Accurate rotating

[itsnathanblake]

Hi guys,

yet again the architects have ****ed up on the coordinates for a new site plan (they never learn) and I'm trying to resolve this, i have managed to get the coordinates bang on but, when trying to rotate im wondering if theres a command that allows me to select different points that will rotate this for me as i dont want to eye it in for obvious reasons.

Below's where I'm at currently -

[vinodkl]

Hi @itsnathanblake 

Are you looking to Rotate objects with Reference?

[peterm · Accepted solution]

You can rotate by angle or reference.

ROTATE - pick base point of rotation - R - select rotation base point, then select a reference point on the object you want to rotate, the select the target point where you want to rotate the reference point to.

Hope that makes sense.

[Valentin-WSP]

You may want to look into "Rotate and Scale Objects by Reference":

https://knowledge.autodesk.com/support/autocad/learn-explore/caas/CloudHelp/cloudhelp/2021/ENU/AutoC...

[Kent1Cooper]

[EDIT: Message 4 came in while I was writing this, and its link is more detailed, but here you go anyway....]  In case a different way of explaining Rotation by Reference helps, try the one in Help.

[leeminardi]

I find the rotate by reference command to be one of AutoCAD's least intuitive commands.

Rather than use it I would set the angular dimension command precision to 4 decimal places, add an angular dimension and use that value with the rotate command.

Better yet, use ALIGN.

[j.palmeL29YX]

---

@leeminardi wrote:

I find the rotate by reference command to be one of AutoCAD's least intuitive commands.

---

May be. But it is (besides the align command, this would also have been my second suggestion) a very elegant way to rotate objects to an exact angle to each other. 

If we follow your suggestion (manual type a previously measured value) we may produce inaccuracies in the drawing - and only a few minutes (or days) later this inaccuracy will fall on our feet. 😞 

I'd suggest never to use "measured" values (as result of a dimension/measure command, or the list command, or properties, or similar commands which give an "informative" number [with limited precision]) for following construction steps (as long as we have the possibility to "grab" this values from existing geometry). 

my 2ct

[leeminardi]

@j.palmeL29YX I agree that you would probably get more accurate results (i.e., with greater precision) using rotate reference or align than by entering a value for the angle of rotation.  However, the absolute precision of rotate r and align are a function of the points used and in general will yield higher precision the further apart that the reference points are located and neither may yield "exact" results. The location of the points with respect to 0,0,0 WCS will also affect precision. The real results also depend on the final angle of the geometry (e.g., 45 degrees will be exact other angle maybe not).   Using a rounded angle to 4 decimal places as I suggested will most likely yield more error than using reference points with rotate or align. The question becomes are the results good enough?  With 15 significant digits the answer is most likely yes.