Erics 6DOF Ride Simulator - Ride-u-lator

It's ALIVE!! I got far enough along in the software to make the platform move up and down, about 2/3 of the possible heave. After slowly dialing it up, I finally jumped onto the platform, and it kept going with no significant change in speed and/or noise !!

Everything is talking, but I am having a hard time getting realtime control of the actuators with the degree of control I need. ModBus is slow - 15ms per request - so I am overshooting my targets substantially.

Yes, 15ms is too slow to attempt realtime PID control directly from computer to actuators...!!! Thats why I use external hardware for that in my sim... ;-)

For example the minimum refresh rate for new position updates, on my AMC, is 2ms while the P(ID) is calculated realtime using interrupts that means could do direction change in 200 - 500ns if needed!!!

No pun intended but most intelligent actuators out there are expensive cause contain PID algorithms that can run independently from the refresh rate of the computer... Modbus could be used for END-POSITIONS parameters then!

Regards, Thanos

Thanks for the info Thanos,

My drives are Indexing, so I can tell it to move to a particular location and it will go there exactly and autonomously. But... unfortunately, the index cannot be changed while it is running, so you would have to wait till it completes one index to re-aim it, which would not work in real-time control.

I have decided to model the actuator behavior in memory on the host computer: The drive is predictable (has a definable acceleration, deceleration and velocity), so if you start it jogging in one direction and stop it at the approriate time (decel_distance away from the desired spot), it will land right on the spot.

I now have the jog-to-position working, but I still need to handle redirecting it in mid-jog. Then I will have a working simulator!

Eric

I'm think that they can't really change direction in 200ns - 500ns. They may be able to accept a change of direction command that fast, but there remains a lot of inertia to overcome while decellerating. Unless these motors are giant, they can't stop a moving 300lb platform in 500ns.

Still, I wish these Emerson Epsilon drive had an interruptible Go To Position command. It would save me days of work. I just went home and confirmed that an index cannot be modified on the fly, nor can it be interrupted and restarted without coming to a complete stop.

So jogging is my only ModBus option (though these drives can also operate (more noisily) in Pulse Mode).

Eric

The Ride-u-lator is UP AND RUNNING! Videos are posted in the video area, or check these out:
http://www.youtube.com/watch?v=uqvwUMS3Gdg
http://www.youtube.com/watch?v=8YhKapkT6Lo

I managed to control the whole thing via modbus in C# by keeping my commands very tiny. I basically have a 60ms roundtrip to find position, decide whether to jog (long moves) or index (short moves), and send the commands. That's about 17 updates per second, and it seems to work great.

I had some trouble with drives failing out with following error even with a pretty high tolerance. When the drive would go into fault mode, the platform would often slam to the ground in the direction of the faulted actuator. Eventually, I wrote an auto-recovery routine, which has eliminated the slams to the ground.

We've tested the platform with a payload 270lbs of riders (me and my son) + the wood and the chair. Its right at the limits of the actuators.

I'd like to add preload springs to offset the platform weight plus maybe 150 pounds of rider weight. Then this thing could handle two adults, one really big adult, or maybe four kids.

My wife and youngest son both ride the Ride-u-lator in Rollercoaster Simulation mode.

6 of travel doesn't do much for the hills, but its still a blast to ride.

http://www.youtube.com/watch?v=HMoCr44FYT0

Whow! Really nice. Please keep posting!

Thanks Jo,

You are the first post since I posted the working details a week ago. I figured nobody was interested.

Eric

Hi Eric were interested and very impressed , great work you have done and are doing , a lot of us can only look on in dream about what your doing, does not mean were not there wishing you on. so keep it up good buddy this forum is built on sharing and lurking and loads o dreamers but all in the best possible way

Good to hear wannabe.... cause if you guys aren't interested, I can't imagine what my poor local friends are going through when I talk about my creation ;-)

Hi ebinary. What an awesome simulator you managed to build! Congrats!
I was away for the summer so I just discovered your progress. The sim seems to move right with smooth movement. You managed to find a very good communication procedure with your actuactor, I think.
Now did you work on the limiters for the platform? If you still using the 6dof soft, did you check out the KneeCap angle limiter? Maybe if this is still on it, it can be a good idea to ajust it see if don't limits the movements of your platform too much. There's also the Max deplacement limiter which have to be adjust regarding the size of the platform.
Hope you're having fun with your sim, if you builded it, don't forget to have fun with it!

Kind regards,

Robin

Hi Robin,

Thanks again for your contribution. There is a good chance that I'll eventually integrate similar kinetics directly into my modbus driver (which is complicated by itself), mostly to reduce the complexity of startup and to avoid the jitter of have your program sample ForceProfiler and then my program sampling your data.

But I would have been lost without your program to get things moving

After making the changes to support linear actuators in 6dofmaster, I really just came up with some reasonable minimums and maximums. I try to run so that your program is having the most extreme allowed movements and then I scale the movements by a percentage in my program.

As far as smooth movements go... the informational feed is great. My main problem come from driving my actuators. The drives only have two styles of moves:

• jogging moves (move in a direction at a particular speed and acceleration), which require the driver stopping at a particular moment. With about 60ms modbus sampling rate, its hard to stop perfectly.
  
• indexing moves: go to an exact location and stop. An index can't be changed without coming to a complete stop.

What I really need is an interruptible index move, that can change destinations without stopping first, but that is not available on these drives. I use a combination of jogging moves (for long moves) and indexing moves (for short moves). But the result is, if index moves are used, you will feel little stops along the way. It works great for car simulation, but smoother simulations (like planes) it is annoying.

Hi I understood your problem but right now I have no clue how to solve it!
Could you send me a copy of the 6dof software modified please? I would like to see how it's like with linear actuactor. Maybe I can also check if there are some tiny mistakes.

Keep up the good work,
See ya

Robin

The problem is not with your software... The servo drives do not suppport these moves.

I will send you the modified code after I clean it up a bit. Its not very useful to anyone else right now.

Eric

Hi Eric,

Your project is coming along nicely. Keep the updates coming.

Great job so far.


....Mike

Unfortunately, my 6DOF is disassembled right now, awaiting some TLC. I host an annual halloween haunted house ride in my garage each year and the motion simulator had to move out to make room. I just reclaimed the garage over the weekend.

Raceray: no, I haven't achieved my desired control on the servos yet. They work great for driving sims, but they can't handle more subtle requirements yet.

I am totally new in motion simulation, i am building a small model with RC servos and USB controller to teach myself the basic and i am interested in building a 6 dof platform for my flight simulator. But your platform is exactly what i would love to build. However i don't think that i can match the hardware, for now baby steps will do!
To resume just amazing work :clap: , thanks for sharing.

Ulrich

PS: the dark ride was also very cool! You must be the coolest house on the block for Halloween!! hehe :thbup:

Congratulations! Great job!

You might by chance pass the specifications of all components that were used?

And the control software is its development? I am wanting to build a simulator the same as yours, but I need more information. Could you help me?