Breaking into the world of auto racing is easy. Step 1: Buy an expensive car. Step 2: Learn how to drive it without crashing. If you’re stuck at step 1, and things aren’t looking great for step 2 either, you might want to consider going with a virtual Porsche or Ferrari and spending your evenings driving virtual laps rather than real ones.
The trouble is, that can get a bit boring after a while, which is what this DIY motion simulator platform is meant to address. In a long series of posts with a load of build details, [pmvcda] goes through what he’s come up with so far on this work in progress. He’s building a Stewart platform, of the type we’ve seen before but on a much grander scale. This one will be large enough to hold a race car cockpit mockup, which explains the welded aluminum frame. We were most interested in the six custom-made linear actuators, though. Aluminum extrusions form the frame holding BLDC motor, and guide the nut of a long ball screw. There are a bunch of 3D-printed parts in the actuators, each of which is anchored to the frame and to the platform by simple universal joints. The actuators are a little on the loud side, but they’re fast and powerful, and they’ve got a great industrial look.
If car racing is not your thing and you’d rather build a full-motion flight simulator, here’s one that also uses DIY actuators.
Thanks to tips line regular [baldpower] for this one.
32 thoughts on “Homebrew Linear Actuators Put The Moves On This Motion Simulator”
Sadly, he is not using Inverse kinematics math for the motion cues, and its moving a little rough. Its far from being ready for realistic simulation
Can you show us your version?
If he is the real thanos…You did the worst sarcastic question….
Why is this idea so prevalent now a days? and i specifically mean the “if you haven’t made one or had the experience then you are not allowed to talk” idea, because that is exactly what you are saying.
I mean yeah sure someone who has built one will definitely have more experience and knowledge to talk about the subject, but anyone who has any experience about robotics understands how PID controllers and inverse kinematics come into play. The truth is that it is far from being ready for simulation but i don’t see anyone claiming that it is.
So why bring up ” Its far from being ready for realistic simulation” if no one’s claiming such? The rest is understandable as an outcome of “not trying to be a realistic simulation”, although if one desires to be it’s a good first step.
This is mine:
Not sure if it’s a fair comparison, since I had a much larger budget. ;)
Good “Rock ’em Sock ’em Robots” arena. :-D
It’s a visual dome:
[Wolf] Helo simulator?
[Ren] Yup, that picture is of an AS350 flight simulator.
It looks like they get all the data from the program (game) and others have far better results in the motion so I wonder if what you are alluding to is, perhaps, him not having got the PID controller dialled in for the motors or the wrong settings for exporting the data (which he does say at the end of the forum thread he’s still doing)
I think my suspension might be set a bit harder than your purple pickup :)
Not yet, but it will…
I wonder if he has an emergency stop button I can’t see? Knowing my history of making mistakes in my code and an unequal amount of bad luck. Could easily see myself testing and run into an infinite loop of it slamming me side to side at full speed. Upset a disc in my back and not able to jump off. Cat would just sit there staring at me and be of no help at all. Ya know, the usual stuff.
That a VX220 in his garage?
Having read through most of the forum thread – yes and yes :-)
Going by his wiring diagrams, he’s put two emergency stops on the power-supply line which doesn’t strike me as a great idea as I didn’t think they were suitable for high-current loads but better than nothing and at least not subject to dodgy software.
Good way to do it is a safety switch monitor (like a pilz pnoz). Can be expensive but easy to source on ebay for a couple of bucks.
Yes it seems over specs, but no, once a person is strapped on such a contraption, safety first.
Years ago at an “un-named” airline training facility, we had full motion Level-D FAA certified sims (in other words, the fidelity of the sim was in-separable from flying the “real” aircraft). So we would have folks ‘screwing-around’ and do all kinds of crazy $hit. Like trying to land at 300+ knots, trying to do hammer heads, etc, etc – the sim did NOT like it one bit – it would literally shake and make un-godly noises – from what I heard later, the code was changed to simply terminate the sim if any flight profile beyond pre-programmed limits were deliberately attempted. Since you don’t do aerobatics in a FAR part 25 passenger carrying aircraft.
I feel this is a mistake. You learn from playing at the edges of possibility. The ones monkeying around doing acrobatics and trying to land at crazy speeds are the ones you want at the controls.
Co-pilot: “Oh shit. Our instruments are wrong, and we we are coming in too fast and a gust of wind has flipped us upside down 100 feet from the runway. Engine number three is out. ”
Pilot: “I’ve seen worse. I got this. Lower the landing gear on the left side. We are going to need more drag there.”
“If you’re stuck at step 1, and things aren’t looking great for step 2 either, you might want to consider going with a virtual Porsche or Ferrari and spending your evenings driving virtual laps rather than real ones.”
Go a level up and manage a virtual team.
I think the same can be said for helicopters,
learn some basics in a realistic simulator before flying the real thing.
(I’m reminded of the guy who bought a new helicopter and couldn’t wait for his Instructor to arrive, and tried to fly it solo… CRASH!)
Question for the experts: Can you accurately simulate the G-forces of a car going through a corner on a device like that? Is the skill that I gain from a motion simulator actually transferable to a real car?
Well, “seat of your pants” flying under IFR, could end up very badly, as trusting what you “feel” without visual or instrumental input can lead to disorientation. There was a cockpit simulator that did not tilt or rise, (but would move side to side or rotate on the vertical axis) “experienced” pilots would often “crash” if they trusted the “seat of their pants” over the cockpit instrumentation, because they “felt” the plane was making a dive or otherwise no longer flying level and attempt to compensate for something that really wasn’t happening.
What I’m trying to get across is: you may not feel the actual G-forces, in the simulator, but you hopefully will have gained experience in other areas of operation, that encountering G-forces in the real thing will be just one new thing to adapt to, and not a bunch of new things.
Not on that contraption. Even on a professional Level D sim, the ability to simulate G-Forces are not there. USAF F-22, USN F/A-18 sims I’ve flown were only “procedural” trainers (ie. the functionality of the aircraft was fully present, and the goal of training was to familiarize pilots with checklists, and aircraft response to control inputs). There is no way in heck (current sim technology) to duplicate the sweat/fudge factor of trying to land on a small piece of moving/heaving real estate in the middle of the ocean.
Perhaps an automotive sim engineer can speak up, but it would seem the advantage of, say a Formula-1 sim would be to acquire the skills to know how to enter turns and when to accelerate/decelerate to avoid mishaps, etc. Provided the tactile control feedback (ie. steering resistance) and fidelity of the motion / visual cues were similar to FAA Level D sims.
As I understand the use in F1 as a complete non-expert, it’s about memorising the circuit as you cannot within the rules get the time to test on the real thing with the real thing.
And yes to everything else you said about the technology level employed, and probably beyond that with mm levels of accuracy for circuit mapping in all axis.
“[Fred G. Sanford] Even on a professional Level D sim, the ability to simulate G-Forces are not there. ”
Depends on the aircraft…
Our Level D Full Flight Simulators are able to generate all the G forces needed for commercial aircraft:
You can simulate some g force if the generated video is all the pilot/driver can see by tipping the sim but keeping the visual in the same position relative to the pilot. A flight sim generates an acceleration ‘force’ on takeoff by tipping the sim backwards but not giving the pilot any visual cues that the sim is tipped. The pressure on the back is interpreted by the body as an acceleration. It’s easy to tell from the outside when a sim is in the landing rollout… it tips forward to make the pilot think he’s sliding forward in his seat.
It’s a pretty limited range of g force and you obviously can’t generate sustained vertical g’s without reaaaaalllllyyyyy lllooonnnggg legs.
If you really want g forces… you have to create them: https://m.youtube.com/watch?v=G3ipv6DPXek
I imagine a VERY large room, with a self-contained sphere. Linked in several directions via strong cables, which move. Flight simulators need a lot more “Z” than cars.
Kind of like this one….
I think I remember seeing a flight simulator for jet fighters built into the end of a centrifuge. Couldn’t really do translations and it had probably some false rotation cues as the acceleration changed, but otherwise it could give a pretty realistic simulation of a flight including realistic G loads
Please be kind and respectful to help make the comments section excellent. (Comment Policy)