2DOF simulator based on CKAS Thruxim

[Mertensbart20]

Hello everybody,

I am a teacher electrical engineering and this year I am building a 2DOF racingsim with my students that is based on the design of the CKAS Thruxim. You can see the Thruxim here:

https://www.youtube.com/watch?v=lOzyqhhKTrY

Our design will be identical but our simulator will move faster and will have a pitch and roll of 18° (this one has 8°). Last schoolyear we built this prototype (thanks to the design of Thanos):

https://www.youtube.com/watch?v=q6HtIFWxAf4

Since we got this prototype running I’m sure the project of this year will succeed. The project has already started but since my students are working on it will take about 7 months to complete it. I As soon as we have something moving I will keep the project up to date.

Below you can find a list of the components we will be using (if you want details, just ask):

• a CX9010 PLC form beckhoff with 8 analogue inputs (0-10V), 8 analogue outputs (0-10V), two encoder modules, 8 digital outputs)
• two frequency invertors Siemens G120C (3x400V)
• two SEW-motors, 0.37kW, 15rpm, 152Nm, with break
• two 24VDC-relays for switching the motorbreaks on and off
• 24VDC powerblock for the relays
• two SICK rotary encoders (1024 pulses per rotation)
• steel pipes for the frames
• a motorcockpit according to the plans of ricmotech (http://www.ricmotech.com)
• logitech G27 steering wheel
• 23” LCD display
• 1 cardan-coupling (don’t exactly how it is called in English)
• a VM140 vellemankit
• the game rFactor

The list will be added with safety components. We will also be providing plans (not for the cockpit, those you can buy at the site of ricmotech), PLC-program, … . The students will also have to write a thesis which I will also upload at the end of the schoolyear. But this thesis will be in Dutch so I am afraid it will be hard to understand for most people on X-sim .

I’ll keep it up to date. The project will take serious momentum once the rotary encoders are installed. From than on I will be posting a lot of pictures.

Greetings,
Bart

[Mertensbart20]

Here are the first pictures where you can see the setup of the motors, the invertors (for the moment older ones) and the PLC.
Next up: connecting the encoders to the motors.
In the attachment you can also find the invoice of the motors

[dearn]

Nice to see some pictures! Good work, this is something different and more difficult then my JRK's haha!

[vicpopo]

Hi Bart,
Very good choice for the gearmotors !
Satisfaction garantee!
Best Regards

[Mertensbart20]

Hi dearn,

About the difficulty, you would think that. But our first project (see the first post) was actually more difficult. The principles are the same. But with this simulator there are fewer mechanical problems that can occur. And ofcourse it is far more accurate and faster.

The only major electric difference is the use of the frequency invertor compared to the H-bridge. But if we can configure the invertors a smart hobbyist should also be able do it.

[fooons]

Hi Bart

Very good job! Nice project and good pictures.


Regards fooons

[Riton]

Hello,
Interesting, I'll look at this project, it is interesting to see the performance engine with these characteristics.
engines.

The VM 140 is used for what?

[Mertensbart20]

The VM140 is used to transform the game-signals of X-sim (pitch, roll, ...) to analogue voltages from 0 to 10V. For us these type of voltages are interesting since we use a PLC to control the simulator.

I think most x-sim users would also use the VM140 to drive their simulator by connecting a motorcontroller to the VM140. But than you might get the problem that the VM140 is to slow for the calculations. In our project al the calculations are done with the PLC. So the VM140 only has to transform the signals.

[tronicgr]

The VM140 is used to transform the game-signals of X-sim (pitch, roll, ...) to analogue voltages from 0 to 10V. For us these type of voltages are interesting since we use a PLC to control the simulator.

I think most x-sim users would also use the VM140 to drive their simulator by connecting a motorcontroller to the VM140. But than you might get the problem that the VM140 is to slow for the calculations. In our project al the calculations are done with the PLC. So the VM140 only has to transform the signals.

Its not only about the calculations, the VM140 update rate is terrible. Just a quick example... The motor is running full speed to the end position of the actuator and is almost 100ms away from disaster. The plc gives feedback to the computer that it needs to slow down. The command takes 50-100ms to calculated in PLC and transmitted to the computer, the PC misses the refresh window of 100ms of the VM140 by some nsec... after 198ms the VM140 finally received the stop command. The end...

You see for a closed loop system like a motor control application you need as less delay in the feedback system as possible to be able to react fast in case of emergency or to just be able to handle and contain a fast moving system, like most motors are, into its limits.


Thanos

[Mertensbart20]

Hello Thanos,

I think you are misinterpreting the way I use the PLC. The PLC just receives signals from the VM140. The signals we use for rFactor are:

- roll
- roll acceleration
- lateral force
- pitch
- pitch acceleration
- longitudinal force

These signals are transformed to a 0-10V signal and take each a refresh time of 2ms (according to Velleman). So we have a refresh time of the Vellemankit of 12ms (I'm going to ignore the time needed for the PC to send commands via USB to the VM140). Than the PLC takes over and controls the simulator according to the values of these signals. And the cycle time of our PLC is 10ms max (according to Beckhoff = the supplier of the PLC). So we have a refreshrate of 22ms. The control of the motors is thus independent of the VM140 (except for the refreshrate).
Furthermore the PLC doesn't send signals back to the VM140. So we don't have that problem. I agree using just the VM140 is to slow. But combined with the PLC the VM140 doens't have to do calculations, just transforming signals.

You have ofcourse a lot more experience. Do you agree with the above? Our fist simulator (based on your design) seems to work fine using the same principle as described above.

Greetings,
Bart