Wooden 2 DOF Sim with integrated G-Seat, Fans, and Vibration

Yes, Motion Dynamics is 10 min drive away.
Ok, good stuff that you are looking for motors.

Much more complicated? How? got me curious

Well, at least you need to deal with 2 boards that need to be wired and with software that needs to be loaded onto the Arduino board.
In contrast to this, JRKs are ready to go after a little bit of soldering.

oh yes, of course, the JRKS are almost "plug and play", that and I am code hopeless, I could not even get my head around a game dash code, let alone a motion written code.

Hi again,

several months have past since my last post, but now I finally have something to show. Due to my permanent lack of spare time it is just a tiny step towards my motion simulator, but I still want to share it with you right now.

Instead of having every detail already planned, I want to experiment with my construction, for example by changing the positions of the motors, size of levers etc. and then see what configuration feels best.

For that reason, I have started with the creation of building blocks that I can arrange in different ways.

One of the finished blocks is the pivot, and the other is one of two motor „modules“ with connected lever on one side and potentiometer on the other. I don’t have any experience with welding, and so I just used 21mm birch plywood, screws, and thick steel brackets. My impression is that the pieces are quite robust and stable. At least I hope so, because otherwise it will probably hurt later on.

Yesterday I quickly connected the motor with my JRK and moved the lever via the configuration utility. It is really a plug & play thing and worked at once. I made a short video of the test run (see below). As you see, I can simply disconnect the potentiometer from the motor for manually adjusting the lever position. I think this may be of some value later on.

Of course the configuration settings are not perfect: When making quick movements the lever turns too far and then returns a bit. Besides, the potentiometer gear should be a bit smaller for increasing the lever’s movement angle. I will fix this later. For now I am really glad that it works.

About the motors: I use these ones as recommended by several experienced builders:
https://www.motiondynamics.com.au/worm-drive-motor-12v-24v-200w-180-rpm-20nm-torque.html

Here are some pictures of the motor module:


This is the pivot:


And here is the first test run of the motor:

Since I will probably need some time until I have the next part of my simulator ready, I would like to entertain you with pictures of my previous project. I just found them on an old hard drive.

I did it in 2012, and it worked without any motors. So it is actually kind of off-topic here because it has nothing to do with the xsimulator software. This is also the reason why I don't think it is worth an own thread.

What I did was mounting an old kitchen chair onto a wodden platform, connecting the platform to ropes and hooking it to the ceiling. When sitting on the seat, I was able to roll and pitch the platform by simply shifting my weight. This muscle-triggered movement was transferred to my PC where it was translated into joystick movements that in turn controlled the plane within my flight sim software.


How was that done? Well, I installed a gamepad app on my iPhone. This app worked via iPhone movements and was connected to the PC via its own wireless network. Now I just had to mount the iPhone onto the platform, and voila: shifting my body weight while sitting on the seat moved the iPhone which in turn controlled the plane.


This setup was extended with a TrackIR device.


This whole setup did somehow improve the immersion, but in fact it wasn't exciting enough to go on with the project.
I hope that this will be different with the current approach, and I am really optimistic after reading the experiences in this community.

Hi again,

since my Oculus Rift will probably not arrive before the end of June, I have some time left for finalizing the first version of my motion platform. This is the reason why I did not continue with building a second actuator like the first one, but instead spent some effort with building a linear actuator.

The main reason is that I think that for flight and rollercoaster simulations a larger movement would be beneficial. Besides, I may want to add "heave" at a later point in time.

The pictures show you what I did until today: The actuator is ready so far, except the wiring. And except performance and load testing!! I really hope that it will work. If it does, I will build a second one and make some pictures during the build process for sharing them with you. If it does not work, I will go back to my previous type of actuator for the first version of my platform.



Edit (or better: Added):
It was late yesterday when I wrote this post, but it should never be too late for mentioning sources of inspiration. So here are the projects from which I got most of the information for the actuator:

• http://bffsimulation.com/linear-act.php: This incredible guy has lots of fabulous projects documented on his site and even offers building plans for a price that is more than fair. I did not use the plan, but I took a huge amount of information from his descriptions and photos. On his site I also learned how important the end switches are. And the site drove my decision to go for the chain drive approach.
  
• http://www.simprojects.nl/diy_motion_platform_v.htm: On this site you see the design of a phantastic actuator that is driven by a tooth belt. Also a video is available that shows every detail of the construction.
  
• http://www.xsimulator.net/community/threads/seattimes-6dof-dc-build.6106/: An absolutely professional build, this one using a threaded rod.
  
• I browsed many other sites and also threads within this forum.

Really nice DIY actuator! I like the use of the limit switches and the chain drive. Is that a 10 turn pot for the position feedback? What kind of speed and force are you expecting to get?

Thanks, @cgodwin!

Yes, it is a 10 turn pot. I was looking for a pot with less turns, because the actuator only makes 2.5 turns from one end to the other, but the electronics store around the corner just had this one. It has 20kOhm, so that it comes to the recommended 5 kOhm for 2.5 turns. I could also change the gear ratio, but for testing purposes this setting should be fine.

About the design: I got my main inspiration from http://bffsimulation.com/linear-act.php. On this site I also learned what can happen if you don't use limit switches.

About speed and force:
The data sheet of the motor says:

• Rated torque: 20 Nm
• Speed: 160-180 RPM dependent on load

About the theoretical(!!!) force and speed:

• Force: 20Nm correspond to a weight of 2kg with a lever length of 1m. So with a lever length of 2.75cm (half sprocket diameter) it should be 727 Nm which corresponds to a weight of about 70kg.
  
• Speed: One turn corresponds to 3.14 x 5.5cm = 17.27cm. With 160 RPM, I get 46 cm/sec. (I need to add that I plan for a maximum linear movement of 30-35cm.)

That's the theory, but I have no idea about the real life values. For example there is more friction than I expected when manually moving the rod, although I use bearings for all moving parts and I worked as precisely as possible (at least with my limited set of tools).

And I have concerns that the moving rod may bend under load. It is a stainless steel tube with an outer diameter of 16mm and an inner diameter of 12mm, and inside is a 12mm threaded rod which I use for mounting purposes.

Hi again,

my first linear actuator prototype is ready so far, and I have done some basic tests today.

The wiring has become a little bit more complex than expected at the beginning.

Here is the wiring schematics:


The reason for this rather complex approach is that I wanted to solve the following tasks without uncoupling the motor axis:

1. perform the JRK learning procedure
   
2. release the motor in the case that one of the limit switches has been pressed

These tasks can now be performed via this control panel:


Here’s a brief explanation how it works:

The two switches labeled (1) must be toggled synchronously. I would have prefered a single switch instead, but I needed 6 separate contacts and only had switches with 4 contacts available.
The purpose of these switches is to toggle between manual mode and JRK mode.

In manual mode (both switches (1) in lower position), I can control the motor manually: I simply select the direction via switch (2) and then turn the potentiometer (3) for powering the motor. See this video for illustration:


This manual mode is primarily meant for facilitating the JRK learning procedure, which is demonstrated in this video:


In manual mode, the „up“ movement is automatically stopped by the upper limit switch. But even when the upper limit switch is pressed, I can still move the motor back down, and vice versa.

In JRK mode (both switches (1) in upper position), the motor movement is controlled by the connected JRK board, as shown in this video:


Now both limit switches are active. That means, if the motor exceeds the allowed movement range, it stops automatically.

Once one of the limit switches is pressed, I can release the motor again by pressing the release button (4). This button is hopefully also useful if the simulator is in a resting position in which gravity has lowered the actuator, so that the lower limit switch is pressed. In this case, I can switch on the simulation and press the „release button“, which should bring the actuator in its neutral position.

This final video shows, how the „release button“ works:


About my next steps: Actually I planned to do some load and performance testing and in case of success build a second linear actuator. However, I changed my plan: For now I will discontinue the linear approach and return to my first actuator build (see several posts above). I will do some little modifications for allowing a longer lever, and then I will build my first 2 DOF motion sim based on these simpler actuators.

So long for today,
Markus

Finally, today was the day on which my simulator did its first (unmanned) movements, linked with No Limits 2 Rollercoaster.
I did a very simple axis configuration with just roll and pitch.

What I don't understand is the following:
Initially both motors were in their middle position. When testing roll and pitch with the Simtools Game Engine, the movements seem good.
Then I started NLR2, and at the very moment on which the coaster started rolling, the left motor moved considerable downwards. There was neither pitch nor roll at this moment, and so I dont understand this.
The same effect occurs with different coasters. And the same effect occurs when I configure sway & surge instead of roll & pitch.
Only when exiting a track via the Esc key, the motors move back to their initial positions.

I am using the latest Simtools plug-in and the current steam-based version of NLR2

Does anybody know a possible cause for this effect? Any help is greatly apprechiated.

Thanks and best regards,
Markus

Here are screenshots of my Game Engine settings:


And here are the settings within the JRK Configuration Utility:

Right motor:

Left motor:


Or are there any other settings that I have missed?

Thanks,
Markus

@noorbeast: Thank you so much for your help and detailed information!

I will follow all pieces of your advice and post the results. The next time I will be able to spend some time with the simulator will be in about 2 weeks. Then I will also add some photos, and if it works also a video.

Best regards,
Markus

Hi All,

it's time for a status update.

Today I assembled my seat mover. Below are some photos.

Here are my 3 JRKs. Only 2 of them are in use at the moment. I plan to use the third one for adding a fan that produces an airstream.


As recommended by the experts, I applied heat sinks to the main chip and mounted a fan above the boards.


The 1000 Watt power supply is located below the JRKs.


Here is a picture of the lever. I made it rather long and with several holes for some experiments, because originally I planned a shoulder mount. I changed my mind as you will see below.


The pot is on the opposite side, similar to the pictures of my first prototype that I posted some time ago.


This is the universal joint, mounted to the baseplate and ready to hold the seat plate.




Seat plate mounted:


The rods connecting the levers with the seat plate:


And finally: The entire rig:


The whole construction is a bit too high, so that I need a rather high footrest. The reason is that I wanted the option for long levers at the motors that can make a full turn without hitting the baseplate. These long levers seemed reasonable for me especially when thinking about a shoulder mount.
Now I mounted the rods to the back of the seat, and I will probably have to drill a new hole into the levers for making them even shorter. At the moment the length is 6cm, and I have the impression that the motors are sometimes overloaded when I am seated.

After some PID tuning and experimenting with the axis assignment, I took a seat and made my first virtual, motion-enhanced rollercoaster ride with No Limits 2. I will post a video soon.

Here is a video of an unmanned No Limits 2 rollercoaster ride.

There seems to be a delay in the motor movements, which I will try to reduce by further PID tuning.
By the way: @noorbeast: Your PID explanations are great and very helpful. Thanks for your effort and for pointing me to this thread.
And @Avenga76: Also your video is phantastic. Thanks!

@noorbeast: As you recommended, I configured single axis at first and tested them separately.
Rolling works rather good in my opinion, but surge behaves really strange.

Here is what I observed: When applying surge as the only force, the seat moves considerably backward when the coaster enters the ascent with the chain drive. But actually at this time there should not be any surge accelertion. There is only a positive pitch movement. Then, when entering the highest point of the track, i. e. the point at which the chain drive ends, the seat moves forward. Also at this moment there should not be any change in surge force, but only a negative pitch movement.
Does anyone have an explanation for this? Maybe the No Limits plugin mixes surge and pitch in some way?

Best regards,
Markus

Ok, I will try this.

I already have the rod in a hole which is 6 cm away from the lever axis. Since the motors seem to be overloaded (they sometimes don't react on the JRK output signal when I am in the seat), I will shorten the length to 4 cm.