Different values for Z on a single line....
-
Right! I may have forgotten to take my medicines today - but still..
In a fairly complex drawing in SU Pro 7.1.6860, I drew a line about 8-10 meters. Units are set to centimeters with 0,0 and no length snap. Then I change numbers of decimals to 0,000000 and check the Z-value by means of the "Add Height Text"-plugin.
Beginning of line Z=0,000000 End of line z=0,000016 and in between I get values for Z ranging around 0,000220 to 0,000240.
As it happends, I had trouble to get "filled" rectangles created by means of inference, drawing one side at a time. The rectangles wouldn´t close (create face..)
And - to continue, if you take reference from this newly created line and place a construction line, this construction line will differ regarding Z-value so it looks like I get a growing error for each reference I make. This occurs if I have altered the orientation of axis at some point. Regardless if axis are reset to original orientation or not. It seems to me that there is some kind of approximation going on here and I´m doing something really bad to make it noticeable in a strange way?(Edited by coloring for clarity. No other changes)
-
Is not some perverse to use such number of decimals?
You will must Scale *10000 or more (for avoid problem) then make your function then Scale / 10000 for return
That is not very productive
-
Hi folks.
SU cannot fill a surface when one of its side is smaller than about 1/16 inch or 1 mm.
The workaround is to scale up and, when done, scale down accordingly.
Just ideas.
-
But z=0,000016 and z=0,000016 in millimeters, centimeters and meters is still under 0,0 millimeters, right? Seems really perverse to me for SketchUP's capabilities!
-
Sorry, if I would attempt to give advice to someone - I would try to read the question until I had a grasp of the nature of the problem.
@ Pilou: You didn´t even bother to acknowledge the fact that I drew the line with settings of 0,0.
That is ONE DECIMAL and can hardly qualify as "perverse".I´m looking for a reason why rectangles with 10 meters side didn´t close.
When I increased resloution to 6 decimals (again, "perverse"?) to check what was going on, I found that the line does not sit on even Z-values. Why do you think my search for ev. causes of a concived anomality qualifies as "perverse"?
I was looking for an insight on internal approximation or something like that.
My bad - wrong forum... -
There is some kind of "tolerance" in SU when first something seems to be on axis but later it turns out that it is not the case. There are endless discussions about such seemingly minor issues but eventually they can become really bothering.
In the below file, there is a rectangle of 10m x 10m, edges set to display by axis. All is fine, the edges are red and green and if you PP the rectangle, the verticals will be blue, too despite one corner being just 0,05 mm higher than the others. There is not even an autofold line so SU definitely "adjusts" itself to such small differences.
I have to say that these "games" ar most popular on the Beta Forums and SU has been working this way ever since I can remember but we "refresh" these topics every time there is a new version.
Nilblo, never mind the critiques; these extreme, little things are that make discovering SU the most exciting thing (and also, plese, guys, be a bit more tolerant with others especially here in the Newbie Forum).
Advertisement