6DOF Project, design considerations

You've probably noticed that we are currently working on a 3DOF motion platform. But you also surely know the effect that while you're building you learn new things and get new ideas. And of course, everybody always wants more... So I'm already planning the next project. I'd like to share my thoughts and ideas but I don't know if anybody is willing to join the discussion. If not you can see this as kind of theoretical BLOG which may or may not turn into an implementation in the real world, later.

As with every DIY project lots of decisions have to be made. So let's start with the first one: The main purpose will be helicopter flight simulation. I have full respect for the car race simulation community. It can surely be fun if you're adicted to motor sports but I personaly find surface vehicles boring. If you have ~$10,000 to spent for a motion rig you could also buy a used sports car and drive it in the real world. However, I flew (as passenger) in a small helicopter once in my life during my military time and I'll never forget that.

So I think I'd need at least 4DOF (heave, roll, pitch, yaw) to get a realistic feeling and the difference in cost and work is not far away from a 6DOF rig. It might be actually easier because machining 6 times identical parts is less complicated than to make different parts for a 4DOF rig.

Next question: VR goggles vs. "touch and feel" cockpit?
VR pros:

• gives an amazingly realistic feeling
• saves cost, work and cabling for control gadgets
• cheaper than multiple wide screen TVs or beamers
  
• easy to change to different models or completely different types of vehicles
• "open frame" rig sufficient

Cons:

• no haptic feedback, controls (apart from stick/yoke) are difficult to find
• can cause motion sickness if head vs. rig movement is not properly compensated

Although the pros outnumber the cons I personally tend to the "real cockpit" solution. I find it very frustrating and distracting having to reach for a switch with a "space mouse" multiple ties until I finally manage to hit it. Second, building a replica of a real cockpit can also be satisfactory work.
Of course, building a real cockpit means that I ideally have to close every gap so the pilot can't see anything of the outside world. The ultimate solution would be to use a 180° screen and 3 beamers (Absolutely amazing @GRAP !) But that's beyond my space & money budget.

There are different ways to attach the actors to the cockpit.

Commercial simulators often use the "overhead" arrangement (top left). The cockpit sits completely on top of the actator plane. This requires more vertical space but allows more surge and sway movement for a given actuator stroke. DIY rigs usually use an arrangement where the actuators are horizontally further away from the center (topp right picture). I don't know the exat reason but I guess one of them is to avoid requiring a ladder to mount the rig. The disadvantage is that yopu need more actuator stroke to achieve the same roll and pitch angles and it limits the horizontal movement and also roll&pitch angle. Horizontal spacers/struts have to be added to avoid that the actuator rods hit the nacelle (bottom picture). This requires even more space and actuator stroke.
Example for DIY rig: @PMVCDA
Example for commercial rig: Delft university

My advice is build the 6dof

You will always wish you had done it when the 4dof is finished.

The cost of 2x actuators will be insignificant over time

Hello

180 screen requires 2 projectors with short focal length.
I have 3 because it's a 180 with 1 more straight meter on each side. It's a horseshoe so I can see the sides all the way to the wings. It's easier than making a total curve of 240.
If you really have a lot of space, you can project onto a 360 sphere and have a visual of everywhere.

yes, a 6DOF is longer but not harder to build than a 4DOF because there is only 1 simple base and 1 platform to build.

Thanks for the hints but as I said I don't have the room for such a big screen. The whole rig has to fit in half a garage which is my biggest limitation (apart from the money and my wife complaining). The ceiling is only 2,5m high. If I can build the cockpit no higher than 1,5m that leaves 0,5m above and below it which is more than enough for heave. But if the screen is protruding far out it will hit the ceiling when pitching up.

I think I have to do some "simulation of the simulation" by building a cardboard model or by drawing everything in the 3D CAD. Would be a waste of money if I buy 6 actuators and then find out that they don't fit.

what projectors do you use ? what sort of costs are they ?

I bought the BenQ TH671ST DLP full HD 1920X1080 short focal and 1 zoom

thats a neat projector

do you have any pics / video of it running ? - what do you use for the projector screen ?

is there any noticeable delay in the image ?

Regards

Delay it s one game beamer, 19 ms, so it s good.
i m still in test. i just fisish the actuators last week.

screen ?
simple bed sheet to make the curve with a circle at the top and bottom.
I'm going to look at a soft plastic to make it better.

my simulation is in the dining room and everything has to be dismantled with just 1 person.

with the 3 projectors, I can use FS2020 fully in the options. 4080 super graphic cards

I just have 2 or 3 test photos. I use fly elise program to display the curved image.

Thats a cool idea ! Very easy to pack away. I bet its good for browsing the web too !

I use my sim mostly for driving sims, not sure if 19ms would be noticable ?

OK, back on the original topic. I still have to find out step by step

1. what are the dimensions of my cockpit
2. how much movement is possible without hitting the ceiling (simultanous heave and pitch up with a big screen in the front is the critical limitation)
3. what arrangement do I choose for the actuators
4. how much stroke do I need

#1 should be easy. I'll buy the seats, put them on a beer crate or something, and check the comfortable position of the pedals. I've found a good writeup for #3 and #4. It was quite hard to find on page 10 of the DIY actuator thread. Don't know why it's not sticky. Maybe most of the forum members (including myself) are not very comfortable doing math and theoretical analysis. I think it's still better to build a model from LEGO bricks or wood and cardboard than to draw and animate it in the CAD.
I've collected some pneumatic cylinders to use them as "model actuators". I mean, just as passive linear bearings to support the model rig and as visualization.

Wow, I just googled "fly elise". That really looks interesting. As I said, I fear I don't have enough space for a full 180° screen and 3 projectors. But something like the Cobra Venom could be possible. IIRC, "fly elise" is also mentioned there (see thread by @peter stolmeier).
Is the software for correcting the distortion of the curved screen and the blending of the overlapping areas affordable?

For 180 screen, you need 2 projectors.

Yes, fly elise is to project on any curves or 1 full sphere if you have enough place.
fly elise does the job but the automatic setup with webcam didnt work so i did the setup by hand. The price with usb key is little too much ...500e ... and update is for 2 years only . I find only this so no choice.

Unboxing time, again. The seats and harnesses have arrived.

To get a better feeling of the dimensions I made a scale model out of cardboard. (1:3)

This is the abolute minimum space required for two persons. Now I have to find the best pivot points for the actuators. A triangle with two corners in the front and one in the rear is easier and leaves more space for doors. But it applies the roll torque at the weaker side of the structure. Two corners in the rear and one in the front center results in more rigidity. But the drawback is the pedals are in the way. I'd have to put the actuators under the floor which limits heave and pitch. I'll try out...
Next thing to check is how much space I have for the screens.

I fear this ain't gonna work, at least not as I planned it, first.

My first plan was to put the joints relatively close to each other. This way I could get more pitch and roll angle and also more surge and sway for the same stroke length of the actuators. But this means that the rig would sit nearly on top of the actuators. Placing the joints under the seats and in the middle between the footwells saves only ~10cm of total height. And the front screen is behind the dashboard console reaching far out. It'll hit the ceiling unless I had at least 3m headroom.

So I have to do it like everybody else, here, and place the actuators in a circle around the rig with the top joints well above the foot level.

But this means that everything becomes much bigger, especially, I need waaayy more stroke for the actuators, something around 700..800mm instead of ~500.

Looking at the cockpit of an R66 I had the idea of arranging the monitors like this.

Using portrait instead of landscape orientation for the two in the middle would provide more screen area at the bottom behind the dashboard.

I added the actuators to my model

It looks a bit asymetric but the idea was that the pilot's head is nearly above the center of the hexapod base circle.

The diameter of the upper joint circle is 105cm (scaled down 35cm), the lower is 140cm (46cm). In the lowest position the bottom of the rig is 70cm (23) above the ground, in the highest possible position the top is 240cm (80) above the ground. So it should fit into the garage with 2,5m ceiling height.

Maximum pitch can only be applied with the COR in neutral position. Pitch and heave cannot be both maximum at the same time, anyway. I have to mount the screens and check if the top edges collide with the ceiling. Maybe I have to make the base circle a bit larger. This way, the triangles formed by the actuators become even more obtuse-angled and the forces rise. But in the lower position there's not much rotation possible so the forces are shared evenly and this shouldn't be a problem.

I also found out that sway causes quite some limitation to the possible roll angle, same for surge and pitch. Especially when swaying to the left, there is still much roll to the left possible but only very little to the right.

This is good for car racing where the rig "leans outward" during turns to simulate the cetripedal forces. But for a helicopter the rig has to do pendulum movements around the rotor head which is above the rig. So when it sways to the left it also has to roll to the right. So an inverted hexapod bolted to the ceiling should perform better. Can anyone confirm this or am I wrong?

https://youtube.com/shorts/p3xqZ69iWTc

Edit: Woah! The youtube video skyrocketed. >6k clicks the last 24h. I should have included the panda then it would be 6Mio.

Hmm. If I simulate motion with @Erik Middeldorp 's spreadsheet and enter my dimensions (105cm platform, 140cm base diameter, 95cm actuator neutral position, 71 min 119cm max) I get fairly good ranges. I can sway or surge +/-200mm and still roll or pitch in all directions by +/-30°. Horizontal movement + yaw by +/-25° is also possible at the same time. This is pretty good, I think.

However, heave limits all other movements quite drastically. With +/- 200mm heave pitch and roll is limited to ~10° and yaw to only 5°. But as I have to limit heave, anyway, due to my low ceiling, it's not all that bad, overall.

With the model it feels like I hit the endstops of the Festo cylinders earlier. I have to check if the cause is the slightly worse stroke/length ratio or if the pushrods collide with some other part of the rig. The ball joints are also not perfect. They have less angular range than my U-joints.