Search

[Xylotica]

Hi !

Since I've been using the "Axis move" component to unwind my axii in the middle of a job, I'm getting lots of kinematics suddenly stopping because this or that axis has exceeded it's max speed.
I wish that KUKA|prc warned me upfront so that I could find other axis values or geometric conditions that would avoid this situation.

I can use the axis graphs to "guess" where this is likely to happen by looking at the curve slopes, but the plugin could do an even better job using the derivative or the angular graph to get the angular speed, and warn the user graphically ; perhaps using an other color than red on the relevant axis.

Cheers,

[Johannes @ Robots in Architecture]

Hello,

The easiest thing to implement that would be to check every position against a change-in-degree-per-second threshold and color it accordingly. Do you think that would do the trick for you?

Best,
Johannes

[Xylotica]

Hello Johannes,

Here I am 6 years later, with the same issue.
I wish I had replied a big " YES ! " to your proposal.

I recently had a milling stop because of high angular velocities, and I couldn't see it coming.
In this case, it happened because I was "close" to the in-line wrist singularity.
This was not detected by KUKA|prc as a singularity (which would have triggered a warning), but still caused a brutal acceleration of axis 4 and 6 which prompted a safety stop.

I think your suggestion is perfect. It would be a very useful addition to the simulation.

[Johannes @ Robots in Architecture]

Hello,

This has been implemented like that for a while, seems like it did not meet the threshold in your case.

There can be multiple reasons for that, for example just a very slight difference between the physical and digital environment can have significant impacts on singularities, i.e. the difference between high and infinite/too high speed.
Or as we simplify the robot models per reach, a higher-payload robot may not reach the speed of a lower payload robot.
However, as the speed calculations are kind of rough anyway as the acceleration ramps of the motors etc. are not known, it's always an approximation.

So in summary, the easiest approach may be to allow the user to change the threshold when the notification triggers. Can you send me your file via mail? I'm curious how much change it would need.

Thanks!
Johannes

[Xylotica]

Hi Johannes,

Here's the exact moment when it happened, just for the record : https://photos.app.goo.gl/Yr8S6bGs4vQcN4jx7

[Xylotica]

OK, I made a custom robot with the actual angular speeds of my specific robot.
I added a slider to reduce proportionnaly all these values, to see at what point get a warning in the analysis at some point, using the definition that previously caused an emergency stop for A4/A6 exceeding their speed limits.
The result was a bit surprising : when I went from 100 % to 1% angular speeds, it completely changed the simulation.
These values actually influence the simulated rotation speeds, which means that the speed of the TCP is much slower than the one requested, since the overal duration time is much longer.
In these conditions, I don't see how setting the max angular speeds differently can influence any warning in the analysis graph.

I'm perplexed.

[Johannes @ Robots in Architecture]

Hello,
I just had a look and it seems a bit strange indeed, I'll have to take a closer look at that tomorrow!
Thanks,
Johannes