7 Replies Latest reply on May 22, 2009 7:01 AM by 1-IRI32N

    freeform questions for Mark B

    Neil Larsen
      I must say I quite like the additions to surfacing from 05 to 09.
      Well done SW folks...its almost a pleasure to play with at 3.8ghz... however I have a few questions about freeform things today I hope you might answer..(hey come on its xmas)

      How can I specify a precise location for a curve and point on a surface?
      I cant see how to do this ATM..thinking back this is a little like the situation when splines really couldnt be specified properly re handles etc.- they just sort of floated out there in space somewhere in the general direction of up/down and around
      Can we not have a realtime measurement along say an edge and then after placing, along the established curve to aid triad placement? maybe including say snap at x mm increments or better yet a numeric input with it?..
      what I am thinking about is possibly similar to inserting a reference geometry point on a curve - distribute evenly etc.
      possibly it might incorporate something like an instant 3d ruler handle or spline end like widgets...dunno..

      I made a freeform experiment that used a previously established spline with a few points with the intent to drive the freeform...
      turns out that if you go back and edit that spline the freeform points arent locked to it requiring manual updating of the triads although they will once again snap to the points...
      Is it not technically possible to have a coincident relation rather than a snap? (or as another option) ,or is that a bug?
      Also I found triads would not snap to any aforementioned reference geometry points if made...could these be included in the snap options?
      The same scenario happens with snapping triad points say to and underlying box corner - ie works but wont make a coincident relation.
      I think it would be cool to have an adaptive surface in this fashion - for instance using multibodies one body could be like a 3d sketeton sketch for a surface freeform deform

      hope you know what I am saying there..

      Can we have a partial reset for triads?
      At present undo undoes the whole 'errant' manipulation
      What if you only want to undo the last triad move and not the last 5 which you were careful about and happy with?
      How about therefore undo steps and undo all...

      Seasoned greetings to SW staff

      Edit: attach are some concept pics of a realtime running distance box to aid curve and point placement -just one approach that came to mind you understand
      In this example you pick an edge and slide the cursor along until the desired point
      - the box gives the running position and the overall length
      Having placed the curve the box then follows the point placement in similar fashion
        • freeform questions for Mark B

          The only percise mechanism is the ablity to snap a freeform point to an external point as you've discovered. Otherwise, there is no way to percisely place control curves or points. If you can use symmetry on the face, then the control curve will snap to the mid-plane of the symmetry.

          1) The intent of Freeform was to visually (and thru curvature analysis) "tweak" exsisting surfaces and solid faces. I find it most valuable for tweaking boundary and fill surfaces were as you can not achive quite the desired intent in those original features and/or rotate the flow of the surface by rotating the control grid of the freeform face.

          2) I don't quite understand your comment about "adaptive surface... using multibodies

          3) please submit an enhancement request for partial reset of triads. You do know that the triads behave like general triads in that you can align them to selections and that the UNDO in Freeform will incrementally undo triad's individual moves.
            • freeform questions for Mark B
              Matt Lombard
              Along the same vein, I think SW has a great "opportunity" for improvement with connectors in lofts and boundary. Same deal as the curves and points in freeform. Would be nice to be able to parametrically locate them either with respect to other geometry via sketch relations or dimensions or through a u-v parameter, percentage like locator (.5 puts it in the middle of the curve). That part is probably happening behind the scenes, but it would be nice to make it explicit.

              As you said, visual editing is really the point of the freeform, but precision is really the point of a CAD program. It wouldn't hurt the visual editing to enable the precision. Maybe you could make the "override dims on drag" option more obvious for people who forget about it.
              • freeform questions for Mark B
                Neil Larsen
                hmmm well in my typical off beat way I guess I was wanting to use this in a different fashion than it was perhaps conceived in the labs..

                I was using it to whip up a creamy sorta shape from a throw away general purpose bit of surface I had laying around rather than just tweak something a wee bit that I didnt quite make the grade from the usual construction.
                I was intending to develop my work radiating away from the sculpted freeform rather than refining a bit of the whole finished item...if that is a novel approach or highly 'unusual' or inconceivable to regular SW coders minds I apologise..there is always one...

                In the same contrary way I was wanting to drive the freeform from other geometry.
                I suppose you might think of this as similar to a loft between 3d splines.
                After the fact in that case you can go back and grab/edit the spline points and the lofted surface redefines/updates.
                If the freeform 'snap' points were 'coincident relations' instead any change to the geometry that locates those triad points would update the freeform surface.
                Presently it just seems to borrow the 3d location to use as a once off..

                Considering multibodies - think of a rectangular block.
                If a freeform surface were constructed in proximity to it it would be possible to 'attach'(snap) some triad points to the vertices of the block.
                In this manner a change to the block would deform the surface
                Sorta like a balloon with glue inside or reverse shrink wrap if you like.
                The freeform surface would adapt to the driving geometry.
                Perhaps think of plastic cover panels fitting over a quadbike or such..although you would perhaps offset to give clearance (or the snap points would already allow for the offset)

                Sorry I am not so good at explaining myself or writing - Matt probably says these things better in fewer words..being a clever chap that he is with technical terms..

                And...Matt is correct loft connectors suffer from the same indeterminant imprecise status.
                Both of these areas are a bit airy fairy for my liking at the moment.
                I really would like to have a way of specifying these placements with 'precision'
                If folks dont want the precision they can just ignore the info and pull them around as they are inspired to do but I feel that this is something that would benefit from being definable or at least have a number of options for
                You may remember how people struggled to define a spline for machining/measuring purposes at one time because you couldnt be precise about angle of handle, length etc...
                A test for this completeness might be say: could I describe how to replicate this surface to someone over the phone?
                and the answer is : presently no - I could describe the the boundary edges accurately but I couldnt describe where exactly to put the curves and points...(or loft connectors either..)

                To me although this tool offers fluidity in visual interaction (and it is kinda neat to use ) is not a lot ulimately a lot different from say using a surface fill with patch boundaries and a constraint curve - that is although it is very plastic/elastic vs hard and resolved it does just define a desired surface and should have the option to being contrained and located precisely - or at least in a more deliberate way if desired than exists presently.

                aghh... I'll give up talking I'm even confusing myself now and repeating things

                Loft connectors definitely need a revisit - its too easy to end up with a tangle of connectors in no particular position at all other than something about... there.
                Perhaps it is only Matt and I who think this way but often with loft I have wished for something that can be quantified/fixed where I want it even if it doesnt seem as though there is any great importance to that to observers.
                I would rather just have it at 5mm or 1/2 way or 3 evenly spaced or something

                regards etc

                BTW because boundary code will eventually replace the old lofts etc and whatever is built on top of that -indent?deform?? it might just be an area that some thought could be seriously given to
                as Matt so ably puts it a 'great opportunity for improvement' awaits..

                and no I didnt appreciate the triad behaved as other general triads re alignment - thanks

                  • freeform questions for Mark B
                    Neil Larsen
                    BTW Mark here is your Xmas present -

                    I keep sticking this "kite" idea in your faces from time to time in the hope you might pick it up and wear it (I finally stopped giving people ties last year..brought 24 wide polkadot ones in a garage sale in '83)

                    Here I made a rough mockup for you in 09 using some ordinary sketch tools...
                    The result is fairly flexible but a little clumsy improvised like this - you can move the arms/strings (yellow) around and drag the rectanges (dk blue) in 3d position and the loft follows it around and re-orientates.
                    I think I messed it up slightly playing with it but you can probably get the idea
                    Done properly there would just be the arm with base point and a plane with a triad point on the ends

                    The idea is to be able to place a loft profile in space hanging off other geometry with the min effort -be it for wiring/piping or surfaces in proximity as here.
                    Of course you could arrange things so that there was a lock on the distance for clearance purposes and just leave it gimballed..
                    The kite would work a bit like placing a balloon in a drawing but in 3d space with a triad.
                    Another analogy instead of a kite would be a car wing mirror ( if you remember those..)

                    So- insert a 'kite' feature ,attach/snap base end to existing geometry, drag out the triad and rotate/move the sketch plane to the desired orientation, add sketch,.. revisit triad and sketch as required for correctly shaped and positioned loft

                    In use the kite would retain the distance and angle relationship at its base point end so that if that end moved the whole kite would move.
                    Resizing the block would update the loft.
                    seems intuitive and useful to me..but then I do tend to have an unusual point of view


                    (sorry couldnt find any gift paper that hadnt been torn or crumpled at this particular time...I hope a zipped folder is acceptable - If this happens to be the same present I gave you last year you might be pleased to know I have only 2 left and they are past breeding age )


                    edit: attach is a better pic of what I think it should be like borrowing some visuals from live plane
                • freeform questions for Mark B
                  Hi there,
                  I'm new to Solidworks & have a freeform question. Mark Biasotti's video on creating a peapod hull form has got me puzzled on one aspect. After creating the freeform surface and placing the curves on the surface relative to the station planes I cannot place points/triads on the keel end of the curves to look or perform anything like that which is shown on the video. I'm using SW2009

                  Roger Dahlberg
                    • freeform questions for Mark B

                      the workflow is as follows:

                      1) Pick face or surface
                      2) select desired boundary condition for the boundary edges (i.e. contact, tangent, curvature cont. or moveable.)
                      3) Create control curve/s
                      4) Create control points on those curves.
                      5) Pull on point/s as desire.

                      step 3 is not required if step 2 is "moveable"
                      Step 4 is not required if step 2 is "contact" - then you can control the surface direction and tangent magnitude at the boundary.

                      Roger, the Property Manager for FF is somewhat laid out from top to bottom in the sequence that you would do these steps.

                      Hope this helps.