brunnhilde's fresh water level sensor failure

Just because it's made in Germany doesn't mean it is good. It is the case with Brunnhilde's fresh water level sensor. Westfalia has been using a similar sensor for a long time now, ever since the Vanagon Camper. The one in my a few months to 25-year old Vanagon Camper still works but not the one in Brunnhilde.

Although I never disected one I inferred that the sensor is based on a number of reed switches with a float that contains a magnet. As the water level changes the float activates one or more reed switch(es). The activation of the reed switch(es) changes the resistance across the sensor terminus. This conveys to the display the water level. While the principle is quite simple there is some mind twisters as I would come to find out.


The failure mold in mine is the display panel always indicate a full tank. I inspected the sensor and I can see many hair line cracks on the plastic tube that contains the reef switches and resistors. I suck at the cracks with my mouth to see which one leaks but I cannot detect obvious leak. However I know my crude test would not detect minute crack that allow water entrance that can cause corrosion.

i cut open the sensor tube using a dremel tool with a mini cutting disc - i was careful to preserve the PCB; the entire PCB is potted in a silicone latex caulking

i carefully cut away the caulking with a sharp utility knife; there are only 3 reef switches to convey 4 possible levels - very clever and optimized design

Pretty soon I was able to determine the circuit. The reed switches are all normally open type. S1 and S2 both became defective from corrosion caused by water. It may seems reed switches with the contact enclosed in a hermetically sealed glass case would be quite immune to water. The truth is they fail very rapidly when exposed to moisture because the terminus corrode readily and it became the weak link of the failure mode.

The mind bender is with the magnetic field caused by the float the magnet will only activate the reed switch within its field. With this sensor the reed switch are quite far apart. The distance between S1 and S2 is the greatest because there is actually a spot (called stuffing option) for a reed switch that is not installed. When the magnet is between two reed switches it would appears to the display panel that the tank is empty. For this not to be the case the reed switch has to be very sensitive. I know they cannot be made so. So how does the designer work around this challenge?

i scrounge through my electronic junk pile looking for relays with reed switch - it been so long since I use them I had a hard time remember where I might find them

eventually i managed to find three types but one (the blue one below) turns out to be normally close type; the second one looks too delicate to be open intact

 success! i found these security sensors

 it has the right normally open reed switch inside and the same 14mm body size

The replacement reed switches have very poor sensitivity and it only activates when I place the float right next to it. I know I can order more sensitive reed switches from digikey for ~$0.6 each. I still have to resolve the puzzle how does the design maintain the closure of one switch while activating the adjacent one.
yes, my test show that for the display panel to show change of water level it has to maintain the closure of one switch while activating the adjacent one

It took me a while to realize the significance of the silicone latex potting. It must have some ferrous metal filing mixed in making it mildly iron like. The weak ferrous property acts as the conductor for the magnetic field effective "stretches" the effective field of the magnet thus keeping one close while moving on to activate the adjacent one. By this time I had already cleaned up the shop bench and threw away the cutting of the caulking. Wanting to find out the puzzle I manage to find a few pieces from the stinky garbage can. Sure enough I was able to find a segment that has iron like property. I use a magnet to confirm it.

So trying to repair the sensor would be too much work. The teardown is however a worthwhile intellectual exercise for me. By then the local rep has already called me back confirming he can order the replacement.

at the mean time i have to cap the hole with a rubber plug

no - forgetting about using the poop sensor; however I am entertaining using it to replace the single level sensor for the grey water tank; i have yet to determine if it is feasible

at the mean while the chinese made Dorman turbo resonator arrived; peeking the inside it is built like an exhaust resonator

i sold the aluminum one to someone who wants it