New Digital Tank Monitor

[Zeppelinium]

OK, mostly done with the house, so I'm back. I just finished an Arduino-based tank monitor that can replace the old lamp type monitor used prior to 1975 and the newer one that used meters. This is applicable to both the fresh/black/gray systems and the ones that only had fresh and black. Here are the preliminary specs:

1. Shows the fluid levels in 2 or 3 tanks, using the original direct contact sensors (screws).
2. Monitors battery voltage
3. Monitors 6 temperatures. Currently that is envisioned to be:
a. Outside
b. Inside
c. Freezer
d. Fridge
e. Fridge Chimney
f. Flame tube (will light a warning light that can be seen in the rear view mirror so you know if your fridge has blown out while towing)

All this will be displayed on a 4 line by 20 character LCD in three pages. The photos below provide a good idea of the size of the finished unit. The LCD and computer are mounted on the plastic bezel as one piece.

If there is sufficient interest, I'll document the project. Docs should include the software file, the PCB design, the files for 3D printing the bezel and other supports/brackets, and wiring instructions for both the earlier and later systems. Total cost should be under $35.

Please indicate interest by posting to this thread.

[rmkrum]

Interested for sure. Would have to reprogram a bit for my setup, but that’s trivial around here. Like the bezel idea as well.

[Zeppelinium]

The new electronics seem to work as intended. The connectors on the ribbon cables can be mounted to the Airstream cabinetry, allowing semi-permanent hookup to the tank wiring and any temperature wiring that a user might install, yet still provide easy removal of the computer.

I need to get this installed in an Airstream and calibrate the tank level sensing voltages. Can't do that until it gets warm enough to put water in the tanks. Maybe another 6 weeks or so.

For those unfamiliar with the old tank monitors, the pre-1975 had 5 sensor wires and two "stimulating" wires. The 1975 and later had three sensor wires and three stimulating wires. This unit provides the appropriate connections for either system.

Zep

[Zeppelinium]

Here's how it works, addressing the pre-1975 version (with the individual light bulbs) first. These Airstreams had two tanks, fresh water and black water. The fresh tank had 3 sensor lines and one ground line, the black tank had 2 sensor lines and one ground line. The schematic for the fresh tank looks like this (all three analog lines have the pullup resistor, only one shown):



If you hook up an Arduino analog line with the internal pullup resistor "on," the sensor line will sit at nearly +5V until the water level touches the sensor. At that point the sensor line will drop to about 1.5V. This is easily detected and can be displayed in whatever desired format.

The logic, in this case, is that the fresh tank sensor lines are at about 1/4, 1/2, and 3/4 full. When all the sensors are dry (and their voltage high) the tank is "empty." When the 3/4 is wet, the tank is "more than 3/4." Etc. That's all you know.

The black tank is similar, but just two sensors. They are at approximately 1/2 and 3/4 full.

When you hook up your wires, you need to figure out which one is connected to which sensor, or if it's the ground (lowest in the tank) wire. I don't have the maintenance manual for those years, so I don't know what the wire colors are--it would take some trial and error.

[Zeppelinium]

The 1975 and later models (the ones with the meters) have only two wires from each of the tanks. At the tank there are 4 sensors and one ground line, which allows for more refinement in detecting the tank level. The wires are connected together through resistors as shown here (sorry about the upside down wire colors):



As each sensor is immersed, the current through the tank varies, based on the series resistance in its wire. This provides a way to create a voltage at the computer that provides an idea of the tank level. One complication is that the conductance of the water in each tank is different due to the impurities in the water, or lack thereof. As an example, the "resistance" of the water in the fresh tank might be 10K Ohms or more, but the black tank might be 5K. This will still allow very discrete voltage levels to be detected, but there might need to be some adjustment of the logic to allow for variations. (very simple, it's just one number in the code.) Here's an example of data from an experiment I did a few years ago on the voltage as the tank empties:



Note that the voltage "jumps" as each sensor is uncovered. But it's not a sharp jump. This is due to the fact that as the amount of water covering the sensor changes, the effective resistance of the path through the water changes slightly, so you get a bit of a curve as the water surface nears and then goes below the sensor. Not a big deal, but you can see that the meters were really pretty much for show, since the curve is not smooth.

Here's the schematic for the 2-wire tanks:



The logic for this system requires some idea of the voltages from each tank at each level, but other than that it's also straightforward. The plus for this system is that there is no polarity--the two wires can be connected to the computer one way or reversed--it makes no difference.

Note: the 500 Ohm resistor feeding the combined tank wire is only for current limiting in case the wire is shorted to the shell at some point. The analog pin in limited to .020 Amps, so a short to ground would destroy the computer. The resistor limits current to .010 Amps.

[Zeppelinium]

In post #3 above, the small printed circuits at the ends of the ribbon cables are for connecting the tanks (on one) and temperature sensors on the other. The Arduino has 6 available analog pins. In both designs one of them is used to monitor (and display) battery voltage. The pre-1975 tanks have five sensor wires (and 3 grounds), so that works. The newer models only have 3 sensor wires and 3 wires providing excitation voltage to the sensors, so that works, too.

The temperature sensors use digital pins and the second dongle PCB is configured to hookup up to 6 sensors.

The same design can be used for both types of tanks, with only two jumper wires. When one of the jumpers is used to tell the computer which kind of tanks, even the same code can be used.

In a couple weeks I'll try to get the 3D printer files, the printed circuit files, and the code documented and posted.

[Zeppelinium]

Here's the data needed to print the bezel. The design code is for Sketchup (.skp file) and the .stl file can be sliced in any slicer and then printed locally. You'll have to PM me for the files since they are not an allowed type for uploading.

The best thing about the bezel is that it supports the LCD display and the microprocessor, so the unit is small, integrated, and has few stray wires.

[dbj216]

Glad to have you back participating in AirForums. I look forward to new adventures.

While you were out building houses, I went and moved to Colorado. Then I went a bought a 75 Overlander. It has a fancy "Control Center" that controls the water pump on or off, not much else. The control center is pretty well made with good quality meters in my view. My tank monitors did not work, my tanks were broken and cracked, and the tanks were woefully undersized by today's standards; like a 10 gallon gray tank.

Here is a crummy photo of my black tank. You can see the sensors and wires. The tank was U shaped to insure it never drained completely. Clever huh? The marketing folks liked the increased volume. The engineers wanted a rear drain manifold.

Had I known about your project I may have waited on my SeeLevel purchase. Oh well, live and learn.

David