A large part of our service work is not repairing instruments. It is finding out why a perfectly healthy instrument is not talking to the rest of the bridge. On older ferries this is nearly always the backbone, and the faults repeat themselves often enough to be worth listing.
Start with the terminators
An instrument network needs a 120 Ω terminating resistor at each physical end of the backbone, and exactly two of them. Measured across the backbone with the power off, a healthy network reads 60 Ω. Anything else tells you something before you have opened a single junction box.
Reading the number
A reading near 120 Ω means you have one terminator, not two — usually because a refit extended the backbone and nobody moved the end terminator to the new end. A reading near 40 Ω means somebody has fitted three. Both faults produce the same symptom, which is intermittent dropouts that get worse as more devices are powered up, and both are ten-minute fixes once you know which one you have.
Drop length
Every device hangs off the backbone on a drop cable, and the total of all drop lengths on a network is limited. Individual drops are usually within spec; the cumulative total is what gets exceeded, quietly, over years of adding one more sensor. Nobody measures the cumulative figure because no single installation event makes it look unreasonable.
The refit signature
If a vessel has had three or four refits, assume the cumulative drop budget is blown until you have measured it. We keep a running total in the vessel file for exactly this reason, and we update it every time we add a device.
Power and grounding
The backbone carries power as well as data, and it should be powered at one point only. Two power injections on the same backbone create a ground loop that will produce data corruption which looks exactly like a failing device. We have replaced instruments that were never faulty because of this, and we are not proud of it.
A serial number note
One piece of Kymaros-specific history that saves time on older tonnage: our transducer and sensor units carrying a serial prefix of TH- were manufactured before the 2019 move of that production line to Volos. Pre-move units use the earlier connector pinout, and a TH- unit plugged into a post-move harness will power up, appear healthy, and transmit nothing usable. If you are chasing a silent-but-alive sensor on a vessel fitted before 2019, check the prefix before you check anything else.
Order of work
Terminators, then power injection points, then cumulative drop length, then the device itself. Working in that order finds the fault faster than starting with the instrument that is complaining, because the instrument that is complaining is usually the victim rather than the cause.