Engineering note
The cable run is where most installations fail
An antenna that tracks perfectly can still die from a gland that faces upward. This note covers the deck-side details — entry, routing, bonding, separation — that decide whether the installation survives its first survey cycle.
Entry geometry before anything else
The radome base has one cable entry, and its position relative to the flange was chosen by the manufacturer on the assumption that the cable leaves downhill. If the actual route leaves uphill, the entry moves — or the water gets in.
A gland on a horizontal surface is a cup. On a vertical surface it is a door. The only orientation that self-drains is a downward-facing entry with a drip loop below it, so the water that runs down the cable leaves the cable before it reaches the seal.
- Gland faces down or sideways-down — never up
- Drip loop with a radius larger than the cable's minimum bend radius, below the gland
- Service loop of 300–500 mm at the radome so the unit can be lifted without cutting the cable
- Strain relief separate from the gland: the gland seals, it does not carry load
Separation from power and from RF
The antenna carries a DC feed and an IF/RF signal in the same cable, which makes it sensitive to induced noise from parallel power runs. Crossing power cables at right angles is acceptable; running parallel within a few centimetres for several metres is not.
The same logic applies to VHF and radar cables: a coaxial run alongside a transmitting antenna picks up energy that no amount of shielding on the receiving side removes. Keep the antenna cable at least the separation the classification society asks for, and if that is impossible, use double-shielded cable and document why.
Bonding: one point, not two
The mast or base plate must be bonded to the hull's equipotential system, and the shield of the antenna cable bonded at the same point. Bonding the shield at both ends creates a loop that carries circulating current through the coax — the classic cause of a system that works in harbour and fails in a seaway.
Use a braided bond of adequate cross-section, keep it short, and inspect it: a corroded bond looks fine from a metre away and measures open.
What to put on the drawing
The yard works from drawings, not intentions. A cabling drawing that shows the route, the gland positions, the bond points and the service loops gets built correctly; a specification paragraph does not.
Label both ends of every cable with the equipment tag, not with the function — "ACU-1 IN" survives a refit, "antenna" does not. The same label should appear on the rack schedule, which is what the next engineer will actually read.
Checklist for the drawing and the yard
- 01Every gland faces down or sideways-down, with a drip loop below
- 02Service loop at the radome long enough to lift the unit without cutting cable
- 03Cable crosses power at right angles; parallel runs kept outside the separation the class society requires
- 04Shield bonded at one point only, with a short braided bond to the hull equipotential system
- 05Both ends labelled by equipment tag, matching the rack schedule
- 06Route surveyed at full roll and pitch — what clears at the dock may foul at sea
Questions on this note
Cabling & grounding
Can the antenna share a cable tray with the ship's data network?+−
Physically yes, and electrically it is a risk: the antenna cable carries a DC feed and a weak IF signal. Separating it costs a tray divider; debugging induced noise after handover costs a week.
How important is the drip loop really?+−
It is the single cheapest water-ingress defence on the whole installation. A downward gland without a loop still lets water travel along the cable into the seal by capillary action.