Field Line Spin Force not working in vertical plane

Hi!

I am trying to generate field lines on a vertical plane (wall) using spin force. It gives an error '1. Field line could not be solved'. It generates field lines but the lines are not in the same plane as the bounds.

Thank you for he help!

Attachments:

Field Lines Vertical Error.gh, 7 KB

Replies to This Discussion

Permalink Reply by David Rutten on July 26, 2017 at 2:24am

Seems to be two different problems. You can get rid of the error by using a different random seed value for your 100 random points. You're using the same seed, meaning the first 3 points out of your 100 are in exactly the same spot as your three spin force centres. This is somehow causing the error. Change the seed to 2 and the error goes away.

However all the resulting lines only contain a single segment, rather than the requested 1000. I'm looking into why this would be.

Permalink Reply by David Rutten on July 26, 2017 at 2:32am

Whatever the problem is, it's something to do with the vertical orientation. Until I fix the bug you can switch to horizontal and it'll work.

Attachments:

FieldLinesVerticalError.gh, 15 KB

Permalink Reply by Tejas Chavan on July 26, 2017 at 2:43am

Thank you David! Looking forward to the fix! Will be using the horizontal orientation for now..

Permalink Reply by David Rutten on July 26, 2017 at 6:32am

Ouch, the force evaluator for spin fields is just a total mess. Only works if the field plane z axis is exactly vertical, otherwise there's a lateral component. Which explains both why the single segment field lines were all poking out the side of the plane and why every field line only generated a single segment (because it broke out of the bounding box right away).

I'll fix it, but it will remain broken forever in Rhino5. Next week's Rhino WIP release should be fine though.

Permalink Reply by Tejas Chavan on July 26, 2017 at 7:24am

Thank you David!!

Permalink Reply by Tejas Chavan on July 26, 2017 at 9:11pm

I had another doubt David..Would it make sense to add a bounds parameter to the FVector component? Maybe then we could use multiple FVectors on a set of points such that the resulting fields have a gradient of changing directions..

Permalink Reply by Tejas Chavan on July 27, 2017 at 2:47am

Oh found a different way of doing the same..Used 2 attractor points with zero decay instead!

RSS