Patric's Stewart Platform (6DOF, AC, FlightSim)

Hello Everyone,
a new kid on the block trying hard to join the club.

My goal: successfully designing and building a 6DOF motion platform to take flight simulation to the next level.
This project already has a little bit of history, yet I am still far from taking off with that sim. Lot's of thoughts and prototyping has already been done and I would like to use this forum to share some of my ideas, discuss some option and keep a build log for myself and others.

Looking around for resources on the web I found lots of papers and theory on the design specifics for motions platforms. Being a passionate DIY guy I decided to start from scratch.

First thing was a design tool which I implemented to study 6DOF platform kinematics and parameter variations. I will try to upload the tool to the cloud, for now checkout this video:



Notice the servo arm animation, from the input attitude and position the required servo angles are computed. Requires quite some 3D vector calc. Image turns red if platform goes out of envelope. I have also added plots to explore the work space (6D space projected on 2D by chose pitch against roll, heave against sway etc.).

Experimenting with a lot of different configuration I came up with a first full scale design. The lower platform is nearly ready for production. I will work on the upper part once I have a better understanding on cockpit layout, seat dimensions etc.



Motors are 0.75kW ACs + 1:52 gearbox by Washtec (car wash technology given some real challenge here!).
Have not yet decided which VFDs to use I need some input from you guys, but lets discuss the later...

I'll be back with some cool footage on a proof-of-concept subscale which I built

A proof of concept seemd necessary, I was especially interested in getting the forward kinematics to work properly and showing exact angles and translations. A scaled down version was designed using of the shelf RC servos:



The original design is already 1.5 years old. In the mean time I was heavily involved in other projects, but coincidentally my primary job offered an interesting use case for the subscale model. We are developing an automatic landing system for helicopters and we are using full scale motion platforms to end-to-end test landing phases on ships in heavy sea. So here come the sub scale model to become the basis for a trade fair booth demonstrator:



The deck illumination is a bit overdesiged and distracting from the actual kinematics but you get the idea...

UPDATE: First servo test completed.

My test setup is using a single motor + 1:52 worm gear. I had purchased a Hitachi WJ200-004SF inverter rated for 0.75kW. I mainly chose this unit as it features a PID controller and can be fully controlled via modbus protocol on RS485. Ultimately I will daisy chain all 6 inverters and run them directly from the PC. No need for intermediate embedded controllers and such.

The test subject is to verify the basic servo concept, primarily playing with the inverter (insane amount of parameters available) und put some realistic loads on the mechanics.

Overall setup:



The inverter is setup in PID mode using a position encoder (10k poti) and receives position demand by a second poti. The hitachi features a 10V analog input (used for position encoder) and a 4-20mA standard current input. Just hooking up the poti to the current input does not work properly as the produced current is not linear (I=U/R). Instead I needed to put together a howland current source (https://wiki.analog.com/university/courses/electronics/text/chapter-4) to get usable linear current control.



Position encoder is held in place by some flexible plastic to compensate for any shaft misalignments.



Close up of the servo arm mounting. The final version will be a 15mm aluminum lever.


This is a free run without load. The rotation angle is limited only by poti angular range (ca. 240 deg). I will fit two endstop switches that will engage the DC brake.

SAFETY WARNING: make sure the lever can freely turn 360 degs. I messed up the settings several times and overran the intended angles. I may actually take this into account for the 6DOF design. The motor is extremely powerful and may easily chop off body parts. Make sure there is enough clearance around lever.



Now comes the load test. 18kg on a 18 cm lever arm, about 32Nm. The motor is wired in star configuration which gives 70% margin to delta config. No problem developing the necessary torque. Though when starting from rest with lever horizontal (max load), the motor stalls when applying small acceleration (note the whining noise). The 6DOF design will feature gas springs that should avoid such cases.

Did this project make progress ?