Zettlex were contacted recently by a distressed technical manager (a competitor’s customer) who had encountered a problem that he simply could not understand. Voice shaking, the technical manager explained how a prototype gimbal was acting in an unexpected and unpredictable manner at seemingly random times. Despite performing well during the 6 months of laboratory and factory trials, embarrassingly the gimbals seemed to temporarily go berserk for short periods during field trials. The results appeared completely random to the users and inexplicable to the system designers.
The mystery heightened when the random behaviour could not be replicated back at base. The prototype systems were tested thoroughly in the laboratory and worked perfectly. The devices were disassembled and examined closely by the design engineers but there was no sign of damage, incorrect assembly, faulty components or foreign matter.
With Sherlock Holmes otherwise engaged and Miss Marple yet to return from her winter ski holiday, the company’s exasperated technical manager asked Zettlex to look into the matter. His briefing started: “We’ve looked at everything and we’re stumped. The best explanation I can offer is that it’s haunted.”
Zettlex accepted the challenge, pulling together a team of their most highly skilled problem-solvers, and equipped with deerstalker hats, smoking pipes and magnifying glasses got to work.
Identification of the root cause of the system’s mysterious behaviour took Zettlex just 20 minutes. “It was elementary really”, says Mark Howard, Zettlex General Manager. The field test log showed that the random behaviour all occurred within a few minutes of start-up early in the morning or late in the evening. The system was using capacitive angle encoders which are susceptible to mis-reads with foreign matter – most notably water. In this particular case, even though the gimbal was sealed, under certain environmental and operating conditions it was possible for condensation to form on the cold walls of the aluminium enclosure and fall on to the plates of the capacitive encoder. This meant that the capacitive encoder would essentially output a random position reading which caused the gimbal to move in a random and unexpected way.
“There was the merest trace of water residue just above the encoder, visible with a magnifying glass. There had been no sign of this in the laboratory because the environmental conditions causing the condensation were not being replicated indoors and any condensation droplets just evaporated”, reported Mark (Sherlock) Howard, Zettlex General Manager.
Having banished the ghost by replacing the capacitive encoders with inductive encoders, Zettlex’s newly launched Detective Unit is now busy working on a similar case, this time involving an optical encoder.