Dual heads on a 3-axis machine, without interfering?

[some bloke]

TL;DR

I'm looking to design something which requires 2 heads to move simultaneously in 3 axes, without their running-gear interfering or colliding with one another. Is it possible?

As the description says, I want to make something which moves 2 heads on a 3-axis basis. The heads will ideally rise from the flat top of whatever contraption will be needed - it's not an option to have one rise from the bottom and the other descend from above.

Aside from the obvious collision if the 2 heads attempt to occupy the same space, I want to make it so that the heads can move past each other, move in circles around each other - all without their running gear (worm drives & actuators etc) from colliding.

My main thoughts at present are:

1: The 2 heads can rotate about the centre, extending inwards from the circumference, and the whole bed can also rotate, which would allow them to move past each other - if it looks like they will collide, the bed rotates to avoid it, and the heads are held in the same location.

2: One head is branching off of the first -so one is a traditional x-y-z, and the other is and x-y-z based off the first head, rather than the base of the machine, and as such would not interact with the running gear for the first.

Has anyone seen anything which does this? Am I missing a trick? I'm concerned it's like the old puzzle of getting gas, water and electric from 3 factories to 3 houses without crossing the pipes - without going over the page and poking back through, it can't be done (and I think wormholes are a little out of my skillset so I'll definitely avoid that approach!)

 

[jrmichler]

I saw a video once of two robots alternately reaching into the same space. They handled the collision problem by programming paths with clearance between the two robots. Modern servo systems all have position error functions that shut down the entire system if anyone servomotor is away from the programmed position by more than a predetermined amount. That amount can be less than 0.001 inch. This method works if both heads are traveling a small number of different paths, where you can step through each path and manually make sure they never collide.

If the heads need to follow a large number of different paths, a different approach is used. The surface of each head is mathematically defined, and an algorithm looks for interference. This is done in high end FEA packages, and is computationally expensive. I'm not familiar with the specific algorithms, but do know that they are not simple.

The best approach is if one of your main thoughts can be developed into a system that cannot collide.

 

[Baluncore]

I think we all know that two heads are better than one, but why do you really need two?
Do you need twice the speed or twice the tools?

The turntable is one possible solution, where one head is fixed with the turntable moving below it. The second head is free to follow the table along the circumference but not collide with the fixed head. Also consider an axis like a lathe, but indexed in rotary position, with one head on each side of the axis.

The general situation may be impossible. To see the special case solution we need to know what tooling you intend to use, and what shape and dimension the object is that is being machined.

 

[anorlunda]

If you find it impossible to pre-plan collision free trajectories, you could add proximity sensors to the heads to make active dynamic collision avoidance.

 

[Ranger Mike]

I'm looking to design something which requires 2 heads to move simultaneously in 3 axes, without their running-gear interfering or colliding with one another. Is it possible?

Was done in early 1980’s by many CMM manufacturers



Case in point dual arm body checkers to automatically measure cars – see photo

As the description says, I want to make something which moves 2 heads on a 3-axis basis. The heads will ideally rise from the flat top of whatever contraption will be needed - it's not an option to have one rise from the bottom and the other descend from above.

Computer software let's you do both movement options
Aside from the obvious collision if the 2 heads attempt to occupy the same space, I want to make it so that the heads can move past each other, move in circles around each other - all without their running gear (worm drives & actuators etc) from colliding.



Today’s software let's you virtual run the program so you can detect a crash on line before actual hardware is moving
My main thoughts at present are:

1: The 2 heads can rotate about the centre, extending inwards from the circumference, and the whole bed can also rotate, which would allow them to move past each other - if it looks like they will collide, the bed rotates to avoid it, and the heads are held in the same location.

2: One head is branching off of the first -so one is a traditional x-y-z, and the other is and x-y-z based off the first head, rather than the base of the machine, and as such would not interact with the running gear for the first.

Has anyone seen anything which does this? Am I missing a trick? I'm concerned it's like the old puzzle of getting gas, water and electric from 3 factories to 3 houses without crossing the pipes - without going over the page and poking back through, it can't be done (and I think wormholes are a little out of my skillset so I'll definitely avoid that approach!)
Typically these dual horizontal arm Coordinate measurement machines share a common x,y,z table and over lap by 100 mm as they both calibrate on one common datum sphere to identify their positions in the measurement volume. Each HA arm moves in xyz and the heads rotate in A and B axis to you have a 10 axis movement

 

[some bloke]

Thankyou all for the replies.

I think that based on the information here, my best bet will be the dual robotic arms as per the picture above, but on a rotating bed to allow them to rotate around themselves.

Cheers guys!

 

[Ranger Mike]

i sell these in USA