I am trying to bend an Array Object along a Curve (using a Curve Modifier) that starts at the same 0,0,0 location as the object. But the Curve Modifier changes the Location and Rotation of the first instance in the Array, which I don't want.

I've discovered that if I rotate the anchor point it moves the object in various ways and I can almost line it up to match, but I am building a composite image that needs to line up exactly with the Array Object when it isn't aligned to the Curve.

Is there a way to reset the translation/rotation of an anchor point, or align an anchor point with the 3D Cursor?

Notice how the top cube moves when it is aligned to the curve? How do I avoid this?

enter image description here

enter image description here


  • $\begingroup$ Does this answer your question? blender.stackexchange.com/questions/71787/… blender.stackexchange.com/questions/49265/… $\endgroup$ Commented May 16 at 8:45
  • $\begingroup$ @DuarteFarrajotaRamos I do not know if you bothered to check if your linked answers are any helpful in this case, but for this question changing the Twist Method to Z-Up will only twist the cubes along the curve tangent and changing the Tilt of the control points of course also rotates along the tangent. It does not change the alignment of the face normal. $\endgroup$ Commented May 16 at 10:03
  • $\begingroup$ I may have misinterpreted that from the question, I was under the impression the user meant the rotation of elements. If that is the case, than the array modifier is probably not the adequate tool, and true instancing should be used instead. Either modern geometry nodes, or old style like particles or duplifaces blender.stackexchange.com/questions/68923/… $\endgroup$ Commented May 16 at 10:26
  • $\begingroup$ @DuarteFarrajotaRamos At least it were these statements: "that starts at the same 0,0,0 location as the object" and "that needs to line up exactly with the Array Object when it isn't aligned to the Curve" which gave me the impression he wants to get something like shown in my answer, but I might be wrong. $\endgroup$ Commented May 16 at 10:42

1 Answer 1


The starting point of the curve is... well, a point. The curve is bent and slightly rotated. The first face of the cube gets bent/rotated with its normal aligned to the curve point's tangent around its center, but since the face itself is not a single point of course some corner vertices might be moved away from their starting position. Because if you look at the handles of the control point, this is how the face gets aligned:

control point handles

If you want the first starting face to stay exactly where it was, you have to select the control point and scale it to 0 on X and Y, this way the tangent is pointing straight down and the face keeps its position. The problem is, the rest of the cube gets distorted nevertheless to follow the curve, and of course the shape of the curve changes when the alignment of the handles is changed. But since the title of your question is "Help Understanding Curve Anchor Points": that is the explanation what is going on there.

handles straightened

  • $\begingroup$ Thank you Gordon, this is helpful. I did notice that the rotation/alignment of the first anchor point would naturally position the first cube, but what I still don't understand how to do is align an anchor point to an axis. Is this possible? For instance can you reset an anchor points position/rotation to 0,0,0? $\endgroup$ Commented May 16 at 16:50
  • $\begingroup$ @GlenCandle If you mean control points with anchor points: just as I described if you have a Bézier curve. The curve's tangent is dependent on the handles, when you straighten the handles you straighten the tangent. For a vertical alignment (on Z) both have to be the same location on X and Y. For X alignment Y and Z have to be the same and X and Z for aligning on Y. Or you have to switch the spline type to Poly for example, then you have straight lines between points. In this case the end points of a line should have the same coordinates on two axes to align with the third. $\endgroup$ Commented May 16 at 17:22
  • 1
    $\begingroup$ Ahh, I get it, I wasn't connecting the handles with the alignment. Now that I've had that Eureka! moment I feel dumb because it's so obvious ;) $\endgroup$ Commented May 17 at 17:05

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .