Mini-challenge
-
@pbacot said:
Well if you had a true arc or circle you could find your exact intersection points. But there are no true circles in SU.
My 'TrueTangent' tools have various options, in this case you could place a cpoint exactly at the 'true-intersection' of two SUpArcs...
-
@unknownuser said:
it will never be accurate (second screen shot). The cursor is "sliding along" the guide and does no give an accurate inference.
That will say that any rotation can be made precisely?
You can make zoom during the process!
And re look my last image, there is green point that indicate that the snap is done!
And a red one will be drawn at the end of the last snaping on the group!What do you want more accurate ?
I have verified! Seems you right!
That is very ennoyous and want to say that any constructions made with align / Roation are false!Another headache to resolve!
-
Jeff, I've had to draw diagonals like this a number of times and not found an elegant solution. It's always been via brute force. I forgot about TIG's True Tangents.
-
@dave r said:
Jeff, I've had to draw diagonals like this a number of times and not found an elegant solution. It's always been via brute force. I forgot about TIG's True Tangents.
Pilou's method was quite easy and simple. And also accurate!
-
@thomthom said:
@dave r said:
Jeff, I've had to draw diagonals like this a number of times and not found an elegant solution. It's always been via brute force. I forgot about TIG's True Tangents.
Pilou's method was quite easy and simple. And also accurate!
Yes, his way works, sort of. The second rotation is not perfectly accurate, though. I meant to simply draw it in place.
-
-
Nice challenge by the way, at first I thought this should be easy but then...
You can also perform the rotation as a tangent to the drawn arc, (referencing the vertices on the arc). Of course it is not exact but the angular error is vanishingly small (approximately a^2*t/l ; t,l thickness and length of the rail, a is angular resolution on the arc (total angle/number of segments) in radians)). The final rotation accuracy increases with the square of the number of segments on the arc. After a few hundred segments SU makes the structure ideal at explode.
Ogan -
Thomas, see Pilou's last post in this thread.
-
@dave r said:
Thomas, see Pilou's last post in this thread.
Wasn't that related to Gaieus' guideline experiment?
-
The thing absolutly crazzy is that the Rotation tool
seems accept just the base alignement as only good alignement!
I am totally depited! -
I get the result he shows in his close up when I try to use his method.
-
I see the issue now. doh!
...and I thought I had gotten peace at mind...
-
So how do you calculate it?
-
arg... I keep wanting to make quide circles... I see how difficult this is with native SketchUp now...
-
I'm not sure how you'd calculate it. I think the only dimension you know is the width of the board. the length of the miter and the length between ends will change as the angle changes due to changes in the spacing between posts and the height of the posts. And you don't know any of the angles either.
-
- Width of the board
- Diagonal of the board
Surely this should be something to work on?
-
But you don't really know the length of the diagonal of the board when the miters have been cut. And the overall length of the board is immaterial.
If you working in real wood, you would get an estimated length so you know the piece of wood you offer up is plenty long. You'd nail it in place and cut the ends off flush with the posts.
-
@jason_maranto said:
This is really easy with the native tools -- here's a short tut file: [attachment=0:3gkxsinp]<!-- ia0 -->challenge.skp<!-- ia0 -->[/attachment:3gkxsinp]
Best,
Jason.How do you place the horizontal guides such that you get the fixed width of the diagonal board?
-
Looking at it, it's just as if you skewed a 0,50m board - which is not correct.
-
This is really easy with the native tools -- here's a short tut file: challenge.skp
Oh, I see why this might be a challenge -- you want a specific width to the crossboard... give me a sec.
Best,
Jason.
Advertisement