Scale Issue
Hello,
To cut a long story short, I need to scale a drawing we've received in from a third party.
The length of the drawing needs to be 480. The current length of the drawing is 480.007.
The width of the drawing needs to be 360. The current width of the drawing is 360.006.
Does anyone know scale factor I'd need to use to achieve this, or how to calculate it?
Thanks!
To cut a long story short, I need to scale a drawing we've received in from a third party.
The length of the drawing needs to be 480. The current length of the drawing is 480.007.
The width of the drawing needs to be 360. The current width of the drawing is 360.006.
Does anyone know scale factor I'd need to use to achieve this, or how to calculate it?
Thanks!
0
Comments
-
Hi Elasticater,
Before you scale the drawing, can I ask what the units are? If they are in mm, would reducing 0.007mm make any difference in terms of plotting?0 -
It may not be a plotting-time issue, where yes the error would be undetectable. But in the digital model, if such things aren't spot-on, further modeling errors compound, things don't quite meet up, points not quite coincident as expected. It's worth getting it exactly right, to within the resolution of the digital model.
You need different scale factors in the x and y (length and width) directions. In Microstation for example that's easy, but Brics can only do it by the workaround of converting the model into a Block first; then on inserting it you're offered separate x and y scales.0 -
You need to calculate the relative difference between 480.007 and 480.
The starting value for a scale factor is always 1.
So:
480.007 = 1
480 = ?
So you multiply 480 by 1 and divide it by 480.007
So the formula would be:
? = (480*1)/480.007
? = 0,999985416879337
If you enter 0,999985416879337 as the scale factor the line 480.007 will become 480.
0 -
> I need to scale a drawing we've received in from a third party.
Sounds like originating from a PDF.
Where a PDF is vector based but the PDF's own data Grid is not fine
enough to keep the CAD precision.
So proper scaling for these special X and Y values may still not work
for all dimensions (?)0 -
You can directly—without any conversion—insert the third party's drawing as a block into a new drawing, setting separate x and y scale factors as needed, per Tom's suggestion. If you don't know the exact scale factors to begin with, don't explode the block immediately, then you can select that block to resize it by entering x, y, and z scale factors independently in the Properties panel.Tom Foster said:
You need different scale factors in the x and y (length and width) directions. In Microstation for example that's easy, but Brics can only do it by the workaround of converting the model into a Block first; then on inserting it you're offered separate x and y scales.
The Scale command has a "Reference" (aka "Base") option that I use often for uniform scaling, but to do non-uniform scaling I've had to rely on lisp, such as the attached.
0 -
I do agree with the idea of using the reference feature of the scale command. This method would even make it more accurate than the 4 decimal places, as BriscCAD does its internal math with more decimal place than what you have set to display. ( I think it is 12 decimal places) Though, this method limits yourself to just making one axis precise.
Be aware that there are limits to exploding a block that is not uniformly scaled. (at least there is in my v14). If there are elements, such as circles, a non-uniformly scaled block cannot be exploded. Also, you cannot use the scale, and reference approach to make separate X and Y scales. So, if you want it more precise, you should first set the decimals so they display as many as possible.
Don't get too caught up with being extremely precise. Keep in mind that a water molecule is about 0.27 nanometers across. = 0.0000027 mm. So, if your units are mm, that water molecule dimension is only 7 decimal places.
I suspect that modern CAD systems are more tolerant of non-precise issues like this. Though, in the early 2,000's, I found that rotating a 1x1x1" cube modeled in AutoCAD, and then importing it into Inventor would result in faces that were not considered parallel by Inventor. So, at the time those molecules counted.
-Joe
0