Why did my TCP come out with orientation (0, 0, 0) after teaching it?

Because the three point method only solves the tip location. It sets W, P, R to the default (0, 0, 0) and leaves it there. Finish the job with the six point method or type the angles in directly, otherwise the tool tilt is left unset.

The reported W, P, R don't match what I measure on the part. Why?

The Cartesian value the controller shows is a theoretical position calculated from each axis's pulse count. It's not guaranteed to match real space exactly, and orientation angles are always relative to a frame, so confirm which frame you're reading against before you chase a discrepancy.

Are W, P, R the same as the joint angles of the wrist?

No. W, P, R are rotations about the X, Y, and Z axes of the mechanical interface frame, a Cartesian description of tool attitude. Wrist joint angles are the physical axis positions. The controller can hold the same orientation through more than one set of joint angles.

Motion

Orientation (W, P, R): W, P, R (roll, pitch, yaw about the mechanical interface axes)

Orientation is the tool's attitude in space, expressed as the three angles W, P, and R that describe rotation about the X, Y, and Z axes of the mechanical interface coordinate system.

What it is

A robot position has two halves. The (x, y, z) part says where the tool center point sits. The (w, p, r) part says how the tool is tilted, its attitude in space. Together they fully pin down both location and pose.

W, P, and R are the angular displacements around the X, Y, and Z axes of the mechanical interface coordinate system. Tool attitude control needs them because how the tool is pointed matters as much as where its tip sits.

How it actually works

When you set a tool frame, the TCP position (x, y, z) and the orientation (w, p, r) get handled separately. The three point method solves only the TCP location from three approach touches; it leaves orientation at the default value of (0, 0, 0). To actually define the tilt you use the six point method or type the angles in directly with the direct list method.

At runtime the controller reports orientation right alongside position. In $CART_POS the array holds X, Y, Z, W, P, R in order, with X, Y, Z in millimeters and W, P, R in degrees. That Cartesian value is a theoretical figure computed from each axis's pulse count, so it will not match real space to the last decimal.

How it differs

TCP offset · The TCP (x, y, z) is where the tool tip is. Orientation (w, p, r) is how the tool is tilted at that point. Two different halves of the same position: the three point method sets the tip location and leaves orientation at (0, 0, 0), which is exactly why you still need a six point or direct method afterward to nail the attitude.
Reference frame · W, P, R are angles measured relative to a frame's axes, not absolutes. Change the reference frame and the same physical tilt reports as different numbers. Orientation is meaningless until you know which frame the angles are referenced to.

Where you meet it in the field

Set Up a User Frame with the Three Point Method · A frame gives orientation its reference axes. Set the frame first, then W, P, R have something to be measured against.
Arc welding · Torch work and travel angle are pure orientation control. Holding W, P, R steady along a seam is what keeps the bead consistent.

Common questions

Why did my TCP come out with orientation (0, 0, 0) after teaching it?: Because the three point method only solves the tip location. It sets W, P, R to the default (0, 0, 0) and leaves it there. Finish the job with the six point method or type the angles in directly, otherwise the tool tilt is left unset.
The reported W, P, R don't match what I measure on the part. Why?: The Cartesian value the controller shows is a theoretical position calculated from each axis's pulse count. It's not guaranteed to match real space exactly, and orientation angles are always relative to a frame, so confirm which frame you're reading against before you chase a discrepancy.
Are W, P, R the same as the joint angles of the wrist?: No. W, P, R are rotations about the X, Y, and Z axes of the mechanical interface frame, a Cartesian description of tool attitude. Wrist joint angles are the physical axis positions. The controller can hold the same orientation through more than one set of joint angles.