape. For a more permanent solution I may spray some Flex Seal over it.

There is the circuit ready to be mounted to the wall. (If you'll note behind it is a 9V regulator: Input 12V output 9V. This is because the Arduino circuit I'm using doesn't have the heat sink capacity to drop 12V -> 5V so I put an intermediate regulator in to do 3V+ of the drop.) Yeah I used wire nuts to hold the wires together--what you don't see here is all the electrical tape I covered them with when I was cleaning up the install.

[JamieGeek]

Test running the app:

Wow its already reading 45% full before I calibrate it. I have to empty the tank and hit the "Min Cal" button to record the reading on empty and then fill the tank and press the "Max Cal" button to record the reading on full. After that the app will read accurately.

Whew...that only took 2+ years to accomplish ! LOL

(At this point I've only instrumented the black tank. The circuit and the app have the capability of also monitoring the grey tank. Yeah: If you have an Axis or Vegas 24.1 crawl under there and look at your grey tank...!)

[blw2]

good stuff Jamie.... about like some of my projects that never seem to get finished!

That should be a really nice upgrade. So it's bluetooth tho the phone is it?

Looking forward to reading your review after some use!

Hey, approx how much do you figure you have in it? I only ask thinking that, in bulk scale, it probably wouldn't cost much in teh big scheme of things to add something like that as standard OEM....something that actually works!

[JamieGeek]

Originally Posted by blw2

good stuff Jamie.... about like some of my projects that never seem to get finished!

That should be a really nice upgrade. So it's bluetooth tho the phone is it?

Looking forward to reading your review after some use!

Hey, approx how much do you figure you have in it? I only ask thinking that, in bulk scale, it probably wouldn't cost much in teh big scheme of things to add something like that as standard OEM....something that actually works!

Yup uses Bluetooth low energy to broadcast the levels to the phone (not real secure as any phone could receive it but do I care if others know my black tank level?? LOL). Much the same way the LevelMate PRO works.

I probably have close to $100 or so into it (not counting the $$ for the developer studio for writing the android app--although if you use Android Studio that would be free). ~$20 for the capacitive sensor board, $20 for the bluetooth board, $24 for the arduino board, and the case. Note that I think you can get a smaller arduino board with bluetooth built in--yup here it is so in a much smaller package for $30 instead of the $24+$20. Can't forget the copper tape.

Unfortunately, though, I'm pretty sure we only have one more camping trip left for this season LOL. Thus I probably won't mention it much until it gets more use next season.

[Wrpt58]

Jamie,
How much effect on the capacitive sensing do you expect odd things against the tank wall will have?, like say paper laying/stuck against the inside wall?
You and FW28z have put a lot of thought, time, money, and engineering into this and, while I haven't read the complete thread, I really respect what yall have done here.
Some people would say "Well I don't need that accuracy since I just use it for weekends and relatively short stays", but those like me, that stay in the RV for weeks or months at a time in places without sewer hook ups could really benefit from an accurate, relatively fool-proof, reliable level indication.
Again, I really respect the work yall have done on this and it looks like it could be easily transferable to practically any tank, given access to the tank.

[JamieGeek]

Originally Posted by Wrpt58

Jamie,
How much effect on the capacitive sensing do you expect odd things against the tank wall will have?, like say paper laying/stuck against the inside wall?
You and FW28z have put a lot of thought, time, money, and engineering into this and, while I haven't read the complete thread, I really respect what yall have done here.
Some people would say "Well I don't need that accuracy since I just use it for weekends and relatively short stays", but those like me, that stay in the RV for weeks or months at a time in places without sewer hook ups could really benefit from an accurate, relatively fool-proof, reliable level indication.
Again, I really respect the work yall have done on this and it looks like it could be easily transferable to practically any tank, given access to the tank.

Thanks.

I don't expect much to be stuck to the inside wall where the sensor is mounted since its outside the tank with no protrusions in the tank (like the conventional sensors). Even a little driving should slosh stuff around enough to get an accurate reading. However, so far all my testing has been using plain water; not sure if the different chemical composition in the tanks will change the readings at all (heck from fill to fill the composition will change).

We can't take all the credit: commercial systems for tank level monitoring using this technique have been available for a while--they just are too expensive for RV use. We were just trying to DIY it at a more affordable level (and now that a DIY sensor board is available that made all the difference).

I would think that anyone with an RV would be interested in more accurate level readings from all 3 tanks.

[JamieGeek]

Update: Labor Day camping.

Our annual Labor Day trip is the only trip we go on without full hookups. Thus accurate black tank readings are more essential for this weekend than any other trip we take.

After a few days the panel in the coach had already read 2/3 full LOL but the DIY sensor only read 21%. I didn't check again until we were ready to leave: Opening the hatch and tapping on the tank (on some Axis/Vegas units you can get direct access to the black tank by opening a side hatch) I figured the tank was about 1/2 full. Check the panel: 2/3 full LOL, check with my phone: DIY sensor was reading 45%--that is about 1/2 full.

Since we were ready to leave and still had plenty of fresh water I turned on the bathroom faucet and watched the black tank reading go from 45%, 50%, 55%, 65%, 70%. Again tapping on the side of the tank..yup about 3/4 full. Panel reading: still 2/3 full LOL. At this point we left and dumped. DIY sensor after dumping: 0%, panel after dumping 1/3 ! LOL.

[blw2]

yes, truly LOL

good news that your project seems to work well. It will be interesting over time to see if things like wet TP stuck to the sidewall might affect it, but i'd guess safe bet even if it does it will be the temporary sort of issue, not the constant "wrong" we get form the OEM meters....

I really wonder why they even bother. Well I understand it might sell to some newby buyer, but these things really do become a laughable joke. A real waste. I seriously would rather not even have the panel at all....but to know that I paid for all that wiring, labor, etc.... ugh.

[JamieGeek]

Well now that my "fall" projects on the RV are done LOL I'm back to this: You'll note above that the app shows both black and grey tank levels but that I did not connect up the grey tank. (Mostly because its further inboard of the black tank which makes accessing it quite difficult, another reason is that its right next to the black tank with very little space where I would install the sensor.)

Lately I've been giving thought of installing the sensor on the grey tank, but there is one issue with how the app would read the level. Here is a cross section of what the grey tank looks like in our 24.1: (this would be looking at the front of the tank towards the rear of the coach)

I would be installing the sensor on the long vertical edge (noted by h1, and h2). Thus when attempting to read out a linear % based on the level at that side the % values above 100*h1/(h1+h2) would be different from the values below that level.

I'll have to put in some scaling to the reading (note that the math would assume the top part is just another rectangle and neglect the sloped bottom--not as accurate but much more accurate than a straight linear progression).

I do get back a linear value from the sensor so scaling it shouldn't be that difficult in software--its just math; something these fancy things called computers kind of excel at.

(I did notice that the conventional sensors on that tank take forever going from 2/3 to F...now I know why! LOL)

[JamieGeek]

After some playing around with Excel here is the kind of scaling I've come up with:


The bottom (X-Axis) is the input value in % (so 1-100) the vertical (Y-Axis) is the output value. The diagonal blue line is the linear response (1% in = 1% out) the orange line is scaled based on the shape of the tank.

Since a lot more of the volume of the tank is above the H1 level filling the bottom bit of the tank doesn't increase the % full until that section is full. After that point the % full increases at a greater rate than linear until it catches back up at 100%.

In this instance, since I don't have the tank measurements, I just used some swag values to get the idea of what the math will look like.

[Gowfster]

I built a similar unit on my workbench using 555 running on 12V and using copper tape. I had an issue with algae formation in the tank that created a film on the interior that was conductive and rendered all my readings as full as long as there was any water in the tank. As soon as I wiped down the film inside my container in the area of my copper patch, the readings returned to accurate. This was in dark basement and only after 6 days.
Any issues with the camper installations? I’m concerned now with installation on my four tanks.

For what it’s worth, my design includes 8266 module which is WiFi which allows me wireless connectivity to status panel I designed or to phone or tablet.
I may post the approach here if I decide the algae (or other material)film problem doesn’t prevent me from completing the installation.

[JamieGeek]

Originally Posted by Gowfster

I built a similar unit on my workbench using 555 running on 12V and using copper tape. I had an issue with algae formation in the tank that created a film on the interior that was conductive and rendered all my readings as full as long as there was any water in the tank. As soon as I wiped down the film inside my container in the area of my copper patch, the readings returned to accurate. This was in dark basement and only after 6 days.
Any issues with the camper installations? I’m concerned now with installation on my four tanks.

For what it’s worth, my design includes 8266 module which is WiFi which allows me wireless connectivity to status panel I designed or to phone or tablet.
I may post the approach here if I decide the algae (or other material)film problem doesn’t prevent me from completing the installation.

Interesting: I can only see algae being a problem if you mount the copper strips on the inside. In my application here the copper tape/sensors are on the outside.

[Gowfster]

My square foot of copper is outside the tank. Inside the tank was a copper probe for ground, but copper is an algaecide (I use copper sulfate to control algae in my pond). The next experiment I will not leave my hoses in the tanks (I used a pump to transfer water into and out of the tanks to watch the gauges), in case it was actually the rubber washing machine hoses that were involved. I’m encouraged if you haven’t had the issue with your rig and I’ll forge ahead.
I’ll post some results in the spring when I’m installed and running.

[Thehobe1]

Originally Posted by Gowfster

I built a similar unit on my workbench using 555 running on 12V and using copper tape. I had an issue with algae formation in the tank that created a film on the interior that was conductive and rendered all my readings as full as long as there was any water in the tank. As soon as I wiped down the film inside my container in the area of my copper patch, the readings returned to accurate. This was in dark basement and only after 6 days.
Any issues with the camper installations? I’m concerned now with installation on my four tanks.

For what it’s worth, my design includes 8266 module which is WiFi which allows me wireless connectivity to status panel I designed or to phone or tablet.
I may post the approach here if I decide the algae (or other material)film problem doesn’t prevent me from completing the installation.

I guess we'll have to look at using strain gauge sensors, weighing the contents! Most bathroom scales are using this technique effectively. Who ever thought that strain gauges would get so inexpensive over the years?

[JamieGeek]

Originally Posted by Thehobe1

I guess we'll have to look at using strain gauge sensors, weighing the contents! Most bathroom scales are using this technique effectively. Who ever thought that strain gauges would get so inexpensive over the years?

My capacitive sensor worked great last season. No issues so far (if you look a few posts up you'll see that I'm hoping to get the grey tank wired up this spring--we'll see).

[ButchSmartini]

JamieGeek (and FW29z, of course),

I just read every word of this thread, not daring to skip to the end to see if it worked. I've been working on exactly the same thing for my boat off and on for some time now, and I was hoping that you guys had figured it all out. It looks like you may have, but it appears that you made a significant change near the end - switching to the two vertical strips of copper tape, and (I think) using the FDC1004 as the interface between the strips and the Arduino.

I have tried strips of copper tape in three different places on the outside of the tank (like your large square of aluminum foil in your earlier version), but have found it impossible to get any sort of reliable consistency. It works great on the bench, but when I actually attach the strips to the tank, the numbers change dramatically, and they get so close together (from all four tape strips - some on the full part of the tank, some in the empty part) that I can't reliably tell what the level is. My next try will be using two parallel strips.

Would it be possible to get your Arduino code?

If you guys are serious about monitoring all kinds of things on your RVs (and it seems like you are), you should check out an open source project called Signal K. It was designed for boats, but there's nothing at all boat-specific about it. And another project called SigkSens - for reading data from various types of sensors and sending it via wifi to Signal K. Once your data is in Signal K, you can access it via wifi from any browser. Both are on GitHub if you're interested.