- Home
- Applications
- Motion Compensation Platform
Application · Ships, vehicles and any base that will not hold still
Motion compensation platform, judged by what is left, not what it can reach.
Every other platform on this site exists to produce motion. This one exists to remove it — so the number that decides whether it worked is not travel, acceleration or bandwidth. It is the motion still present at the payload while the base underneath is moving.
Hoping for a residual figure in the first reply? You will get a range and the assumptions it rests on. An honest number depends on your sensor and your loop delay at least as much as on our mechanism — the three-link section shows why, with the arithmetic.
Motion compensation platform
Same phrase, four different machines
Four things are sold as motion compensation, and only one of them is a platform
Enquiries arrive asking for a compensation platform when the right answer is a gimbal, a winch, or something the naval architect should have solved. The four sit at different heights in the same stack, remove different motion, and cost different amounts of money for reasons that have nothing to do with each other.
The distinction that matters is where the machine acts. A stabiliser acts on the hull, so everything on board benefits a little. A heave-compensated winch acts on the rope, so one suspended load benefits a lot. A gimbal acts on the mount, so a line of sight is held but the payload still translates. A compensation platform acts under one defined payload, and nothing else on the vessel notices.
We build the fourth. Where the requirement belongs to one of the other three, that is what you will hear back, because a platform underneath the wrong layer is an expensive way to relocate a problem.
Compensation is not a property of a machine. It is a statement about which layer of the stack you chose to fight on.
Not sure whether the fix belongs at the vessel, the rope or the payload? Describe what moves and what must not, and we will name the layer →
Four motion-compensation layers showing where hull stabilisers, winches, gimbals and payload platforms act
The 4-Way Compensation Crosswalk Which machine removes which motion, and what it leaves behind
| Machine | What it removes | What it leaves | Do we build it? |
|---|---|---|---|
| Vessel stabilisation — fins, gyros, anti-roll tanks, dynamic positioning | A share of hull roll, and station-keeping in the horizontal plane | Heave essentially untouched, and pitch largely so | No — naval architecture, not our discipline |
| Active heave compensation on a winch or crane | Vertical motion of one suspended load, over metres of stroke | The attitude of the load, and everything not on the hook | No — a rope machine, and a different industry |
| Stabilised gimbal or pointing head | Angular error of a line of sight, in two or three rotations | All translation; the payload still heaves with the deck | Sometimes — and where a gimbal is enough we will say so |
| Payload compensation platform | Motion of one defined payload’s own frame, in three or six axes | Everything outside that payload, and everything past the stroke | Yes — this page |
Row three is the one we talk enquiries out of most often. If nothing in the requirement is sensitive to translation — a directional antenna, a camera with a long working distance — a two-axis gimbal is lighter, cheaper, faster and easier to install than any platform we could build, and we would rather lose that order than deliver a six-axis machine to hold a bearing.
The part that decides the answer before we quote anything
Residual motion is set by the worst of three links, and two of them are usually yours
A compensation platform can only cancel motion it knows about, in time, within a stroke it has. Those are three separate limits and they belong to three different parties. A specification that names only the third — the platform — is not describing the system that will be installed.
The 3-Link Residual Chain The residual is the worst link, not the average of them
| Link | What sets its floor | Who owns it in practice | What it costs to improve |
|---|---|---|---|
| Measurement | How well the base motion is known, and where it is known — the sensor’s own specification plus the lever arm from where it is mounted to where compensation happens | Usually the customer: most projects already own an inertial unit | A better unit, a shorter lever arm, or an honest re-statement of the target |
| Delay | Everything between the physical motion and the platform answering it: sensor filter, transport, command interval, servo response | Shared — the sensor and the network are yours, the controller and the drives are ours | A shorter filter, a dedicated link, or prediction; each has a price in accuracy or risk |
| Actuation | Whether stroke and acceleration cover the tail of the record rather than its average, at your load and centre of gravity | Ours | A larger platform, or a narrower band you agree to compensate |
The unhelpful version of this conversation is a supplier quoting a residual figure from the third row alone. We would rather send you the first two rows as questions and wait, because a residual promised without a delay budget is a promise about a system nobody has described yet.
Link 01 · What the sensor knows
A compensator cannot be more certain about the sea than the instrument watching it
Marine motion sensors are honest about this in their own datasheets. Heave accuracy is commonly published in the form 5 cm or 5 %, whichever is greater, and that figure is a property of the unit, not of whatever it is connected to. If the sensor feeding the loop is specified that way, no mechanism underneath the payload can honestly claim a residual heave below it — the platform would be correcting a number it does not have.
Angles behave differently, and this is the single most useful asymmetry on the page. Roll and pitch are typically specified an order of magnitude tighter than heave, which is why angular compensation reliably outperforms vertical compensation on the same vessel with the same equipment. Requirements written in degrees are usually achievable. Requirements written in centimetres of heave need the sensor discussed first.
Where the payload sits away from the sensor — and it always does — the lever arm between them turns the vessel’s angular motion into translation at the payload. That transform is part of the configuration, it needs the geometry to be measured rather than estimated, and it is a common reason a system performs on the bench and disappoints on the deck.
Link 02 · What the delay does
Past a certain delay a compensation platform does not underperform — it amplifies
Take a base moving sinusoidally and a platform that answers it with exactly the right amplitude but a little late. The residual is not “slightly worse”: the two waveforms are offset in phase, and the leftover is the difference between them. Written out, the residual amplitude is 2 sin(π f Δt) times the original, where f is the motion frequency and Δt the total loop delay.
That expression crosses 1 when f × Δt = 1⁄6. At that point the platform is working hard, drawing current, and leaving the payload exactly as badly off as if it had been bolted straight to the deck. Beyond it, the correction and the disturbance add rather than subtract, and the compensated payload moves more than the uncompensated one would have.
This is arithmetic, not a claim about our hardware, and you can check it in a spreadsheet in two minutes. It is also why the honest first question about a compensation project is not how much travel do you need but how quickly do you find out that the deck has moved.
The Delay Ceiling Total loop delay a target reduction allows, at five motion frequencies
| Base motion | For 90 % reduction | For 80 % reduction | Break-even | So what |
|---|---|---|---|---|
| 0.1 Hz — long swell, 10 s period | 160 ms | 320 ms | 1.7 s | Delay is not your problem here. Stroke is. |
| 0.2 Hz — where most ship motion energy sits | 80 ms | 160 ms | 830 ms | A well-built loop reaches this comfortably. |
| 0.5 Hz — short seas, deck machinery | 32 ms | 64 ms | 330 ms | Sensor filter settings now matter more than the platform. |
| 1 Hz — vehicles, structural response | 16 ms | 32 ms | 170 ms | A shared general-purpose network has already spent the budget. |
| 2 Hz — machinery, road and rail input | 8 ms | 16 ms | 83 ms | Needs a dedicated sensor and link, or prediction. |
Read this as a budget, not a specification. It assumes pure delay, correct amplitude and no prediction — real systems beat the last column by predicting ahead and fall short of the first by getting the amplitude slightly wrong. What it is good for is arithmetic you can do before anybody quotes: add up your sensor’s output filter, your link, the command interval and the servo response, multiply by the frequency you care about, and see which column you are in.
Do you know your loop delay, or only your sensor’s update rate? Send both and we will return the delay budget with our share of it filled in →
How compensation work actually divides
The 4-Configuration Compensation Stack
Four configurations, and in each one a different property governs the design. Reading which one you are in is more useful than a payload figure, because it tells you which requirement will end up dictating everything else.
C1
Sensor and antenna compensation head
Pointing governs · commonly under 300 kg on the moving frame
A small, stiff, fast platform under an instrument or terminal. What is being bought is angular accuracy at the aperture, and the number that matters is stated in degrees or milliradians, not millimetres.
Inside the supply
- Mechanism, controller, drives and cabinet as one delivered unit
- Compensation reference point set at the aperture rather than the base plate
- Interface to your motion sensor, with the coordinate convention frozen in writing
- Loaded acceptance against a base-motion record you supply or agree
Not included: the instrument or terminal, its radome, and the pointing or tracking logic that decides where it should look.
C2
Personnel or cargo platform
Safety and stroke govern · deck, structure and rated occupancy as one load
Once a person or a load crosses onto the platform, the design question stops being accuracy and becomes what happens at the edges: at the end of the stroke, on a lost sensor input, on a stop, on a power failure.
Inside the supply
- Envelope mechanically limited to what the operation was assessed for
- Hard-wired stop circuit and a defined resting attitude on stop
- Named behaviour on loss of base-motion input, agreed before manufacture
- Status signalling to the host so the operation knows the window is closing
Not included: the gangway, walkway or transfer structure as a certified product, the operational procedure, and the class or authority approval of the installed system.
C3
Continuous-duty compensation table
Duty cycle and thermal margin govern · hours, not minutes
A test platform works for ninety seconds and rests. A compensation table works for a watch. The mechanism is similar; the sizing, the thermal design, the bearing selection and the service interval are not.
Inside the supply
- Sizing run against a duty profile rather than a peak demand
- Thermal and service margin stated for the operating pattern you describe
- Environmental specification written down — salt, humidity, temperature, ingress
- Condition and status data available to the host during operation
Not included: the process happening on the table, its tooling and fixtures, and the vessel’s power quality and supply arrangements.
C4
Large deck or custom geometry
Structure and installation govern · beyond 2,000 kg, or a shape the deck already decided
Where the platform is large, or where a deck, hatch or existing structure fixes the footprint, the geometry is drawn around constraints that cannot be changed later. Installation is designed at the same time as the mechanism.
Inside the supply
- Geometry laid out around the space and the payload that exist
- Structural analysis against the mounting arrangement, not a nominal one
- Interface loads issued for the deck or foundation designer to work against
- Installation, commissioning and acceptance on site
Not included: the deck structure and its reinforcement, the seafastening, the electrical installation, and any survey or approval of the vessel.
C3 is the configuration most often bought as C1. A platform sized against a peak demand will meet it and then overheat halfway through the first shift, and the difference is settled by a duty profile in the enquiry rather than by a warranty claim afterwards.
Between two configurations because the operating pattern is not fixed yet? Send the worst realistic day and we will size against that →
Which mechanism sits under the payload
The platforms these configurations are built on
In compensation the axis count is a bandwidth decision as much as a geometry one. Every axis you add is mass the actuators have to move before they can do anything useful, and mass is the enemy of the one thing this application needs: answering quickly.
Three axes
3DOF motion platform
Heave, roll and pitch from a serial mechanism. Where the disturbance that hurts is vertical and angular — which describes most sea states — three axes carry it with less moving mass than six.
Six axes
6DOF motion platform
All six axes moving together on a parallel mechanism. Necessary when surge and sway reach the payload, when the lever arm turns hull rotation into translation, or when a person is crossing the platform.
Hexapod geometry
Stewart platform
Six legs laid out around holding an attitude rather than travelling. Chosen where the payload must return to a stated pose and stay there, and where the pivot belongs at the aperture rather than at the base.
The Axis Read from the Disturbance Start from what the base actually does, not from what sounds thorough
| What the base motion carries | What to specify | Why |
|---|---|---|
| Roll, pitch and heave, with the payload close to the sensor | 3DOF | Three axes answered quickly beat six answered late, and late is the failure mode here |
| A long lever arm turning hull rotation into translation at the payload | 6DOF | The payload sees surge and sway the hull barely has; three axes cannot remove them |
| A line of sight only, with translation irrelevant | Gimbal, not a platform | A platform would carry mass and stroke that the requirement never uses |
| Holding an attitude and returning to it, rather than following a path | Stewart | The figure that matters becomes repeatability at a pose, not travel between poses |
The trade-off is worked through on 3DOF vs 6DOF. Electric vs hydraulic matters more here than anywhere else on the site: a compensation platform runs continuously in a space where people work, and a servo electric machine has no fluid to leak onto a deck and no power unit running through a watch.
Have a motion record but no view on axis count? Send the record and we will report which axes actually carry the energy →
Delivered compensation platforms
Three projects, and what actually governed each design
Customers stay unnamed. What travels is the sector, the scale, the job the platform was doing and the one requirement that ended up dictating the build. In none of the three was it the payload figure.
MARINE UAV OPERATIONS
6-DoF Wave-Compensation Platform for UAV Take-off and Landing
A six-degree-of-freedom platform receives an approved vessel-motion reference and generates corrective motion to reduce deck movement during approach, touchdown and launch. The governing requirement is not UAV mass alone, but the residual motion allowed at the landing surface—together with sensing delay, prediction horizon, platform travel and the safe response when the disturbance exceeds the approved envelope.
MARINE SURVEY / OFFSHORE RESEARCH
Continuous-Duty Stabilisation Table for Shipborne Instruments
For survey instruments, inspection equipment, calibration fixtures and other precision payloads, the platform provides a more stable working reference while the vessel continues to move. The system is sized against a realistic watch or duty profile rather than a short peak-motion test. The governing requirement is sustained operation: disturbance spectrum, moving mass and centre of gravity, thermal margin, marine environment, service interval and residual motion at the approved payload point.
OEM SENSOR & PRECISION EQUIPMENT
Motion-Compensated Base for a Shipborne Payload System
The platform operates as a subsystem beneath an OEM sensor, antenna, optical assembly or precision tool, exchanging reference, command, status and fault data with the host controller. The governing requirement is the interface: coordinate convention, time synchronisation, compensation reference point, latency budget and behaviour at saturation or loss of input. Where only line of sight must be held and translation is irrelevant, a gimbal may be the more appropriate architecture.
We do not publish customer names, vessel names or project identities without asking first. Where a tender needs references, we approach the customer concerned and pass on whatever they are willing to put in writing — slower than a logo wall, and it survives being checked.
Recognise one of these three arguments in your own project? Name it and we will send how that one was resolved →
Filmed on a moving base, not against a white wall
What compensation looks like when it is measured rather than described
Three clips, each showing something a datasheet cannot. A platform filmed moving smoothly on a workshop floor proves the motors turn, which nobody doubted.
Both numbers in one frame. The base motion being applied and the residual at the payload, measured at the same instant by an independent instrument — not a before-and-after cut.
The saturation case, deliberately. The base pushed past the platform’s travel so you can see how the correction rolls off and what the host is told while it happens.
Input pulled mid-operation. What the platform does the moment the sensor stream stops — the behaviour that has to be agreed in design, filmed end to end.
Want to see the failure case before the demonstration? Ask for the saturation and input-loss clips first →
The input most compensation enquiries leave out
Where the base motion comes from, and how late it arrives
Almost every enquiry states a residual target. Almost none states how the platform is supposed to find out that the base has moved. Those two are the same question asked from opposite ends, and the second one has the answer in it.
The 3-Source Base-Motion Ladder Plus the fourth thing people mean when they say compensation
| Where the motion comes from | What it gives you | What it costs | What it cannot fix |
|---|---|---|---|
| Your existing inertial unit | No new hardware, and a sensor already accepted on the vessel | Its output filter and the link it runs on are both inside the loop and outside our control | Its own published accuracy, which becomes the floor of the whole system |
| A sensor dedicated to compensation | The shortest and best-known delay, and a mounting position chosen rather than inherited | One more device in the scope, commissioned and maintained | Little — where the target is demanding, this is the row that gets you there |
| A prediction fed from outside | Delay bought back, sometimes enough to change which column of the ceiling you are in | The prediction’s own error joins the residual, and it fails differently from a sensor | Confidence. A platform acts on a bad prediction exactly as firmly as on a good one |
| Nothing — a stored profile | A recorded motion replayed on demand, repeatably | Almost nothing; it is the simplest thing on this list | Today’s sea. This is playback, not compensation, and it belongs on the testing page |
Row two is worth more than most projects expect, and row three is worth less. A dedicated sensor mounted where we ask removes two unknowns at once — a filter setting nobody can change and a lever arm nobody measured — and it does it for a fraction of what a larger platform costs. Prediction is the opposite: it looks like free performance and it is really a transfer of risk from the mechanism to a model.
Already own the motion sensor this will run on? Send its make, output rate, filter setting and mounting position →
How to write a number that survives a witness test
A residual figure means nothing until five things are attached to it
“Residual under 10 mm” is not a specification. It does not say where on the payload, under what base motion, in which axes, as a peak or an average, or measured by what. Change any one of those and the same machine produces a different number, honestly.
This matters commercially more than technically. Two quotations that look ten per cent apart are usually thirty per cent apart on conditions, and the cheaper one is often quoting a root-mean-square figure at the base plate under a mild spectrum while the other quotes a peak at the aperture under the record you sent.
We write the five conditions into the quotation, into the acceptance protocol and into the delivered record, in the same words each time. It makes our figures look worse than they need to on first reading, and it makes them mean something at the FAT.
A residual quoted as one number is a marketing figure. A residual quoted with five conditions is a test you can witness.
Comparing two compensation quotations right now? Send both figures and we will tell you which conditions are missing →
The 5-Part Residual Figure What has to travel with the number, every time it is written down
| Part | Why it changes the number | If it is left off |
|---|---|---|
| The point it applies to | Residual grows with distance from the compensation reference point; an aperture two metres up is not the base plate | The figure is quoted where it looks best rather than where the work happens |
| The base motion it was measured under | A spectrum, or a named record, with its band and amplitude — the same platform reports different residuals under different seas | Acceptance runs under whichever motion makes the number pass |
| The axes it covers | Angular residual and translational residual are different quantities and usually differ by an order of magnitude | A good angular figure is read as a whole-system claim |
| The statistic and the duration | Peak, RMS and a 95th percentile over the same run can differ by a factor of three | Two quotations are compared that were never measuring the same thing |
| The instrument that measured it | A residual read from the platform’s own encoders cannot see what the platform got wrong | The system marks its own homework |
Two outside references are worth reading before you write a target. Marine motion sensor datasheets commonly publish heave accuracy as “5 cm or 5 %, whichever is greater” — check the unit you already own, because that figure sets the floor for translational compensation. And for a worked example of how residual is reported in published work, Mu, Zhou, Li and Liu (2026), Journal of Marine Science and Engineering 14(2) 187 report heave variation held within 1.6 cm on a 1:10 scale gangway model in a wind-wave basin, with the sea state, the scale and the method all stated alongside it. Those are their figures on their apparatus, cited as such, and the reason to read it is the reporting format rather than the number.
The behaviour that gets decided in commissioning if it is not decided in design
Every compensation platform runs out. What matters is how it does it
There is a sea state, a road or a manoeuvre beyond any envelope. A compensation platform that has not been told what to do at that moment will do something anyway, and it will do it while people are working on top of it.
Failure one
Out of stroke
The base asks for more travel than exists. The wrong answer is to track until the mechanical limit and stop dead, because the payload then gets a step it never got from the sea.
Failure two
Out of acceleration
Stroke is available but the platform cannot get there in time. The residual grows smoothly rather than suddenly, which makes it the failure nobody notices until the data is analysed.
Failure three
Input lost
The sensor stream stops, or goes wrong quietly. A platform still acting on stale data is more dangerous than one that has stopped, because it looks like it is working.
None of the three is a defect and none is avoidable by buying a larger machine. What is avoidable is discovering the behaviour on the day. Each of the three is written into the interface document before manufacture, demonstrated at the FAT rather than described, and filmed — the second and third clips in the video section exist for this reason.
The 8-Line Compensation Brief What we need before a residual figure means anything
| Line | What it changes on the platform | If it is left blank |
|---|---|---|
| Base motion: a measured record, or a spectrum with height and period | Axis count, stroke, acceleration, and whether the job is possible at all | We quote against an assumed sea and it is the wrong one |
| Where the compensation point sits on the payload | Every residual figure the project will ever quote | Figures stated at the base plate, which is not where the work happens |
| Payload mass and CG height above the interface | Actuator force, achievable acceleration, and how much travel survives | Sized for a low centred payload; a tall one spends its stroke on tilt |
| Motion sensor: make, output rate, filter setting, mounting position | Two of the three links in the chain, and the lever-arm transform | The delay budget cannot be assembled, so no residual can be promised |
| Total loop delay you can sustain, or the parts of it you know | Whether the target is reachable before any hardware is chosen | A target is agreed and later found to be arithmetically impossible |
| Duty profile: hours per day, continuous or intermittent | Motor and drive selection, thermal design, service interval | Sized against a peak and then run through a watch |
| Required behaviour at saturation and on loss of input | Controller architecture and the safety design around it | Decided during commissioning, by whoever happens to be present |
| Environment and any approval scheme the installation must satisfy | Materials, sealing, documentation and the whole test regime | A machine that cannot be accepted on the vessel it was bought for |
Eight lines, and compensation enquiries typically arrive with two: a payload mass and a residual target. Write “unknown” against the rest — an unknown gets asked about, an estimate gets built to and then defended at the acceptance test.
Cannot fill in the delay line because nobody has measured it? Send the sensor and the network and we will assemble the budget with you →
From a motion record to a number somebody can witness
Five steps, each ending in a document rather than an agreement
Compensation projects go wrong at the two ends: a target agreed before anybody resolved the base motion, and an acceptance test designed after the machine was built. These five steps put both at the front.
The 5-Step Residual Verification Path What each step produces, and who it is for
ST 01
Resolve the record
Your measured motion or an agreed spectrum is transformed to the compensation point through the geometry as measured, and we report which frequency band actually carries the energy.
Out: resolved base motion
ST 02
Budget the delay
Every element in the loop is listed and added up against the band from ST 01 — sensor filter, link, command interval, servo response — with your share separated from ours.
Out: delay budget
ST 03
Check the envelope
Stroke and acceleration are run against the tail of the record at your load and CG height, not its average, and the saturation rule is chosen rather than inherited.
Out: envelope & saturation note
ST 04
Replay and measure
The base motion is applied and the residual measured with an instrument independent of the platform’s own encoders, in the axes and at the point the target names.
Out: residual record
ST 05
Loaded acceptance
Repeated at the delivered load, witnessed on site or by live video, and issued with all five conditions attached to the figure rather than beside it.
Out: acceptance record
Need the acceptance protocol before you can raise a purchase order? Ask for the blank protocol with your five conditions written into it →
Where the machine stops and the operation begins
What we supply, what the integrator owns, and what we will not sell you
Compensation scopes go wrong in a particular way: a supplier agrees to everything, and the gap between “a platform that compensates” and “an approved system on a working vessel” is found by whoever is standing on the deck. These four zones go into the proposal so the gap is visible while it is still cheap.
Ours
- The mechanism, controller, drives and cabinet
- The compensation loop from your motion input to platform motion
- The interface, at the level fixed in the interface document
- Workspace protection, mechanical limits and the stop circuit
- The residual record, with its five conditions attached
Never ours
- The vessel, the deck structure and the seafastening
- The motion sensor as a certified navigation device
- The decision to begin, continue or stop an operation
- Class, flag or authority approval of the installed system
- The payload itself, and what it is being used for
Outside our class of machine
- Hull-level stabilisation — fins, gyros, anti-roll tanks
- Heave-compensated winches and cranes as rated lifting appliances
- Certified walk-to-work gangways as a complete product
- Sub-micrometre or nanometre positioning
Signed by both
- The base motion the design is against
- The compensation point, and the residual target with its conditions
- The delay budget, itemised, with each side’s share
- Behaviour at saturation and on loss of input
We would rather be left out of a scope here than be the reason an operation looked acceptable on paper.
Not sure whether your scope needs a platform or a certified system? Describe the operation and we will mark the line where our supply ends →
Before the specification is written
Reference reading for a compensation specification
Four pieces that answer the questions this page raises but does not settle: how many axes, what the duty costs, where the same hardware is used to create disturbance instead of removing it, and the vocabulary a specification has to use.
Axis decision
3DOF or 6DOF, read from the disturbance
What three axes cannot remove, what six cost in moving mass, and why moving mass is more expensive in compensation than anywhere else.
Trade-off
Payload, stroke and acceleration under a duty cycle
Why raising one of the three lowers the others, and why a machine running a full watch is sized against different numbers than one running a test.
The other direction
The same hardware, used to create the disturbance
Reproducing a base motion in a laboratory so equipment can be qualified against it — the mirror image of this page, and often the step before it.
Vocabulary
Motion platform glossary
Gross moving load, motion reference point, combined-axis travel, residual — defined once, so a specification can be written without ambiguity.
Two outside references are worth the time before writing a compensation scope. IMCA’s commentary on raising standards for walk-to-work sets out how ownership, design standardisation and competence are being tightened around motion-compensated transfer — useful whether or not people are on your platform, because it is where the documentation expectations are heading. And the 2026 Journal of Marine Science and Engineering study of a vessel-borne compensated gangway is a clean worked example of stating a residual with its conditions. Both are other people’s work, linked because they are good, not because we were involved.
Asked before a specification is written
Motion compensation platform FAQ
What residual motion can you guarantee?
Not a number, on its own, and not before we have seen the base motion and the sensor. Residual is set by three things: how well and how fast the base motion is known, how long the whole loop takes to answer it, and whether the stroke covers the worst of the record. Two of those three usually belong to the customer. What we will do is take your motion record and your sensor details, return a delay budget with our share itemised, and quote a residual with the five conditions attached — at which point, under which base motion, in which axes, as peak or RMS, measured by which instrument. A supplier who gives you a single number before seeing any of that is quoting the mechanism and hoping the rest of the system cooperates.
How is a compensation platform different from a simulator or test platform?
The hardware is the same family; the objective is inverted, and that changes three things. Sensing: a simulator needs no knowledge of the outside world, a compensator is useless without it. Delay: a simulator can be a few tens of milliseconds late and nobody notices, a compensator that is late makes things worse in a way that is arithmetic rather than opinion. Duty: a test platform runs a profile and rests, a compensation platform runs for a watch, which changes motor sizing, cooling and service life. A platform built for one is not automatically suitable for the other, and the difference is mostly not in the mechanism.
Can you use the motion sensor we already have, or do you supply one?
We integrate the unit you already own in most projects, and that is usually the right call — it is already installed, already accepted, and already trusted by the crew. What comes with it is that its output filter and its accuracy specification are inside our loop and outside our control, so both go into the delay budget as your rows. Where a target is demanding, a second sensor dedicated to compensation and mounted where we ask is often cheaper than the larger platform that would otherwise be needed, because it removes two unknowns at once. We can specify and integrate one. We do not certify any sensor as a navigation device.
What happens when the sea is bigger than the platform’s stroke?
Something, always — so it is designed rather than discovered. The behaviour is agreed before manufacture and covers three cases: running out of travel, running out of acceleration, and losing the motion input. In each case we fix which axes keep priority, how the correction rolls off instead of clipping to a hard stop, what status the host is given while it is happening, and where the platform comes to rest if it has to. A platform that tracks to its mechanical limit and stops dead hands the payload a step it never got from the sea, which is worse than not compensating. All three cases are demonstrated at the factory acceptance test rather than described in a manual.
Can you supply a certified walk-to-work gangway or personnel transfer system?
No. We supply the motion element and the records that describe it. A transfer system that carries people between a vessel and a structure is a certified product in its own right, and that certification belongs to the company that builds the system, not to whoever supplied the platform under it. The operation itself sits under industry guidance, with the asset owner, the vessel operator and the system supplier between them. Where we are the motion supplier into such a scope, we will say exactly that in writing, and we will not let a proposal be worded so that our supply reads as the certified article.
Do you work to classification society requirements — DNV, ABS, CCS and the rest?
It has to be in the specification from the first day, because almost none of it can be added later. A scheme changes material certification, welding and inspection records, documentation, the factory test regime and often the electrical design. Tell us which scheme and which scope at enquiry, and you will get a direct answer in the first reply, including where the answer is that we cannot support it. What we will not do is accept a scope on the assumption that the paperwork can be arranged afterwards.
Can it run continuously, or is it a machine that works in bursts?
Both exist and they are not the same machine. A platform sized against a peak acceleration will meet that peak and then heat up halfway through the first watch; one sized against a duty profile carries more thermal and service margin and usually less peak. Which you get depends on a line in the enquiry: hours per day, continuous or intermittent, and what the realistic worst day looks like. Continuous-duty machines are configuration C3 on this page, and the acceptance test for one is run over duration rather than at maximum demand.
Why is your quotation higher than a supplier offering the same payload and residual?
Compare the conditions before the figures. Ask at which point on the payload the residual was quoted, under which base motion, in which axes, whether it is a peak or an average, and whether it was measured with an instrument or read from the platform’s own encoders. Then ask whether the quotation contains a delay budget at all, what it assumes about your sensor, and what the machine does when the stroke runs out. Those questions account for most of the gaps we see. Where a competing quotation genuinely covers the same conditions and costs less, we would rather hear about it than guess.
Commercial questions — lead time, payment terms, spares, installation — are answered in the full FAQ.
Send the base motion. We will tell you what is left.
A motion record or a spectrum, the payload, the sensor you intend to use and what has to stay still are enough for a first pass. If the requirement belongs to a gimbal, a winch or a naval architect, that is what you will hear back — and hearing it in week one is worth more than a quotation.
What comes back first
- Whether the fix belongs at the payload layer at all, naming the machine class if it does not
- The delay budget with our share filled in, and which column of the delay ceiling the project sits in
- Which of the four configurations it is, and whether three axes or six carry the disturbance
- A residual target written with its five conditions, plus a blank acceptance protocol