When I worked in the CAD CAM industry our plotter had a repeatability error of no more than 3 microns over a movement of about 1/2 a meter. Using lead screws as well. Interesting video.
I think you and the rest of the replied commenters misunderstand what they demo's here. They are demoing the triCal system, not the robot? That's the 3 micrometres in a jig thing with the changing numbers.
@@zeitgeist909 Your an engineer? Just guessing based off your statements. Your playing a glass half full/half empty game. You weather your demo'ing a "fixture" that checks calibration of true positioning, or demo'ing a robot that can place an object in true oreintation with repittion, it does not matter. You still need both and are reliant on eachother. You dont design a fixture for the purpose or not checking the reliability robot. I myself come from the feild of fixing things like this robot, or the ones that have to use gauges in real life applications on real world robotics.
@@Hallettjs7957 It's more of a RU-vid snippet than a demo. I would assume that they probably published a lengthy paper on this since it was designed and made at a university affiliated engineering lab.
He's not playing any game. He correctly identified a comment which is off topic and flagged it as such. They are not both reliant on each other, they are independent of one another. This is to test any industrial robot. Also as an engineer, repeatability under load after 1 million loads will only be due to the wear on components, not an additive error. These robots have direct feedback, so any wear in mechanical parts does not affect the output more than one step of each of the output motors, or more than the lead/lag of the output and how it is programmed.
@@OleksandrStepanenko Expected so, I seem to see a repeatability of around 0.006 -0.007 mm or so, not bad. Absolutely within 0.01 mm that is within the range of normal for these.
@@OleksandrStepanenko wait this is CGI right? The colors and reflections look too perfect to be real... (Edit: well im dumb i missed the full title lol)
There seems to be a bit of a drift but that is still incredible precision. I imagine you could use a similar setup for a calibration feedback loop for prolonged operation even under load.
Yes, because measuring repeatability is not just reaching to a point from same direction all the time. Moreover, it needs to be loaded with its payload as well.
@Armin Tamzarian It looks good in video. But practically any robot cannot be delivered as a product if its repeatability is identified and certified using this approach.
Assume robot tool ball moved in only direction of Y axis dial gauge and abusively we can see deviation in Y direction dial gauge ,but due to spherical geometry of robot tool ball ,we will get deviation in direction of x and z dial gauges also .How you are consider/compensate it ?
I really wanna know what do you do professionally , ok I am sorry I didn't even notice I am commenting on every video at the same time but I just found out know about all this stuff but please answer a last question what do you do professional ?
That's cool, so that arm has a repeatable accuracy range of 0.003mm or 3 microns (3 measurements were .008, .006, .009) on that top micrometer sensor. I wonder how complex the algorithm is to keep that 3 micron range, I bet there's a lot of variables they had to account for. Must have been difficult.
Up to 0.1 of a mm is good. If there is a bit of force in the works these robot arms get the 'shakes.. very quickly you have 0.2mm repeatability.. For many applications still a well designed simple cartesian machine is better.
@@TheRainHarvester The overhand is huge on arms. Imagine it has a payload of 20 Kgs... or 60+Kgs.. Even the big arms get the shakes; The trajectory planner on the high end tries to ease that but there are limitations.
@@combatcorgiofficial Riiiiiiiight! IF you only knew. I can tell you one thing, though. If that is a standard axis resolution kuka robot, I'm guessing here, then one rotation of the resolver on any of its axis has a cool 1 million counts. Our machines have 4 million per revolution. The next generation equipment coming in here will have 12 million count per rev. So... Yea, KUKA ain't shit... I've watched them make over a milion car bodies back when I worked in a car mill and they do a fine work spot welding. They don't compare to any decent CNC machine, although the germans will make that claim "it's perfect and has great repeatability".
ah! your weakness is rough textures when its all smooth textures it looks so realistically fake but with with rough textures it just would look like any tech demo (i might be dumb cos i see the world differently and understand things less than most people do)
