Right I understand that I can discover the failure at some point later on, but the question I'm asking is *why is roboDK failing to filter each time?* Is there a readout of which of these reasons (close to axis 5 at 0 deg, or arm fully extended, etc) is causing each failure?
In a single program I might have 40 lines that fail, spread out along the program
e.g.
Line 25, Line 456, and Line 3770 failed to filter.
What was the reason for each of those lines failing?
At the moment I'm forced to guess at a solution by adjusting the approach angle, the retract height, the cutting depth, or the transition or any of 100 possible things. Then I have to regenerate, export, and drop it into robodk to find out if it will filter or not.
It is currently not possible to have more information when you filter entire robot programs. However, by looking at the joints you could see if the robot is close to the joint limits or a singularity.
I have another problem with this same cell, we rotated the spindle vector by -60deg around nulltool Z to try get around the frequent "failed to filter" problems.
If we run a small sample operation approaching from the X-1,Y1 direction, the program fails to filter our approach and retract points. It filters the machining operation fine. The joints are at least 30deg from their limits, and the limits are set appropriately. There aren't any fixtures in the cell.
Can you filter this example operation (we just drag the .src file into the cell and click "filter only") and tell me anything more about why it isn't working, or what I can do? Do I need to rotate the spindle back the other way (+120 around nulltool Z), maybe?
Otherwise, if I get the script working, what would happen with the "fail to filter" errors? Would it just skip them, and not adjust the value? Even that could be a workable outcome, it always seems to be the transitions, and not the actual operations. Does RoboDK struggle with large reorientations?
When you try to import the KUKA SRC file instead of filtering it you can see that the first linear movement is not reachable. Something may be wrong with the setup, for example, the way RoboDK understands the tool, the coordinate system or the position of the turntable.
Thanks for the response, the problem I have is that the first position is actually reachable. I can run the program without filtering with no issue, and the failed point isn't close to the joint limits (about 30 deg away from the limit on the closest joint).
We have the same tool vector defined in the robot, cam, and robodk. I believe the table matches up fine, as when it does filter, it matches up correctly. We've had to shorten our approach/retract distances to less than 50mm to get programs to filter. Is there a maximum distance setting somewhere?
Any other hints about why it wouldn't be filtering, and only failing on the approach/retract movements?
If you program your robots to be close to robot singularities or axis limits you may see issues like the one you described (unable to filter certain points). I would recommend you to use APT files instead.
Another option is to increase the sinularity tolerances to allow the robot to move closer to singularities. You can edit the singularity tolerances by following these steps:
