Before the file: a few honest caveats you should know upfront. I can't browse the web, so everything below is compiled from training knowledge (current to ~early 2025) — every price, spec, and URL needs re-verification with suppliers. Critically: **I cannot point you to a verified, purchasable Chinese-made RIM drive over 100 lbs thrust.** The Yamaha HARMO (Japan) is the only mass-production rim drive I can cite with confidence; Chinese "rim drive" listings exist on Alibaba/Made-in-China but most are mislabeled ducted props, so I've built you a verification toolkit (cross-section checklist, RFQ template, red-flag list) rather than pretending to certify vendors. The physics sections confirm your intuition — with one important nuance — and the interactive charts/table let you test scenarios (including back-solving the CdA that reproduces your observed ~1 mph with the mixers).
Save this as `seastead-propulsion.html` and open in any browser (fully self-contained, works offline):
```html
Seastead Propulsion Study — RIM Drives, Mixers & Thruster Alternatives
Seastead Propulsion Study — RIM Drives, Submersible Mixers & Thruster Alternatives
Comparison table · China sourcing playbook · Thrust/power-vs-speed theory & charts · Transit power calculator
⚠ Read first: This document was compiled entirely offline from model training knowledge (no live web access; data current to roughly early 2025).
Every price, specification, and URL below is approximate and must be re-verified with the vendor before any purchase decision.
Rows marked ●○○ are leads, not vetted products. In particular: no verified, purchasable Chinese-made RIM drive
above 100 lbf could be confirmed — see the “Real RIM?” checklist before trusting any “rim drive” listing. Nothing here is a purchase recommendation.
1 · Key findings
Best $/lbf (bollard)
Chinese QJB mixers & trolling motors: ≈ $6/lbf. Hard to beat for pure static push.
Best lb/kW (bollard)
Big slow mixers: ≈ 54–111 lb/kW. Small-unit standout: Blue Robotics T500 (verify W).
Best verified RIM
Yamaha HARMO: 227 lbf / 3.7 kW = 61 lb/kW. Only mass-produced rim drive I can cite.
Chinese RIM status
Listings exist but mostly mislabeled ducted props. Treat as unverified; use RFQ checklist (§3).
Your intuition — confirmed*
At fixed throttle, power does fall as speed → zero-thrust point, plateauing at ≈15–25% of P₀ (spin loss). *Nuance in §5.
Biggest lever
Transit power ∝ V³ and is set by hull drag area. Cutting CdA from ~13 → 4 m² saves more than any thruster swap.
The strategic picture in one paragraph: your existing 2.5 m submersible mixers are bollard-pull champions (huge disk area → superb lb/kW at zero speed),
but they are physically speed-capped: with a zero-thrust speed VZ ≈ 3.2 mph they can never push the hull to 3 mph — that is a pitch/RPM limit,
not a power limit. For 3 mph you need real propellers with higher advance: ideally two large RIM drives (⌀ ≈ 1.2 m, ~8.5 kW each) as prime movers,
while your planned 8 small thrusters are best dedicated to differential control (yaw/roll damping), where they rarely exceed 20% power.
A hybrid “2 big + 8 small” architecture gets you transit efficiency and the stability actuation you want.
2 · Comparison table (click column headers to sort)
Candidates for the 8× differential-control array alongside T500-class units. ●○○
“lb/kW” compares bollard (static) thrust per electrical kilowatt — the right metric for station-keeping, only a rough guide for transit.
Ranges shown where sources conflict; data-v sort keys use midpoints. Confidence: ●●● well-attested · ●●○ probably right, verify · ●○○ lead only.
3 · Is it a real RIM drive? Verification kit for Chinese purchases
Anatomy (what you must see in a cross-section drawing)
Permanent magnets embedded in the ring at the blade-tip radius — the rotating rim itself is the motor rotor.
Stator windings sit in the duct wall facing that magnet ring.
No propeller shaft, no hub bulb, no gearbox. The center is open water.
Direct drive only. If anyone mentions a gearbox or shows a central shaft — it is a ducted prop, not a RIM.
Red flags (fake “RIM” listings)
“Rim” refers only to an outer guard ring around a conventional prop+shaft.
No bollard-pull data, or thrust quoted without simultaneously measured DC watts.
Only beauty-shot renders; no cutaway, no dimensions, no weight.
Price dramatically below motor-cost plausibility for the claimed torque.
“Send a cross-section drawing showing exactly where the rotor magnets are.” (Must be a ring at blade-tip radius; stator in duct; no shaft.)
“Send a bollard-test video: spring scale/load cell + DC wattmeter, full throttle, 60 s, fresh water.”
Thrust-vs-RPM table; efficiency map if available; continuous (not peak) rating in 25 °C seawater.
Construction: blade material, magnet bonding + retaining sleeve (carbon/SS), bearing & seal type (magnetic seal preferred), IP rating, anode provision.
Controller: included? FOC? PWM/CAN/UART input? Bus voltage (48 V?).
Pricing at qty 1 / 8 / 16; lead time; spares (seals, bearings, blades); warranty terms.
Certs: CE, EMC, salt-spray hours.
4 · China sourcing playbook
Platforms:Alibaba (exporters, Trade Assurance) ·
Made-in-China ·
1688.com (domestic wholesale — typically 30–50% cheaper; needs an agent or sourcing service, e.g., Yiwu/Shenzhen-based).
Adjacent industries to mine: e-bike/e-scooter outrunner hub-motor OEMs (a RIM is geometrically an outrunner with the rotor at blade-tip radius) and ROV thruster shops in Shenzhen.
Process: sample order of 1 unit → your own bollard test (spring scale + wattmeter) → then commit to the 8-piece batch. Pay via Trade Assurance; 30/70 TT typical.
Benchmark: strongly consider buying one HARMO-class (or TSL) unit as a reference to test Chinese contenders against on your own dock.
5 · Thrust & power versus speed — the theory behind the charts
For a propeller (a RIM drive behaves like a ducted prop with the motor in the shroud), performance is described by
advance ratio J and the thrust/torque coefficients:
J = V / (nD) · T = ρ n² D⁴ KT(J) · P = 2π ρ n³ D⁵ KQ(J)
Linearized around bollard: KT = KT0(1 − J/JZ) , KQ = KQ0(1 − J/1.2JZ)
where VZ is the zero-thrust speed (advance ≈ 0.8–0.9 × pitch). Three operating modes — all three statements below are true simultaneously,
they just answer different questions:
Fixed throttle (open-loop): T/T₀ = 1 − V/VZ and P/P₀ = 1 − V/1.2VZ.
→ Your intuition is correct: as speed approaches the zero-thrust point, power draw falls, plateauing at
≈ 17% of P₀ (linear model; real units 10–30%) — that residue spins the prop against water friction. Power only reaches zero past freewheel (1.2 VZ).
Fixed power (closed-loop on kW): the controller raises RPM as speed builds, so thrust decays slower:
T/T₀ = (1 − V/1.2VZ)−2/3 · (1 − V/VZ) — up to ~40–110% more mid-range thrust than fixed-RPM, same zero crossing (Chart 3).
Fixed thrust (closed-loop on force, e.g., station-keeping): required power rises with speed, approaching the useful limit T·V
(actuator-disk law: P = T·V·(1+√(1+CT))/2). Holding 100 lbf:
Speed
⌀0.32 m prop (A=0.081 m²)
⌀2.5 m disk (A=4.91 m²)
0.5 mph
0.86 kW
0.26 kW
1.0 mph
1.00 kW
0.46 kW
1.5 mph
1.15 kW
0.68 kW
Ideal-disk lower bound; real hardware lands 60–80% of it (mixers further below, due to low blade-area coverage — hence the empirical anchors used in the charts).
Along the natural drag equilibrium, CT = CdA/Adisk (constant!), so propulsive efficiency is fixed by the disk-area-to-drag-area ratio,
and required power follows P = ½ρ·CdA·V³/η — the cubic law that dominates everything.
6 · Charts & transit-power calculator
Calibration: your reported ~1 mph with 2× ~216 lbf mixers back-solves to CdA ≈ 13 m² (bulky wetted structure). Press that preset to reproduce your experience,
then slide down to see what clean wings buy you. Drag model: D = ½ρ·CdA·V² (wave-making negligible below ~3 mph at this length).
Chart 1 — Available thrust vs speed (fixed throttle), single units, with hull drag curve. Dots mark the equilibrium speed each device alone achieves.
The mixer’s curve dies at VZ ≈ 3.2 mph — it cannot reach 3 mph regardless of power. Higher-pitch props (HARMO-class, small-RIM array) keep pushing.
Chart 2 — Power draw vs speed at fixed throttle (% of bollard P₀). Confirms the falling-power behavior you predicted; at each drive’s VZ (dotted lines)
it still draws ≈17% of P₀ just to spin. Real KQ curvature puts the true residual anywhere in 10–30%.
Chart 3 — Fixed-power vs fixed-RPM thrust decay (normalized). Closed-loop power control buys meaningful mid-range thrust; the zero-thrust speed is unchanged.
Transit power calculator — “what does it take to make X mph?”
Each option is checked against both limits: electrical power available, and thrust-available-at-speed (T₀·(1−V/VZ)) vs hull drag.
ηeq = 2/(1+√(1+CdA/Adisk)) — the actuator-disk efficiency set purely by geometry. Green = feasible, Red = limiting factor shown (T = thrust-bound, P = power-bound).
Reading the table: at CdA ≈ 13 (your current hull) nothing here reaches 3 mph — that is physics, not shopping failure.
With wings bringing CdA to ≈ 3.5, Option D (2× ⌀1.2 m RIM @ ~8.5 kW) makes 3 mph on ~5–6 kW draw, and Option C (8× small RIMs) brute-forces it on ~16 kW.
Illustrative energy budget at CdA 3.5: 3 mph ≈ 6 kW → a 10 kWh bank runs ~1.7 h; 6 kW of solar ≈ 25–30 kWh/clear day → roughly 4 h of cruise per sunny day.
7 · Recommendations for the winged seastead
Adopt the hybrid: 2 big + 8 small. Prime movers: 2× ⌀1.2 m RIM drives (~8.5 kW each, T₀ ≈ 500 lbf each, VZ ≈ 8 mph) — this is the “Option D” spec to hand Chinese OEMs.
Control array: your 8 small thrusters (T500-class or CN equivalents) mounted in opposing pairs per float/wing for differential yaw and roll damping; they’ll rarely exceed 20% power.
Don’t ask mixers to transit. Keep them (cheap, rugged) for heavy station-keeping/current-holding if you like, but their VZ ≈ 3.2 mph caps hull speed near 1.4 mph. It’s a pitch limit, not electronics.
Spend on the wings first. CdA 13 → 3.5 cuts 3-mph power ~3.5×. No thruster purchase competes with that.
Cavitation & noise: at 1–3 mph tip speeds stay under ~10 m/s — cavitation is a non-issue; RIM’s low vibration and inherent blade-tip guarding are real safety/comfort wins for a living platform.
Corrosion & fouling: isolated (galvanically broken) mounts, Mg/Zn anodes per unit, antifoul coating inside ducts; biofouling can quietly eat 20%+ of thrust. RIMs have no exposed shaft seals in the stream — a maintenance plus.
Electrical: specify FOC controllers (supplier’s or VESC-class) with CAN daisy-chain; synchronize slew rates across paired thrusters so differential commands don’t fight; fuse each unit separately; 48 V bus keeps copper sane at 15–20 kW peaks.
Procurement sequence: (1) buy 1 HARMO or TSL unit as benchmark → (2) RFQ 3–5 Chinese OEMs with §3 template → (3) bollard-test samples dockside → (4) batch-order 2 big + 8 small.
```
Three things worth pulling out of the file in plain text:
1. **On finding Chinese RIM drives >100 lbf:** I couldn't verify any specific purchasable model — the honest paths are (a) the Alibaba/Made-in-China "rim driven thruster" search links with the §3 verification kit applied ruthlessly, and (b) commissioning one from China's outrunner hub-motor industry, since a RIM drive is geometrically just an outrunner with the rotor relocated to the blade-tip ring. Buy one HARMO or TSL unit first as a physical benchmark.
2. **Your power intuition, refined:** at fixed throttle, yes — power falls as you approach the zero-thrust speed, plateauing around 15–25% of bollard power (spin loss), not zero. But if you close the loop on *thrust* (station-keeping), required power *rises* with speed toward T·V. Chart 2 shows the first behavior; the mini-table in §5 shows the second.
3. **The uncomfortable arithmetic:** your observed ~1 mph with the two mixers back-solves to CdA ≈ 13 m². At that drag area, 3 mph needs thousands of pounds of thrust-at-speed — no thruster on the table gets you there economically. Getting CdA down to ~3.5 with the wings is what makes "2× ⌀1.2 m RIM @ 8.5 kW → 3 mph on ~6 kW" work. Try the "Match my mixers" and "Clean-wing target" preset buttons to see the flip yourself.