I need help with measuring TP

The angular rotation is a basic and I'm thinking Polar Rad isn't what you want to report ? Having a tough time wrapping my head around this part but the holes are at " another angle " as well ? Another defined BASIC ?

When I measure holes on an angle I create a circle from the intersect of the cylinder with the face it's on, this is what the print implies. From your report your PA is good but your PR is out, if it's still out after creating intersect circles then the bolt circle pattern needs to be adjusted smaller. Also, create a bolt circle from the 36 holes & verify it's at 31.50.

Don't have all the print views I need, but here we go:

The surface the hole is defined in is either also defined by basics, or there is a plane down from -A- where that Ø31.500 bolt circle is hitting, as it clearly isn't the intersection point with -A-.
That point MUST exist, or you are under-defined on the drawing. If the face is defined by non-basics, I probably wouldn't argue it much with design, but you want to make manufacturing hold it REALLY close so you don't get a different answer than your customer during inspection and not be able to argue why you are right.

Align ABC, which I presume puts you in the middle of a main bore, flat on -A- face, rotated so that -C- is pointing along an axis (I'm going to say it is pointing +X for the balance of this, but it could be -Y, you'd just have to adjust what you translate).
Rotate 5° around Z, the basic rotation, bottom right.
Translate down wherever the Ø31.500 is contacting surface (I hope that is a basic in a view you didn't provide, otherwise, you'll have to calculate it off the definition of that angled surface).
Translate X 15.75, that is half the basic diameter.
This puts your XYZ zero right at the mouth of where a perfect hole would be.
Rotate around X around Y 120° (you could rotate Z around Y 30°, but if you do what is explicitly on the print, you don't have to think about what you did later during review).
This has your alignment zeroed at the mouth of the hole AND rotated along the axis of the hole.

Output in Legacy the Position of the hole.

Tell it to give you BOTH start and end. You'll know where your cylinder is, if it is crooked, and why you have an issue.

If you are programming off the CAD, and the alignment to the CAD isn't QUITE perfectly in agreement, the software might think the hole is somewhere other than it is supposed to be.

you'll see this in the legacy command, as your axes should be both at 0. No numbers as a theo, you moved to the perfect (True) location. That is the concept with basics, you move to perfect and report Actual Position against the True Position. If your theo's have a number, something is wrong with your alignment, or something is wrong with your CAD, or, something is wrong with your alignment relating to the CAD.

If the theos are zero, and the hole is off, you'll see if it is off at the top or bottom.

Is there a burr.
Did you go too far out the bottom of the hole and you are shanking.
Is the probe angle off enough that you are shanking.
Did you take three levels in that cylinder?
Did the software solve the cylinder in the correct axis, or did it rotate it? (three levels will help prevent this, though it isn't 100%, especially if you shank or have other form error)

Lots of ressons why, but start with looking at hard data you can conceptualize.

I don't know anyone that thinks in polar coordinates.

...

I don't know anyone that thinks in polar coordinates.

hahahahaha!!!! just a certain type of bear!!!!! and Santa, of course

I'd take and create 36 alignments, offset and rotate to where the theoretical perfect cylinder is defined on the drawing then tolerance the position of each hole to perfect 0/0

"Tell it to give you BOTH start and end. You'll know where your cylinder is, if it is crooked, and why you have an issue."

XactMeasure does not give you the option of choosing START / END

My update (sorry but I've been a bit busy)....
​
Thank you so much...I corrected almost everything, my problem was that I gave the presumed position of these holes at an angle, calculated it and gave it the correct one. I have a bigger problem now, every time I measure the same part of DATUM B it turns in a different direction. The diameter of the piece is approximately 33 in, the height is approximately 5 in. Datum B is a cylinder with a depth of about .300, additionally, the holes at an angle are drilled close to this datum, and the part is made of aluminum.
I know I have 36 holes, but they asked me to measure 4, which are opposite on both sides.
Really need help. How to stabilize DATUM B.
I created 6 circles every .050 and then concentrated them into a cylinder (the circles were ok, but the cylinder was not).​

measure datum 'A' plane
measure datum 'B' circle
measure datum 'C' circle

create a line between circle 'B' and circle 'C'

align datum 'A' in Z and datum
align the created line in 'Y' axis
datum circle 'B' in X and Y axis

Do the same thing again in auto, When you do datum 'C' use a sample hit inline with the hole ( ie 12 clock )

when you measure the angle holes use a sample hit on the angled surface or even better measure the angled hole as a cylinder and measure a line on the angled surface at the center of the hole and intersect the 2 features.

As the angled holes are 10' apart you should be able to do a paste with pattern.

Hope this helps.

There is one problem...

DATUM A - PLANE - Z
DATUM B - CYLINDER - facing up - can controls X and Y
DATUM C - on the side - can controls Z and X

What you said is TECHNICALLY true, but not possible given what you are told to do by the FCF.

-A- plane, controls rotation about X and rotation about Y (two rotaions) and it controls translation in Z. Because it CAN and the engineer called it first (primary) so it DOES.
That leaves rotation about Z, translation in X and translation in Y.
Doesn't matter what the following things CAN do, they are only allowed to control that which is still available after the primary datum takes its chunk.

-Y- cylinder, can control rotation about X, but that is not available, so it doesn't.
Can control rotation about Y, but that is not available, so it doesn't.
Can control translation in X and Y, those are both available so it does.

That only leaves rotation about Z.
That is all.
Whatever comes next, it doesn't matter what it CAN do, only what it has left and is allowed to do.

-C-, cylinder can control rotation about Z, this is available, so it does.
Can control translation in Z, but that isn't available, so it does NOT.
Once rotated, can control translation sideways (you said X), but that is not available, so it does NOT.

Primary takes what it can.
Secondary and tertiary take whatever is left over to them.
Every time.

There is a method an engineer can employ to over-ride this, I've never seen it on a print, you would KNOW it if you saw it.

It is not in the snips of print you posted.

I'd also make the line from B to C as it will provide better rotation, but, in the instance -C- is not on center (and the more off center it is the worse this gets), it is the axis of C that rotates about Z, not a line from B to C. That would look like | A | B | B-C |
So long as -C- is sufficiently close to the center-line of -B-, a line will work and be representative of reality.
"Sufficiently" is dependant on your tolerances, actual product and how risk averse you are.

Using an axis that is .5" long (or however thick your wall is with -C- in is) to rotate a 33" diameter is going to carry it's own risk.