G-levels on Robotic arms...

[Peru]

Can anyone give me an idea what g-levels a robotic arm carrying a 125 lb laser would reasonably expect? Laser cutting speeds are relatively slow perhaps up to 250mm/s, and probably g's less than .1 but when non-operational, any ideas?

Thanks

Mel

 

[tygerdawg]

It's not clear what it is that you are requesting. "g-levels" ?

That would imply some sort of acceleration applied to your robot + laser, but resolved in units of 32.2 ft/sec/sec or 9.81m/s/s. Doesn't make sense: an industrial robot carrying an auxiliary load doesn't get subjected to external accelerations unless there's an earthquake or something. In that case, you're in deep soup anyway.

Industrial robots typically allow speeds & accelerations, either total or defined by individual axes, in the range of 0-100% of a default value. The default value is set by the manufacturer to assure that accels aren't exceeded and damage occurs. If auxiliary loads are placed on the arm, then these speeds/accels are reduced. Manufacturer's manual usually have some sort of chart or table providing this info to the user.

So...what is it that you are trying to determine? And why?

TygerDawg
Blue Technik LLC
Virtuoso Robotics Engineering

http://www.bluetechnik.com

 

[Peru]

G-level is a relative acceleration value compared to 1 G (9.81 m/s/s). So .5 g level is 4.9 m/s/s.

In my application, the laser is the auxiliary load and it would experience whatever speeds/g level's applied to it by the robotic arm. You refer to a default value by manufactures, do you have suggestions for manufacturers that I can contact, or default values that I can use.

I'm working on a specification for vibration testing of the laser but I need to know what the probable field acceleration levels are to define the test. Mil-Std 810 e.g. provides detailed info on this but this is not a military application.

Thanks

 

[tygerdawg]

The people you should be speaking with is the Applications Engineering Department of whichever robot company you are investigating. But...

Your question is much too broad to provide a definitive answer. The arm link carrying the auxiliary load will move differently for different arm positions and robot configurations. To get a max accel estimate of that link, you would have to assume the velocity / accel profile is at max (whatever that is, IF you could determine it). Then determine a 6-axis configuration which would move that particular link the most when motion starts, thus generating the max accel effect on the aux load. And then that doesn't even consider motion in the gravity field, and those effects.

I don't see any reliable method to estimate this with any accuracy. Unless you spend some money and get the Apps Eng Dept to do a motion study for you. Otherwise you can get from the manufacturer the max angular velocity spec of the loaded link axis, and apply a guesstimated time-to-accel time of....0.25 sec? 0.1 sec? 0.05 sec? Then do (DeltaV)/(DeltaT) for max accel.

TygerDawg
Blue Technik LLC
Virtuoso Robotics Engineering

http://www.bluetechnik.com

 

[FeX32]

There is a way. There are some new methods in optimizing the performance of robot mechanics that can be used inversely.
Essentially

do a motion study

is the answer.

If you have the inverse or forward dynamic model of the robot. I can describe a method (high complexity of you are not familiar with advanced dynamics ect. ) Or recommend a research paper.
But, I would suggest not going this route and just getting an estimation from the manufacturer as mentioned above.


Fe

 

[Peru]

Thanks for the info...looks like I found a local outfit that's willing to let me do a motion study on their equipment. Now, I'll get some specific field data...

Thanks for the feedback.

Mel

 

[FeX32]

Np.
You are going to conduct the motion study? Sounds like fun.



Fe

 

[BrianPetersen]

As an aside, I speak from experience in saying that the accelerations involved at the end-effector are pretty darn high when you smash the end-effector into whatever it is that you are trying to pick up, but which happens to not be quite exactly in the place where you think it is.

Just sayin' ... It would be a shame to go to the trouble of splitting hairs with a motion study in the interest of making everything really light and spindly, only to smash it all to smithereens the first time someone presses a wrong button or leaves a foreign object (i.e. a hammer or some other big heavy tool) on the fixture ...

 

[Peru]

On the contrary, instead of making it really light and spindly I'm testing it to see how rugged it is...or needs to be without turning it into a panzer tank. Also, I'm not trying to pick stuff up but shoot it with a laser beam (for industrial purposes only). So assuming everything is supposed to be where it is, a motion study might be quite helpful and interesting to boot...

 

[MikeHalloran]

One thing you might want to measure in your motion study is 'ringing'.

We had our laser cutting vendor try to mark 12 gage stainless sheet blanks with the same laser that cut them. They dialed down the power and dialed up the speed, but the laser left a hole wherever a character had a corner, because whatever elements were moving would wobble around a bit when asked to change direction abruptly. So they had to cut the parts on one machine, and mark them on a machine with much less laser power.

I was never clear about whether the ringing occurred because the moving parts were relatively massive, or if the servo system was out of adjustment, or if some parts of the machinery were nearing end of life.




Mike Halloran
Pembroke Pines, FL, USA

 

[FeX32]

but the laser left a hole wherever a character had a corner, because whatever elements were moving would wobble around a bit when asked to change direction abruptly

This sounds like it may be a lash induced resonance of the robot. The configuration of the robot could be changed to an end effector pose that increases the stiffness during motion to observe is the "wobble" diminishes or not.
At least, this would be first attempt at what you described.


Fe