DIY Capacitive tank sensor

[blw2]

Originally Posted by FW28z

......

Similar or identical to this one:

http://www.amazon.com/dp/B003D3WPGS/...29c02f6c912d_S

The ones I found are $30~50 for a gauge, which is too expensive for testing.

......

Yes, that's the same thing I have on my '13 31L

I have to dig in to see if that part is bad, or there's something amiss in the panel. The mechanical gauge correctly reads much lower than the electric indicator. If I come up with more warranty work, I might just leave it to the dealer to address, but I'm not thinking I want to go to the delay and hassle of taking it to them just for this....

[FW28z]

Originally Posted by blw2

Yes, that's the same thing I have on my '13 31L

I have to dig in to see if that part is bad, or there's something amiss in the KIB panel. The mechanical gauge correctly reads much lower than the electric indicator. If I come up with more warranty work, I might just leave it to the dealer to address, but I'm not thinking I want to go to the delay and hassle of taking it to them just for this....

Hopefully there will be an adjustment potentiometer on the back of the display unit, but I am not holding my breath. I will open it up and check when I have a chance though. The lame owner's manual does not state any kind of adjustment.

I wonder if they would send me a schematic so that I can replace the offending resistor with a potentiometer...

I am going to ask. All they can say is no.

[JamieGeek]

Stopped at my local home improvement store on my way home this evening and picked up some foil duct tape (the kind used to join HVAC pipes together).

I can see why this is the choice of material for use in that marine system. The foil tape easily doubled the capacitance of my test container from any of the other materials I've tested. So much so that I may have to reduce the value of the resistor (as the time it takes to charge/discharge the RC circuit to measure takes a while if the tank is full).

I'm guessing the reason for this is the fact that the foil adheres closer to the "tank" than my other two tests (I had noticed that the capacitance did increase if I could get the material tighter in on the tank).

[FW28z]

I am also thinking this might be some useful stuff:

http://www.amazon.com/dp/B000IC7ZQ2/...=ISR803ILD4W06

It's basically conformal coating in a can.

I am thinking it will waterproof the foil strips. My tanks are fully exposed to the under body of the coach.

[JamieGeek]

So is mine. When the plate tested out well I was thinking to spray paint it with some rustoleum but with the foil solution my initial thought was simply tape over it with clear shipping tape but that coating may be a better solution.

[FW28z]

This stuff might even work better as it's copper... and you can solder the leads directly onto it:

http://www.amazon.com/TapeCase-1126-...ve+copper+tape

It uses conductive adhesive. The marine unit uses these copper strips that are soldered to the sensor leads. Then, the strips are attached to the aluminum strips on the tank. They probably do this for economy, but with a small production run, you won't realize much savings by using the aluminum. Might be just as easy to use a full-length copper strip.

Conformal coating is the stuff that circuit boards are dipped in to make them waterproof.

If you use duct or shipping tape, there is a chance that moisture could get trapped under it, which would affect the accuracy I think.

[JamieGeek]

I'm familiar with conformal coating (at one of my prior workplaces they used a dishwasher to remove all the remaining flux before applying the conformal coat). The local home improvement store sells the "as seen on TV" "Flex Seal"! LOL If it isn't conductive it may make a good cover as well (much like a spray on bedliner would for a pickup truck applications).

The adhesive seems to be conductive. After I applied a few overlapping strips of the foil tape to my "test tank" I checked for conductance across all of them. The test involved 3 overlapping strips which yielded the best results so far. I even had to step down the resistor (to about 3M) so that the measurement can be completed in a reasonable amount of time.

I have yet to attempt to solder to the tape (the next step when I get a chance).

This is the tape I picked up:
Shop 3M 2.5-in x 150-ft Silver Foil Tape at Lowes.com

[FW28z]

I wonder if the dielectric properties of the conformal coatings vs. other waterproofing materials would affect the readings at all?

Another option might be battery terminal protectant spray.

[JamieGeek]

Going a little overboard (from the earlier conversation): Bluetooth:
https://www.sparkfun.com/products/12576

[FW28z]

That is exactly my problem. I keep improving things until they become too complex.

[JamieGeek]

Well soldering to the foil tape purchased at Lowes was a no-go. Even after sanding the tape (in case there was some coating on it preventing that) it didn't work. I ordered the copper tape yesterday and it is showing up today! (shipped from Madison Heights! LOL)

Will try experimenting with that tonight...

[JamieGeek]

As you can see it is very easy to solder to the copper tape. There really isn't much copper tape on the spool though so my coach may resemble the test "tank" above: The 3M tape on the tank with a single strip of copper covering it all. The adhesive on both tapes must be conductive or there is enough of the metal making contact to short it all together: Using my meter I measure about 1 ohm between the end of the wire that is soldered there and different points around the foil.

[JamieGeek]

Closing in on a final circuit:

The left side is the circuitry used for the sensor wire; the right side is the circuit used to measure the tank level.

R6, R7, and Q4 will be used to wake up the Arduino. When the button is pressed on the panel that will allow some current through R6 turning on Q4 grounding the wakeup pin (Putting the Arduino to sleep should draw less than 2ma--at least it was when I measured it earlier). I don't really know the size of R6 yet: I want it to be large enough that the current through Q4 is enough to wake up the Arduino but small enough that it doesn't affect the measurement.

Q1-Q3 are used to "simulate" the sensor probes to the panel. When all three are off (logic low on PD2-PD4) the tank will indicate "Empty". Turning on PD4 should signal "1/3" full, turning on PD3 should indicate "2/3" full and turning on PD2 should indicate "full". I figured it would probably be easier to just put all the resistors here and only connect 1 signal wire than connecting each probe's wire to the board.

SW1 is used to wake up the circuit when standing next to the tank (instead of having to run inside and press the button). Having the button here implies that I'll also have a display showing the level. That circuit isn't on the diagram...yet.

[FW28z]

Why not use a programmable resistor?

https://www.sparkfun.com/products/10613

I am not sure if you can find one withing the range that is needed though.

[JamieGeek]

Because the Arduino I'm using doesn't have an analog output (no D/A converter..only A/D).

[FW28z]

Yea, in my original statement, I forgot the Arduino's "analog" output is actually a psuedo-analog output in that it is actually a PWM digital signal.

So I remembered programmable resistors... I think I have one in my parts bin that I bought a couple of years ago, but never had an analog output project to use it.

Sorry for the confusion.

[FW28z]

Originally Posted by JamieGeek

I figured it would probably be easier to just put all the resistors here and only connect 1 signal wire than connecting each probe's wire to the board.

You actually have to do that. The monitor only has a single wire for each sensor's input.

[JamieGeek]

Originally Posted by FW28z

So I had the idea of a programmable resistor... I think I have one in my parts bin that I bought a couple of years ago, but never had an analog output project to use it.

I looked a few of those up too but they were all 5V parts (at least the ones I found).

I could wire up this guy:
https://www.sparkfun.com/products/12918
but I think this is getting a little complicated... LOL

[JamieGeek]

Originally Posted by FW28z

You actually have to do that. The KiB monitor only has a single wire for each sensor's input.

Unless you disconnect the wires from the probes on the tank...

[FW28z]

Originally Posted by JamieGeek

Unless you disconnect the wires from the probes on the tank...

The three tank probes go into a resistor network in the wiring harness at the probe. The output of the wiring harness is a single wire to the monitor. If you used the three tank wires to connect to the Arduino, they have a ladder network on the other end that would be in the way. Best to just snip out that mess.

I actually found I had 4 of those 4131-103 in my parts box. Unfortunately though, they are only 10KOhm.

5V should not be a problem. The ATMega can run on either 3.3V or 5V, and I think you will find a 5V regulator already on your board you could maybe use. Not sure if it can handle the current demands of the programmable resistor or not, but the UNO boards have a high power regulator for powering other devices.

See, I like to over-complicate things...