|
Post by rchristensen on Jul 11, 2016 13:06:43 GMT
Hi, I am building a linear actuator based 6dof simulator. I am curious if anybody has any example of setting up the Thanos 6dof plugin in X-Sim (will eventually use BFF also) for a linear system instead of a control arm system. If this isn't possible, how is it recommended to interface a 6dof linear system with X-Sim? I have the hardware built, working on a few issues with the position sensors but the final step is software configuration. Thanks in advance!
|
|
|
Post by hexpod on Jul 11, 2016 19:53:11 GMT
Hi, I am building a linear actuator based 6dof simulator. I am curious if anybody has any example of setting up the Thanos 6dof plugin in X-Sim (will eventually use BFF also) for a linear system instead of a control arm system. If this isn't possible, how is it recommended to interface a 6dof linear system with X-Sim? I have the hardware built, working on a few issues with the position sensors but the final step is software configuration. Thanks in advance! Maybe I could go for an alpha version.
Give me your actuator stroke and base/platform diameter in mm?
|
|
|
Post by rchristensen on Jul 11, 2016 21:12:35 GMT
Stroke 12" (305mm) Base triangle side: 4 feet (1220mm) Platform triangle side 3 feet (915mm)
Thanks!
|
|
|
Post by tronicgr on Jul 12, 2016 15:55:48 GMT
You were asking some questions for the multi turn potentiometers on your email. If you haven't shorted that out, please post some details here to figure the electrical part of it.
Thanks Thanos
|
|
|
Post by rchristensen on Jul 12, 2016 16:02:26 GMT
Yes, I was just asking about modifying the min/max range of the distance sensors on a per actuator basis in the software. I am using a string potentiometer to detect the distance of my actuators. The potentiometer that came in the kit is 10 turns, but i'm only making it through about 4.5 turns with my 12 inch stroke. I know that I would lose precision by not being able to use the whole 0-5v range, and I'm looking into replacing the pots with 5 turns instead (even looking into hall effect for added durability but they are very expensive). I am open to other fairly inexpensive suggestions on measuring the position of a 12" linear actuator. My actuators are pneumatic so that rules out the methods used for motor controlled actuators.
|
|
|
Post by tronicgr on Jul 12, 2016 20:12:33 GMT
The amc1280usb has an option in the LCD menu called feedback limit I think, it's set to 10% by factory ( by me, lol ). You can decrease that to 1% and you can then use more range from your pots.
Thanks Thanos
|
|
|
Post by hexpod on Jul 12, 2016 22:08:15 GMT
Hi, I am building a linear actuator based 6dof simulator. I am curious if anybody has any example of setting up the Thanos 6dof plugin in X-Sim (will eventually use BFF also) for a linear system instead of a control arm system. If this isn't possible, how is it recommended to interface a 6dof linear system with X-Sim? I have the hardware built, working on a few issues with the position sensors but the final step is software configuration. Thanks in advance! Maybe I could go for an alpha version.
Give me your actuator stroke and base/platform diameter in mm?
tried to follow the procedure from this thread www.xsimulator.net/community/threads/6-dof-master-developing-corner.1698/page-2#post-17153 and I got the 3d model moving correctly by variation of the length of the rods.
Unfortunately the platform is not moving at all.
I have no idea why.
It might be the PWM servo calibration values have to be rewritten considering the variable change.
In the cranked design the "servo" (servo rotation) variable is clearly set for calibration. I tried to replace with "L2" (rod length).
Maybe Thanos could give a try...
cheers
|
|
|
Post by tronicgr on Jul 13, 2016 11:04:25 GMT
Hi Ignacy,
The amc1280usb will work with either linear or rotation data being sent to it (it really does not know the difference, just so takes it as target position ).
I think the issue lies to the fact that the motion data are linked to the rotation variables (crank) instead of the rods. Should you follow the motion data values in the code, you will see how they are transform to rotation values. Robin suggested to replace those with L2 but L2 is not get the motion data feed anymore if you remove the rotation calculations.
Thanks Thanos
|
|
|
Post by hexpod on Jul 13, 2016 23:02:49 GMT
Hi Ignacy, The amc1280usb will work with either linear or rotation data being sent to it (it really does not know the difference, just so takes it as target position ). I think the issue lies to the fact that the motion data are linked to the rotation variables (crank) instead of the rods. Should you follow the motion data values in the code, you will see how they are transform to rotation values. Robin suggested to replace those with L2 but L2 is not get the motion data feed anymore if you remove the rotation calculations. Thanks Thanos "L2 is not get the motion data feed anymore if you remove the rotation calculations."
I agree.
with this, the rods are nicely calculated:
L2 = sqrt( pow(A-PBT[0],2) + pow(B-PBT[1],2) + pow(C-PBT[2],2) );
Logically, one could imagine that exchanging the rod length variable with crank rotation variable would move the platform.
But it is not so obvious where exactly.
I am stuck.
Keep me posted if you find it...
|
|