*MonkeyJump
Message 1 of 3 (30 Views)

# REPOST: fitted curves

30 Views, 2 Replies
03-15-2000 06:16 AM
If you try to draw a fitted curve through 3 points which you know are on
a parabola, QC7 doesn't draw a parabola at all. Normally, if you define
"n" points you would expect the fitted curve to be a polynomial of the
(n-1)th degree...at least that is the mathematical definition. However,
QC7 treats fitted curves differently as it lets you jump anywhere on the
screen so that you can even cross the curve unto itself, this- in
disagreement with the analytical model. I am just wondering what
mathematical formulation these curves have internally. There also seems
to be no way of snapping to them, except at the endpoints. Sorry for the
repost but this question has remained yet unanswered. :-)
*Becker, Don
Message 2 of 3 (30 Views)

# Re: REPOST: fitted curves

03-17-2000 08:13 AM in reply to: *MonkeyJump
so here goes.
The command you're talking about you have to select 3 points, right?
Well, I would guess the three points on your parabola are the (1) the vertex
(is that what you call it?) then two points, one resting on either side,
opposite each other.
First the Quickcad command is trying to make a circle. Or part of a circle
(an arc)
so here's what I'd do to get a parabola( note: the author of this post would
like to tell all that he's not sure this will work, it's just a thought, but
a phrase to all that would critisize- post something then!).
First, get that vertex(?) point.
Then you need 2 points on ONE SIDE of the parabola.
Draw the fitted curve.
Then mirror the half parabola you just drew to get the full monty, er,
parabola.

About the snapping thing, try highlighting the item in question,
then right-click, go to convert- take it from there, might work, might not,
you are trying the intersect command, aren't you moneyjump?

Lemme know if all that helps,
Don Becker
MonkeyJump wrote in message
news:38CF9B36.9716A411@videotron.ca...
> If you try to draw a fitted curve through 3 points which you know are on
> a parabola, QC7 doesn't draw a parabola at all. Normally, if you define
> "n" points you would expect the fitted curve to be a polynomial of the
> (n-1)th degree...at least that is the mathematical definition. However,
> QC7 treats fitted curves differently as it lets you jump anywhere on the
> screen so that you can even cross the curve unto itself, this- in
> disagreement with the analytical model. I am just wondering what
> mathematical formulation these curves have internally. There also seems
> to be no way of snapping to them, except at the endpoints. Sorry for the
> repost but this question has remained yet unanswered. :-)
*Support, Autodesk
Message 3 of 3 (30 Views)

# Re: REPOST: fitted curves

05-08-2001 08:03 AM in reply to: *MonkeyJump
control points when creating a splined curve (AutoCAD) or a fitted curve
curve formulae. One popular program that does allow formula input to
generate a curve is Mathematica, that will then allow you to calculate
the area under the curve, for example.

To create the parabola from existing points in QuickCAD use the
Draw/Curve/Fitted with grid snap on (if the existing points coincide
with grid points). Use Nearest Snap if the points are off-grid.

As far as internal mathematical formulation, here is an example of a
parabola created in AutoCAD and LISTed to view the internal geometry
positions of the control points: (Note: QuickCAD does not display this
level of detail for curves properties.)

Command: li
LIST 1 found

SPLINE Layer: "0"
Space: Model space
Color: 7 (white) Linetype: "BYLAYER"
Handle = 93
Length: 15.60
Order: 4
Properties: Planar, Non-Rational, Non-Periodic
Parametric Range: Start 0.00
End 15.23
Number of control points: 5
Control Points: X = 1.00 , Y = 3.00 , Z =
0.00
X = 2.00 , Y = 6.50 , Z =
0.00
X = 4.00 , Y = 13.50 , Z =
0.00
X = 6.00 , Y = 6.50 , Z =
0.00
X = 7.00 , Y = 3.00 , Z =
0.00
Number of fit points: 3
User Data: Fit Points
X = 1.00 , Y = 3.00 , Z =
0.00
X = 4.00 , Y = 10.00 , Z =
0.00
X = 7.00 , Y = 3.00 , Z =
0.00
Fit point tolerance: 1.00E-10

MonkeyJump wrote:

> If you try to draw a fitted curve through 3 points which you know are
> on
> a parabola, QC7 doesn't draw a parabola at all. Normally, if you
> define
> "n" points you would expect the fitted curve to be a polynomial of the
>
> (n-1)th degree...at least that is the mathematical definition.
> However,
> QC7 treats fitted curves differently as it lets you jump anywhere on
> the
> screen so that you can even cross the curve unto itself, this- in
> disagreement with the analytical model. I am just wondering what
> mathematical formulation these curves have internally. There also
> seems
> to be no way of snapping to them, except at the endpoints. Sorry for
> the
> repost but this question has remained yet unanswered. :-)

--
Bob Felton
Autodesk Product Support, USA
WW Support & Services, Autodesk
Discussion Q&A: http://www.autodesk.com/discussion
Post to the Community