Case studies

Case Studies · Delivered project library

Projects we have delivered, recorded so you can check them

A working library rather than a brochure. Filter by industry or by platform type, open any record, and you get the same three things: the requirement that governed the design, the decision taken because of it, and what was measured at acceptance. New records are added as projects complete.

What every record contains

Three parts, or it is not published

A project that cannot supply all three does not enter the library. These are also the three questions worth asking any supplier about any case study — including this one.

Part 01

The requirement

The constraint that actually governed the design, not the industry it belongs to.

Part 02

The engineering decision

What was chosen because of that constraint, and what was traded away for it.

Part 03

The measurement

What was instrumented, at what load, and against what threshold.

The library

26 delivered project records

Filter by end market or by platform configuration — both routes reach the same records. Open any card for the requirement, the decision and the measurement. Records marked anonymised have the customer withheld and the parameters published unaltered.

Seven-axis helicopter cabin training platform

A cabin platform training six crew simultaneously, with a seventh axis mounted in series above the six-axis base to extend yaw beyond what the hexapod alone could reach.

6DOF

Configuration

6DOF

Configuration

6DOF

Configuration

6DOF

Configuration

Flight simulator platforms →

Seven-axis helicopter cabin training platform

A cabin platform training six crew simultaneously, with a seventh axis mounted in series above the six-axis base to extend yaw beyond what the hexapod alone could reach.

6DOF

Configuration

6DOF

Configuration

6DOF

Configuration

6DOF

Configuration

Flight simulator platforms →

Seven-axis helicopter cabin training platform

A cabin platform training six crew simultaneously, with a seventh axis mounted in series above the six-axis base to extend yaw beyond what the hexapod alone could reach.

6DOF

Configuration

6DOF

Configuration

6DOF

Configuration

6DOF

Configuration

Flight simulator platforms →
How to read this library

Three parts, and one disclosure

Most case study pages in this industry are photographs of finished machines with an adjective attached, which tells an engineer nothing. A record is only useful if it lets you predict something about your own project, and that needs all three parts below.

Part 01

The requirement that governed

Every platform has one constraint that determined the design; the rest followed. A 35-tonne marine table is governed by structural stiffness. A docking rig is governed by drift over hours. A driving simulator is governed by perception and duty cycle. Naming the wrong one produces the wrong machine.

Ask any supplier: Which single requirement sized this machine, and what happened to the others because of it?

Part 02

The decision and its cost

Engineering decisions are trades. Raising rated load lowers achievable acceleration for the same actuator. Widening the workspace moves the singularity boundary. A record that lists only benefits has had the engineering removed from it.

Ask any supplier: What did this configuration give up, and would a different choice have suited a different customer better?

Part 03

The measurement, with its instrument

A figure is evidence only if you know how it was obtained. Positioning accuracy derived from encoder resolution and positioning accuracy measured under load with a laser system are different quantities that look identical on a datasheet.

Ask any supplier: Measured or calculated, at what load, at what pose, with what instrument?

Naming policy

Which customers we name, which we do not, and why the difference is marked on every card

Our delivered base sits largely with defence establishments, aerospace and shipbuilding institutes, national research academies and universities. Many of those projects carry confidentiality obligations, and some involve equipment we are not free to describe at all. Publishing only the handful that happen to be nameable would misrepresent the whole; so would inventing consent that does not exist.

  • Named records carry a green badge. These are customers — typically overseas training providers, university laboratories, integrators and operators — who have given written permission for their name and project to be published. Where you see a name, permission exists in writing.
  • Anonymised records carry an amber badge. The parameters, the requirement, the decision and the measurement are all published unaltered. The customer, the equipment the platform represents, and our internal model designation are not — those designations encode the customer, so publishing them would identify it by inspection.
  • The featured project above is named because it was publicly reported and carries no confidentiality restriction.
Measured validation

What the third part of each record rests on

Every record above ends with a measurement. This is the apparatus behind those measurements, and the thresholds our platforms are designed and accepted against — published because a supplier who will not state a number in advance cannot be held to one afterwards.

Method 01

Factory acceptance under a stated load

Acceptance is run with a test load standing in for the real payload, and the load figure is recorded next to every performance number. A performance result without its load is not a result. The blank acceptance protocol — test list, instruments, pass criteria, result format — is available before you order.

Method 02

Laser measurement, not encoder inference

Position and attitude are verified externally with a laser measuring system. Encoder-derived accuracy describes the control loop’s opinion of itself; laser measurement describes where the platform actually is under load. The two can differ by an order of magnitude on a machine with joint compliance, and both are legitimately called “positioning accuracy”.

Method 03

Command against achieved, same run

For anything reproducing a profile, the commanded trajectory and the measured response are recorded together and compared, with amplitude and phase-frequency characteristics produced for road, flight and wave spectra. This is the only format in which a motion claim can actually fail, which is why it is rare. Ask for it from anyone.

Method 04

Analysis that has to agree with itself

Modal behaviour is solved twice by different methods — finite element in ANSYS and multibody dynamics in ADAMS — and workspace is analysed by Monte Carlo sampling rather than boundary assumption. Disagreement between methods is treated as a finding, not smoothed over.

Published design and acceptance criteria

These thresholds appear throughout the records above. They are stated once here so the records do not have to repeat them.

Criterion Threshold Where it applies
Continuous-operation drift ≤ 0.25 mm per cylinder over 12 h Sustained-duty platforms; 0.2 mm on some configurations
Loaded static hold ≤ 0.1 mm and ≤ 0.01° over 48 h Precision positioning and docking platforms
Cross-axis coupling ≤ 1% Six-axis platforms used for test and measurement
Sine waveform distortion ≤ 0.5% Sway, rocking and vibration tables
Random time-history reproduction ≤ 3% Spectrum reproduction platforms
Bandwidth at −3 dB ≥ 2 Hz, 0–15 Hz working range Test and measurement platforms
Single-cylinder steady-state error ≤ 0.1 mm Multi-axis platforms with independent control loops
Supply disturbance ±20% fluctuation or sudden loss No damage; normal restart on restoration of supply
Mechanical design factor > 3 × rated load All mechanical components, with FEA verification on cylinders

Read these correctly. They are design and acceptance criteria for the platform classes named, not a guarantee attached to every machine we build. The binding figure for any specific platform is the one in its factory acceptance record, measured at its load and configuration. If a number here matters to your project, put it in the specification and it becomes contractual — that is what it is for.

References

Asking us for something closer to your project

The library is deliberately general because it has to be publishable. A serious evaluation needs something narrower, and most of that can be supplied without breaching anyone’s confidentiality — provided the request is specific.

Step 01

Tell us the constraint, not the product

“A 1,400 kg cabin, 4.5 m ceiling, six hours a day, five days a week” is answerable. “A 6DOF platform” is not — it describes a hundred different machines in our own record.

Step 02

We identify the closest delivered configurations

From the full 215-record base, not only the 26 published here, with the constraint that governed each.

Step 03

You get the acceptance protocol, in blank

The tests, instruments and pass criteria your platform would be judged against — before you commit. This is the document worth comparing across suppliers.

Step 04

Where a reference call is possible, we ask

Some customers have agreed to act as references; most cannot. We will tell you which situation applies rather than leave the question open.

Step 05

Witness the acceptance test yourself

In person or remotely. This is the strongest evidence available in this industry and it is a normal arrangement, not a special concession.

What we will not do

We will not show you another customer’s acceptance report, project documentation or site photographs — for the same reason we would not show yours to the next enquiry. A supplier who shares one customer’s material freely is telling you what they will do with yours.

We will not describe a project as ours where our part was a subsystem, and we will not present a delivered configuration as evidence for a requirement it does not cover. If your numbers sit outside what we have built, we would rather say so at the enquiry stage than have it emerge during acceptance.

And if another kind of supplier suits your project better — a nanometre-class positioning specialist, or a local integrator for a small installation — we will say that too. The manufacturer market map sets out the categories.

About this library

Questions about the records themselves

Questions about working with us — quotations, lead times, warranty, shipping, installation and support — are answered on the main FAQ.

Because most of them cannot be. The delivered base is concentrated in defence establishments, aerospace and shipbuilding institutes, national research academies and universities, and a large share of those projects carry confidentiality obligations. Overseas customers who agree to be named are published with their names; everything else is published with the identity removed and the parameters unaltered. Each card states which of the two it is.

Yes. Every parameter comes from the delivery records in our product catalogue, which documents 215 delivered configurations across 15 platform lines with dimensions, rated load, travel, angle, velocity and acceleration for each. It is a record of what was built, not a range we hope to sell. Where a figure is a design or acceptance threshold rather than a measured per-machine result, the page says so.

Records are added as projects complete acceptance and the customer position on publication is settled. A record only appears once it can state all three of its parts: the governing requirement, the engineering decision and what was measured. A project we cannot describe in those terms is not published, because a case study missing one of the three cannot be used to predict anything about your project.

The library filters on two dimensions, because buyers arrive from two directions. Some arrive by end market: marine, flight, energy, automotive. Others arrive by configuration: 6DOF, 3DOF, compensation, positioning. Both routes reach the same records. If the combination you need returns nothing, that itself is useful information and worth telling us about.

Where a customer has agreed to act as a reference we will introduce you, usually after a mutual non-disclosure agreement is in place. Where they have not, we will not imply otherwise. Tell us early which application area matters, because it determines whether we have anyone we can approach.

That is the normal case. Of 215 delivered configurations very few are repeats, because the customer base is dominated by research institutes and defence establishments whose requirements are one-off by definition. What matters is not whether an identical machine exists but whether your governing constraint resembles one solved here and whether your numbers sit inside the envelope already demonstrated.

Not from another customer's project, for the same reason we would not show yours to anyone else. What we can supply before an order is the blank acceptance protocol: the test list, the instrumentation, the pass criteria and the result format. That is the document worth comparing across suppliers.

Start here

Bring us a requirement, not a product code

The most useful first message is not a request for a quotation. It is a description of what has to move, how heavy it is, how often it will run and what has to be true when it is finished. From that we can tell you which records here are genuinely comparable, and whether the numbers you need are inside our envelope or outside it.

Six inputs that make the first reply useful

Tell us these and we can usually send exactly what you need without a follow-up exchange:

Gross moving load. Everything that moves — payload, cabin, fixtures, occupants.

Envelope. Floor area, ceiling height, access route into the room.

Motion required. Axes, travel, and any recorded profile you need reproduced.

Duty. Hours per day, days per week, and for how many years.

The acceptance criterion. The one number that decides whether the machine is finished.

Interfaces. What has to talk to it, over what, and at what update rate.

Scroll to Top