Questions about Vector Hits

[B. Jacobs]

I am running into some problems here. I have four points that I created with the autofeature. I have taken my part and did a DCC alignment and ran the part. The part has a bunch of CAD points that I picked and am dimensioning.
I have a slot that is 24.4788 deg in the Y from the bottom of the part. I have taken the points, two on each side of the slot at about the same Y and Z distance. The problem I am running into is that in my X direction, the points are coming out different each time. I have my vectors set so the probe comes straight in at the X direction and takes the hit.
On one side of the slot, the dimensions are telling me that the slot is off in the X-. on the other side of the slot the points are telling the slot is off in the X+. I know this to be wrong because I have measured the slot width with gage blocks.
I have used 2 different tips, one a 1.5mm and the other a 4mm. With the 4mm tip my dimensions are as stated. With the 1.5mm tip, my dimensions are showing to be out the same way on both sides of the slot but the numbers are totally different than with the 4mm tip.
I have calibrated both tips numerous times and get the same results.
Would the vectors have something to with this or maybe probe compisater?
Each probe does have an angle change. It is the same for both tips.
Any help would be great.
Thanks

B/S ONE 7-10-7
3.5MR2

[B. Jacobs]

If I have ya'll cunfused on what I am asking let me know and I will try to explain better.
Thanks

[#2#]

It sounds to me as if your hit vectors are not normal (perpendicular) to the surface. If that is the case, your probe compensation would be all screwed up.

[B. Jacobs]

Here is a picture of the way I taking the hit. The green lines represent the blue vect lines when takeing an autofeature point.
Is the way I am taking them probably my problem. I am going straight in my X direction. I am doing this so my Y dimension is always the same. I need to know where my X hits.

[Matthew D. Hoedeman]

Here is a picture of the way I taking the hit. The green lines represent the blue vect lines when takeing an autofeature point.
Is the way I am taking them probably my problem. I am going straight in my X direction. I am doing this so my Y dimension is always the same. I need to know where my X hits.

If you need to know at an EXACT Y value, you will have to measure (using the correct vectors) 2 lines, one on each side of the 'slot'. Then, create a nominal plane or line at the Y dimension you want, then intersect them.

[#2#]

Here is a picture of the way I taking the hit. The green lines represent the blue vect lines when takeing an autofeature point.
Is the way I am taking them probably my problem. I am going straight in my X direction. I am doing this so my Y dimension is always the same. I need to know where my X hits.

It looks to me like the probe compensation would give you some error. I don't understand what you are trying to accomplish by doing it this way. I think you want the hits to always be at the same point in y. If the slot moves around a bit, the y will change because of the probe moving at an angle? Am I correct? Does your y always have to be the same for some reason? I am not sure how to get what you want. Perhaps measuring each side as a line and then if you can figure out a way to create a point at the intersection of your desired y and the line.

[B. Jacobs]

It looks to me like the probe compensation would give you some error. I don't understand what you are trying to accomplish by doing it this way. I think you want the hits to always be at the same point in y. If the slot moves around a bit, the y will change because of the probe moving at an angle? Am I correct? Does your y always have to be the same for some reason? I am not sure how to get what you want. Perhaps measuring each side as a line and then if you can figure out a way to create a point at the intersection of your desired y and the line.

Yes the Y needs to be the same every time. The reason for the points is the slot is cut on Op 2 and there is stock left on the outsides of the part.The attatchment on my previous post doesn't show it very good but the slot walls are at angles. The points I am trying to get will tell the operator what the deviation is based off of Op 2. This way the remaining stock can be cut off at the right angle and hold the wall thickness call out.
The points I am trying to get are programimg points for the machinist.

[#2#]

If you need to know at an EXACT Y value, you will have to measure (using the correct vectors) 2 lines, one on each side of the 'slot'. Then, create a nominal plane or line at the Y dimension you want, then intersect them.

Yes the Y needs to be the same every time. The reason for the points is the slot is cut on Op 2 and there is stock left on the outsides of the part.The attatchment on my previous post doesn't show it very good but the slot walls are at angles. The points I am trying to get will tell the operator what the deviation is based off of Op 2. This way the remaining stock can be cut off at the right angle and hold the wall thickness call out.
The points I am trying to get are programimg points for the machinist.

In that case, I would try doing what Matt suggests.

[B. Jacobs]

In that case, I would try doing what Matt suggests.

If you need to know at an EXACT Y value, you will have to measure (using the correct vectors) 2 lines, one on each side of the 'slot'. Then, create a nominal plane or line at the Y dimension you want, then intersect them.

I will give this a try.

[B. Jacobs]

Gave everything a try. Didn't seem to work.
When I use the small tip (1.5mmX40MM) the dimensions are showing being out in the same direction.
The large probe (4mmX70MM) still shows the deminsions out the oppisite way. What this means is one wall out in the minus and the other wall out in the positive. This would mean the slot is undersize, but it isn't.
I guess it is one of those things that make you go HUMMM....

[djayne]

Looks like the vectors are not normal to the surface as stated above - just from looking at your pictures. Rotate to the angle of the slot and then take the points.

[Matthew D. Hoedeman]

It is all due to the cosine error of the tips. If you are NOT hitting exactly perp to the wall, then as the size of the ball changes, so does the amount of error and if you are close to zero, it could easily go from one side to the other. Make your hits perpendicual to the edge and you will get correct values for the amount the wall is off, if you don't use correct vectors, you will never get the correct probe comp.

You COULD try tuning off the probe come for those hits and using an ACTUAL THICKNESS equal to the probe radius, but as a negative amount (or would it be positive?) and see what you get that way.

[[email protected]]

Center point between two walls

Hi I have something similar. I needed some help on how to measure this feature (0.052). I'm originally taking a point on both side of the wall and with a variable creating a center point between those two point to take the hit on top. On the drawing the part is symmetrical but on the actual parts those top point are skewed left or right causing a incorrect hits. I have nine of these little ribs that I need to measure. I'm using Xplus as my work plane so can't scan don't have enough room to get in there. Can I use a square ruby instead of using a 1.5mm ruby ball, not sure if one can do this or not? or any tip and suggestions would be most appreciated thanks.untitled.JPG

[JEFMAN]

I think it's very hard to measure the point with a good accuracy... The error of measurement decreases when the ball diameter increases ( the bottom of a square ruby could be assimilated to a infinite diameter), because of the positioning accuracy of CMM. If you can, try to construct the point (intersection for example), and compare to the measurement...

[DaSalo]

You could try measuring the facet on each side as a plane, constructing a third, generic plane perpendicular to them, and intersecting the three planes to get the point (the way the corner point feature works, but with the third plane constructed instead of measured.) That is a pretty small feature so definitely use the smallest ruby available. What kind of accuracy do you need to achieve here?

Depending on what the rest of the part looks like there might be other ways to think about it. For example, are these 9 features close together and do you need to know the height of each as an individual or would it be sufficient to know the height of a tangent plane across the whole group? If the later would work you could place a gage block on top of them, measure that, and subtract the thickness. You might also need to consider alternative measuring methods. You could take an impression of that with Facsimile or Reprorubber and measure the height on an optical comparator, for example.