Tension-Leg Trim Winch — Selection Study

Slow, quiet, computer-controlled cable length adjustment at 2,000–3,000 lbf for the seastead tension-leg mooring system

Bottom line: Yes — there is a winch type that fits this duty almost perfectly: a self-locking, two-stage worm-gear electric drum winch (essentially a miniature mooring/tension winch), spooling a single layer onto an 8-inch drum, with an absolute encoder, an inline load cell, and electronic speed control giving roughly 0.5–12 in/min. It holds full tension with zero power, is nearly silent at trim speeds, is failsafe on power loss, and the stroke can be extended to 10+ ft just by adding cable. Runner-up: a self-locking machine-screw linear actuator if strokes stay ≤ 4 ft.

1 · What the duty actually looks like

Your own numbers define the problem. You stated that a 1 ft change in water level is about 1/7th of total buoyancy, i.e. roughly 3,900 lbf per foot of heave. Pulling the platform down 1 ft therefore produces about 3,000 lbf of total pre-tension — matching your "2,000–3,000 lbs" figure, which I treat as the total across the three corners (~1,000 lbf per corner). The hardware below is deliberately sized at 3,000 lbf working per winch so it is valid under either interpretation (total or per-leg) and leaves room to grow.

Parameter	Value / Derivation
Static tension, total (target)	2,000–3,000 lbf (≈ 700–1,000 lbf per corner)
Design working pull per winch	3,000 lbf (13.3 kN) — covers per-leg reading + margin
Peak/transient pull per winch	6,000 lbf (wave dynamics, ~2× static)
Required trim speed	A 8 ft semidiurnal tide moves at peak ≈ 0.4 in/min. 1 in/min beats any tide on Earth.
Reposition speed (rigging)	~12 in/min, attended
Minimum stroke	3 ft (your Anguilla case)
Recommended stroke capability	12 ft — opens up macrotidal sites later
Holding	Indefinite, unpowered, silent
Control	Closed-loop on tension; 3 corners synchronized; slack alarm
Environment	Salt splash zone; prefer winch in dry locker, cable through deck pipe
Power source	Battery/solar → holding must cost ~0 W

The real enemy is not the tide — it's slack. A wave trough larger than your pull-down depth unloads the line, then it snaps taut on the next crest. Shock loads from snap-taut events can exceed static tension many times over. Your Caribbean plan (small waves, 3 ft capability) is sound; for bigger water you either pull down deeper, add a compliant "snubber" element in the line (§7), or both.

2 · Why "geared way down" is exactly right — the math

Your instinct to gear far down is confirmed by the physics. Mechanical power is force × velocity, and your velocity is glacial:

Required line speed (worst realistic tide)
  8 ft range, M2 period 12.42 h  →  peak rate = A·ω
  = 4 ft × 2π/12.42h ≈ 2.0 ft/h ≈ 0.40 in/min

Power at the drum (3,000 lbf at 12 in/min — 30× faster than any tide)
  P = F·v = 13,345 N × 0.00508 m/s ≈ 68 W

Drum torque (8 in root drum, r = 4 in)
  T = 3,000 lbf × 4 in = 12,000 in·lb = 1,000 ft·lb = 1,356 N·m
  Peak (6,000 lbf) = 2,712 N·m

With two worm stages, 60:1 × 61:1 = 3,660:1, η ≈ 0.30
  Motor torque  = 1,356 / (3,660 × 0.30) ≈ 1.2 N·m   (tiny!)
  Motor power   = 68 / 0.30 ≈ 230 W at full 12 in/min
  Drum speed    = 12 in/min ÷ 25.13 in/rev = 0.48 rpm → motor ~1,750 rpm ✓

Daily energy
  Caribbean micro-tide: ~12 in travel/day → ~6 min of running → ~0.03 kWh
  Macrotidal site (16 ft range): ~0.4–0.5 kWh → trivial against solar + batteries

So a ½ hp-class motor does everything. The gearbox does the heavy lifting; the electronics only ever make small corrections. This is why the solution is cheap, quiet, and sips power.

3 · Primary recommendation: self-locking worm-gear drum winch

RECOMMENDED A high-ratio worm gear is self-locking: the lead angle of the worm is shallower than the friction angle, so the load cannot back-drive the motor. The winch holds 3,000 lbf indefinitely with no power, no brake coil burning watts, and no drift. Payout is also fully motor-driven (the load can't "gravity-feed" out), which gives you clean, symmetric control in both directions and makes runaway impossible.

Per-corner unit specification

Item	Spec	Why
Gearing	Two worm stages, ~3,600:1 total, output ≥ 3,000 N·m	Self-locking; worms are the quietest mesh (sliding contact, no spur whine)
Motor	24 V BLDC, ~500 W, fanless/TENV, IP66, with drive on CAN/Ethernet	Fan is the #1 noise source at these speeds; integrates with your existing computer network
Drum	8 in root dia × 5 in wide, single layer	D/d ≈ 21 on 3/8″ wire (fatigue-friendly); single layer = constant radius = stable calibration
Stroke	12 ft of cable (= 0.48 drum rev; axial wander only ~2.2 in)	Covers Anguilla now, macrotidal sites later
Speed range	0.5–12 in/min electronically	0.5–1 for autonomous trim; 12 for rigging
Feedback	Multiturn absolute encoder on drum shaft (4,096 cpr → 0.006 in/count)	Position survives power cycles; resolution far beyond need
Force feedback	Inline load cell or load pins at fairlead, 10 kip, IP67	Control tension, not just length — cable stretch makes length-only control lie to you
Redundant hold	Optional spring-applied, energy-released brake	Belt-and-suspenders; also guards against vibration unseating the worm in rare cases
Manual backup	Hand-crank / hex-socket drive on gearbox input	Recover from total electrical failure
Materials	Epoxy+polyurethane coatings, stainless fasteners/shafts, synthetic gear oil, isolated from aluminum hull	Saltwater survival; avoid galvanic couples with your aluminum structure

Fig. 1 — One corner, elevation view. Winch lives in a dry locker; cable exits vertically through a fairlead; snubber cushions dynamics; paired helical screws anchor the leg.

4 · Strong alternatives

ALTERNATIVE AMachine-screw jack / linear actuator

A vertical self-locking Acme-thread actuator (industrial "machine screw jack," or marine-grade linear actuators of the LINAK class) pushing/pulling via a sheave. The most precise and quietest option — position accuracy in thousandths of an inch, zero drift, totally sealed units available.

Pros: silence, precision, inherent self-lock, compact, trivially computer-controlled.
Cons: strokes beyond ~4 ft get bulky and expensive; exposed screw threads hate salt unless bellows-protected; higher cost.

Pick this if you commit to the 3 ft Anguilla case and want maximum refinement.

ALTERNATIVE BElectric chain hoist (budget)

An off-the-shelf 1–3 ton electric chain hoist mounted under the hull, chain down to the anchor eye. Chain gives exact length measurement (count sprocket teeth) and unlimited stroke.

Pros: cheapest, rugged, huge stroke, widely available (choose worm or oil-bath brake models).
Cons: clatters when moving; chain needs freshwater rinses + lube in salt service; less refined tension control.

Pick this for the first prototype while you validate the tension-leg concept.

Considered and rejected

Option	Verdict	Reason
Hydraulic cylinder + HPU	✗	Pump noise, standby leakage, seals in splash zone, poor efficiency for micro-motions on a solar budget
Ball screw + servo + brake	✗	Back-drivable — holding depends entirely on a brake; a brake fault = runaway. Its efficiency advantage is worthless at your duty cycle
Capstan / traction winch	✗	Rope must slip on the drum to hold → creep under static load; needs a separate lock; complexity without benefit at 3 kip
Pneumatic	✗	Compressibility = sloppy position holding; compressor moisture/corrosion
Thrusters as station-keeping	✗	Burns kilowatts continuously; fine for hours, wrong tool for seasonal mooring
Fixed-length legs + ballast trim	△ Plan B	Pumps replace winches (fewer splash-zone machines) but add tanks/plumbing, and pump failure strands you at the wrong tide height. Keep as fallback architecture

Head-to-head matrix

Criterion	Worm drum winch (recommended)	Screw jack / linear actuator	Ball screw + brake	Chain hoist	Hydraulic cylinder
Holds without power	★★★★★	★★★★★	★★☆☆☆	★★★☆☆	★★★☆☆
Position accuracy	★★★★☆	★★★★★	★★★★★	★★★★☆	★★★☆☆
Quietness	★★★★☆	★★★★★	★★★★☆	★★☆☆☆	★★☆☆☆
Stroke scalability	★★★★★	★★☆☆☆	★★★☆☆	★★★★★	★★★★☆
Saltwater suitability	★★★★☆	★★★☆☆	★★★☆☆	★★★☆☆	★★☆☆☆
Maintenance burden	Low–Med	Med	Med	Med	High
Relative cost	$$	$$$	$$$	$	$$$$
Fail-safe character	Excellent	Excellent	Poor–Fair	Fair–Good	Fair

5 · Control system

Fig. 2 — Closed-loop on tension; encoder length is the coarse inner variable.

Control tension, not length. Wire rope stretches ~0.5% at working load (~2 in over a 30 ft leg) and HMPE creeps over months. The load cell is the truth source; the encoder is for rate limiting and travel accounting.
Modes: PARK (hold, 0 W) · TRIM (autonomous, keep tension in a ±10% band) · REPOSITION (fast, attended) · STORM (policy: max pre-tension to prevent slack, or ease off to cap peak loads) · RETRIEVE.
Slack guard: tension < ~200 lbf for more than a few seconds → alarm + automatic take-up. Never let it snap taut.
Synchronization: one corner is master; attitude feedback from your existing IMU/computers biases individual corner targets to keep the deck level — the same software pattern you already plan for coordinating thrusters between two linked seasteads.
Failsafe ladder: encoder loss → tension-band-only mode; load cell loss → motor-current torque estimate; power loss → self-locking worm holds position indefinitely; total loss → manual crank.
Logging: continuous tension/length time-series doubles as mooring-health telemetry (anchor creep, cable wear trends).

6 · Rigging details that make or break it

Single-layer spooling only. Keeps drum radius constant so force/length calibration never shifts. At 12 ft of stroke the cable wanders only ~2.2 in axially — a 5 in-wide drum handles it with room to spare.
Fleet angle: place the lead sheave directly below drum center, within ~2 ft, so the cable leaves tangentially and vertically through a deck pipe. Slow, single-layer service forgives small angles.
D/d ratio: wire rope wants drum/sheave diameter ≥ 20× rope diameter. The 8 in drum gives D/d ≈ 21 on 3/8″ wire — right in the sweet spot.
Snubber (recommended): insert a ~6 ft nylon tail (or polymer spring) in each line, ideally near the anchor. At ~10% stretch it absorbs shock, caps peak dynamic loads into the winch and hull, and cushions any accidental slack-snap event. Trade-off: slightly softer heave restraint — make tails interchangeable so you can tune stiffness per site. This pairs beautifully with your heave plates.
Two screws per corner: either a self-equalizing bridle off one winch (halves per-screw load) or — better, matching your triple-redundancy philosophy — two independent winch lines per corner, each sized as above.
Corrosion: hot-dip galvanized or 316 anchors and shackles; isolation washers wherever steel meets your aluminum hull; sacrificial zincs on the anchor hardware; annual freshwater rinse + wire-rope lubrication; synthetic gear oil changed yearly.

Cable options

Cable	Approx. MBL	SF @ 3,000 lbf	Notes
3/8″ galvanized 7×19 wire	~14,400 lbf	4.8	Dimensionally stable, cheap; rinse/lube annually
3/8″ 316 stainless 7×19	~12,000 lbf	4.0	Best corrosion, ~15% weaker, pricier
7/16″ galvanized 7×19	~19,600 lbf	6.5	Choose if the 3,000 lbf is per leg
1/2″ HMPE (e.g., 12-strand)	~23,000+ lbf	7.7+	Featherweight, no rust; creeps under sustained load — fine here because the control loop re-tensions automatically

7 · Scaling to bigger tides

Stroke is nearly free. Going from 3 ft to 12 ft of travel costs only ~2 in more drum width and a longer cable. Even 30 ft stays single-layer (1.2 drum revolutions).
Real constraints at macrotidal sites: snubber length, anchor embedment depth, and the slack-prevention policy in big seas — not the winch.
If loads ever grow past ~10 kip: same architecture, bigger drum + a planetary boost stage. Nothing conceptual changes.
Noise scales down with speed: at 0.5 in/min the unit should be essentially inaudible over water (<45 dBA at 10 ft). Schedule fast REPOSITION runs for midday; nobody hears a 6-minute whisper at 2 in/min at night.

8 · Validation checklist

Bench proof load at 1.5× working pull, 10-minute hold, measure worm back-creep (should be zero).
500 full-stroke cycles under load; inspect wire for drum crushing.
Salt-fog exposure (ASTM B117, 500 h) on coating coupons.
Night-time noise measurement during TRIM at the Anguilla trial — target < 45 dBA at 10 ft.
Verify helical screw torque ratings in actual Anguilla seabed soil before trusting pre-tension.
Annual: gear oil, brake inspection, cable lube, structural bolt torque checks, zinc replacement.

Procurement reality check: nothing off-the-shelf matches this spec exactly — which is normal. The standard approach is to build the unit from stock subsystems: two catalog worm gearboxes (~3,600:1 combined), a 24 V fanless BLDC + drive, a fabricated drum, a multiturn absolute encoder, a 10 kip load cell, and marine coatings. Expect roughly $3k–$7k in parts per corner plus fabrication labor — a rounding error next to the rest of the seastead, for the piece of machinery that lets it sit perfectly still.