How to avoid overstressing the rig without 6DOF math?

From information available on this forum it is quite evident that 6DOF simutators can well work with the current version of Simtools without proper 6DOF math. I understand that this is achieved by finding close-enough linear approximation of non-linear actuator commands actually required to correctly position the moving platform. Since that woudl be still an approximation, the motors will not be all positioned correctly and will be somewhat pulling and pushing against each other, putting additional stress to the rig.

Is there any correct procedure to find correct axis percentages without ripping the rig apart? Or should I just make sure that all my mecanical elements are strong enough to simply absorb extra forces?

I'm not speaking from the experience of doing it because I don't have a 6dof, but generally speaking it is being achieved by keeping movements smaller than would otherwise be possible. So I suggest you would start with very small axis percentages and then work your way up until you are satisfied with the motion or until you start noticing binding issues.

A properly setup Stewart Platform does not work that way, no position of an actuator should cause it to bind. If your rig is binding, using 6DOF math would not help. It becomes a mechanical, not a software limitation

I do not think this is correct. For each possible position of the moving platform there would be exactly one or two correct positions of a motor lever. Any other position of the motor lever would cause that lever to compete with and work against other five levers. Without calculating the exact angular positions of the levers based on 6DOF math it would not be possible to ensure that every one of them is positioned correctly.

Thanks. So it all is about going slow and light.

Just because a lever/actuator is not in the exact calculated position does not mean it will tear the rig apart - like I said before, if there is binding it is a mechanical issue, not a software issue. I have a working 6DOF which uses both BFF (6DOF math) and Simtools. In testing I had every actuator at min/max to ensure that I had no possible mechanical binding issues. There is no combination of actuator lengths that will make my rig bind - a property built Stewart Platform just works that way. Build a model if you don't believe me. Or better still, take the advice of members who would not know....good luck - me out.

My explanation is that the rig 1) has enough slack to absorb geometrical inconsistencies and 2) is strong enough to hold against stresses applied to it by uncoordinated lever movement.
I am quite confident that my rig will also hold well against these things, but in ideal world it would be good not to add unnecessary forces to the ones that exist naturally and unavoidably. Along with stressing the mechanical components, having too much of an error in motor positioning commands decreases efficiency and responsiveness of the rig.

After thinking it over once again, I found an error in my logic: somehow I assumed that the moving platform is "frozen" in the target position and unless all the levers are exactly where they should be -- they will be taking stress. In reality, of course, it is the other way round and incorrect lever position will just put the platform in the incorrect pose.

@SeatTime is correct, there is no competition between the levers and error in lever position just leads to error in orientation of the platform. Therefore, tuning the axes is only about minimizing the discrepancy between game output and actual state of the simulator, no additional mechanical stresses are involved.

That’s why I let it be known I was not speaking from the aspect of actually doing it myself so that could be taken into consideration.

What I’ve observed is this though @SeatTime. Your 6dof rig has more movement than most others I’ve seen that use Simtools. There are two things different about your rig than most other rigs here though in that you have linear actuators and are also using BFF software. So my assumption was the others have less movement because they aren’t using BFF software and maybe more so that they are using non-linear actuators.

So I think what I said was still somewhat true. But they use less movement in order to keep the motion from getting too sloppy with non-linear actuators I assume without the 6dof math.

But I also guess you are speaking from the point of view that as long as non-linear actuators all are moving together as they should, then there should be no binding on a Stewart platform? I didn’t run a test model to check, but it seems counterintuitive to think that you would be able to move one actuator at a time on a 6dof and not have binding somewhere.

Edit: After reading your post again carefully I see that you tested your rig one actuator at a time. I suppose it may put a little more stress on the actuator but now I see how it will still move in some way or the other if the motor is strong enough.

As long as your arms are far enough away from your levers they should not bind. If they do bind then you are losing the maximum movement that your rig can make.

As @SeatTime said it's a mechanical limitation that needs to be fixed you should have a full 90 degrees of movement on each arm with no binding if you have then you should look at moving the arm connection further away from the lever, shorten the length of the lever or limit the movement in simtools.

Thank you, @SilentChill.

I am not sure I understand where this angle should apply. Can you clarify, please? That is not about the lever movement range which, I believe, should be 180 degrees, right? Is it about the angle between the arm and the lever on the motor shaft?

Sorry 90 degrees each way from centre