taking out the trash to the dumpster behind his house when his cordless phone rang.
It was from a medical company who had a problem. The company manufactured a wide
variety of products. One of the products was an antibacterial soap dispenser which
was used in hospitals. The company wanted a method to indicate when a 2 liter bottle
of soap was empty. They tried several methods to detect when the bottle was empty but
none of the methods proved robust enough. Wily was intrigued. He asked for more
The very thick soap was gravity fed to an electrical mixing system which
mixed the soap with hot water. The mixture was then sent under pressure out through
a nozzle when a foot switch was activated. The company had experimented with various
schemes to alert the operator when the system was out of soap but all the methods they
tried failed over time. They tried pressure sensors, optical sensors and flow
sensors. The thick fluid tended to clot whenever there was any restriction in the
flow. Flow sensors and pressure sensors quickly produced false readings. Small
bubbles, which often formed in the fluid stream, produce false readings in several of the
optical sensors they tried. Could Wily come up with a better method18
After some thought, Wily suggested the fluid detection method shown below.
The plastic tubing routing soap to the mixing chamber is split, then recombined, forming a
loop. Two toroidal coils are mounted to the outside of the tubing. One coil is
connected to an AC signal source. The second coil is connected a rectifier circuit.
When fluid fills the ring, AC current is induced into the fluid, which loops around and is
detected by the second coil circuit. When there is no longer any fluid in the ring,
the current is interrupted and the signal appearing at the detector coil drops to zero.
The circuit has several advantages over the other methods attempted by the company.
Since the fluid if free to flow through the sensing loop, there is no obstructions for
clots to form. Small bubbles in the stream would show up as a small drop in the
detector signal strength but only large bubbles would completely interrupt the current
loop. A time delay circuit would only activate an alarm when the fluid detector
signal dropped to zero for a few seconds. Wily’s new client liked the concept and
contracted Wily to develop several prototypes to test the sensing method.