My point was that i checked the sketchup geomtry against another calculator. They matched to 6 decimals.
1 Like
One thing to note is that area calculations in Entity Info are affected by the segment count.
1 Like
That’s good to know…
But in fact a little strange as SketchUp internally works with real circles…
the arc length is shown in the entity info box. why not just create a line of the length given for the arc?
Actually, the situation is a bit complicated and potentially confusing to those who aren’t math or code types. SketchUp represents a circle (or circular arc) as a series of straight segments - always. However, when it knows the series was generated using the circle or arc tool it attaches a block of metadata containing the true parameters of the circle. This metadata allows some modifications of the circle’s segmented representation via entity info, such as changing the radius or number of segments. Alas, some types of further modeling involving the circle force SketchUp to drop the metadata, and then the modifications are no longer possible.
This is why I suggested using the pie tool to create an arc. An arc retains the meta data, but a circle will lose it when you cut it with a wedge.
@slbaumgartner , I know that circles and arcs can get exploded by some tools like Follow-me.
What I don’t understand is that Entity info displays the correct length for a circle and arc, but doesn’t display the correct area as @Anssi told me…
I don’t work for Trimble so I can’t give the design rationale, but for whatever reason SketchUp always calculates areas and volumes from its segmented or faceted representation regardless of whether metadata is available.
2 Likes
SketchUp gives the area for faces that are attached to edge segments. I don’t work for Trimble either, but I understand that to be otherwise, the application would have to maintain a separate algorithm for geometry-based and face-based calculations, and they have decided to keep it simple. For things like room areas the error is tolerable in most cases.
1 Like
yeah.
I did a quick test using an online calculator.
1m radius circle vs 24-gon (standard SU circle) vs 48-gon
here are the areas, perimeters and deviation percentages
-
Circle :
area : 3.14159 m2
perimeter : 6.28318 m -
standard 24 sided circle :
area : 3.10583 m2
% margin : 1.14%
perimeter : 6.26525 m
% margin : 0.29% -
standard 48 sided circle :
area : 3.13263 m2
% margin : 0.29%
perimeter : 6.27870 m
% margin : 0.07%
These result show that a standard 24-sided polygon is already really close to a circle. 1,14% difference of the area is really small.
And if you feel it’s not enough, simply by doubling the sides to 48, you get… well really, 0,29%, it’s a rounding error at this point.
Off course, if 0,29% is still too much for you, then you can still switch back to 2d tools. SU wasn’t designed for aerospace engineering 
I love this thread BUT
My head is exploding!
I am an ex military pilot who is used to working to tight real life tolerances where lives are in critical danger if I get it wrong, ± 15 seconds, seems a long time but is 0.0173611111111111% of 24 hours. So why we are arguing/discussing something that is the smallest possible size for anything in the universe, the Planck Length , which is 1.6 x10-35 m across.
Anyway unless the distances/dimensions are super critical what is the issue with being as accurate as 50+ decimal places of any units used in an Architectural drawing program…
And then we have the issue of line thickness - is the area taken from the inside, outside or median value?
Best wishes for a happy and prosperous New Year - it can’t be as bad as the last 3 years!
4 Likes
If you want as many places as SketchUp Ruby can give you, select a circle or circular arc and then enter this into the Ruby Console:
curve = SketchUp.active_model.selection[0].curve
(curve.end_angle - curve.start_angle)*curve.radius
Note: if the circle or arc has been scaled or is in a group or component instance that has been scaled, the value will need to be corrected for the scaling. That’s beyond what I have time to enter just now, as it depends on the depth of nesting and the scaling at each stage.
1 Like
Oh wow lots of activity here. Thanks for everyone’s help.
I am only looking for the line segment to be as precise as whatever SketchUp considers to be the smallest value it can handle. I also don’t mean connecting points either since SketchUp will connect points even if they are 1/1000" off. I will try out this Ruby code. From what I have experimented with 14 digits (in inches) after the decimal place is the upper limit right?
1 Like
Some stuff to read if you have spare time this year:
1 Like
In other words, the earth ain’t even this perfect. That is unless you’re looking at it from Mars or the like. Then it will appear more smooth than the smoothest Q-Ball you can find in our world. This info is from an astronomer dude on the History Channel.
As an architect, somewhere around 1/32 inch is “close enough” (actually way too close enough).
I leave the accuracy variable to “as provided by others”.
However, in my limited graphic mind, I’m thinking that one could explode the arch and grab a line segment from the arc, rotate it to x-y, then array (multiple copy) it using the number of arc segments as the “X” factor. This info is from a “math dumb” architect…LOL
I began using Cad a long time ago because I could solve these kinds of curve problems there-in as I didn’t know how to mathesize it. The contractors loved me for the “precise” information (1/8" variable).
The whole point of this discussion is that if you do it your way the total length will be shorter because an arc consists of straight lines instead of a real curve.
As you should know a straight line is the shortest way from point to point…
yeah, I get it…due to the faceting of the circle perimeter it’s shorter:
the more facets, the more accurate the result, still like a dog chasing his tail tho.
My only point is that I’d rather solve graphically, than by math cuz I’m very un-mathly.
I also discovered that the distance across at earth’s equator is 8000 miles while
the circumference is something like 26,000 miles when I was trying to understand curves better.
Who knew?
Guess you mathe-ma-dudes do.
1 Like
Couldn’t resist my besserwisser urges. It’s 24 901,461 miles
24 900,— miles
I have always wondered who has been willing to go the extra mile to close the circumference gap
3 Likes
Thank you for the clarification.