Vectors for angled hole

[PowerJunkie]

Hey all....
I have 1 feature on this part that has us stumped over here.

I'm trying to probe a hole which sits at a 5° angle and it's not working out. I can run in to the hole straight in or at 7.5°, neither seems to matter much, as a result I'm getting cosine errors as you can imagine.
We drew the hole up in mastercam and got the vectors. After I probed the hole as a circle, I plugged the vectors in (via F9 / hit targets). When I change the vectors like this and try to run it, the probe runs in to the hole wall and throws a "movement interupted" error.

If you need any other info, screen shots etc, let me know and I'll whip it right up.
Also, I searched and consulted with the help menu already, I just cant find anything that helps.

Thanks!!!


Vecors:
1) -.86273, -.08716, .49810
2) .500, 0, .86603
3) .86273, .08716, -.49810
4) -.500, 0, -.86603

PC-DMIS v4.2 MR1
Global Image

[bw_bob]

Did you do your rotation of 5 degrees so your alignment looks like the diameter angle. Then try & probe it

[JamesMannes]

Did you do your rotation of 5 degrees so your alignment looks like the diameter angle. Then try & probe it

Good idea. But, translate to the hole(x,y, and z) and then rotate, then you should be at a perfect 0,0,0 and perfect 0,0,1 vector.

[JamesMannes]

Hey all....
I have 1 feature on this part that has us stumped over here.

I'm trying to probe a hole which sits at a 5° angle and it's not working out. I can run in to the hole straight in or at 7.5°, neither seems to matter much, as a result I'm getting cosine errors as you can imagine.
We drew the hole up in mastercam and got the vectors. After I probed the hole as a circle, I plugged the vectors in (via F9 / hit targets). When I change the vectors like this and try to run it, the probe runs in to the hole wall and throws a "movement interupted" error.

If you need any other info, screen shots etc, let me know and I'll whip it right up.
Also, I searched and consulted with the help menu already, I just cant find anything that helps.

Thanks!!!


Vecors:
1) -.86273, -.08716, .49810
2) .500, 0, .86603
3) .86273, .08716, -.49810
4) -.500, 0, -.86603

PC-DMIS v4.2 MR1
Global Image

Or, are you using CAD with the part being inspected? Click on CAD and then incorporate sample hits on the surface.

Without CAD use the sample hits and you should be good to go.

[Underspec]

Translate your alignment 5 degrees OR you can use 3 sample hits to compensate for the 5 degrees. Translating the 5 degrees would be best if you require certain basic hole locations. Otherwise if you don't have any use 3 sample hits to determine the angle.

[Matthew D. Hoedeman]

Use an auto feature instead of a learned feature. IF you know the vector and the XYZ nominal, that's the way to go, no worries about all the individual hits or any of that garbage.

[PowerJunkie]

Well I didn't expect that I set the alignment to the hole with 5° rotation and translated xyz, re-probe hole and and I think it's working. It definately looks correct now.
Thanks

One more question. Since I'm not working with CAD it seems that I have to probe the hole, set the alignment, then probe the hole again. Or no?

[PowerJunkie]

Use an auto feature instead of a learned feature. IF you know the vector and the XYZ nominal, that's the way to go, no worries about all the individual hits or any of that garbage.

I'll try that next...

[Wingman]

Use an auto feature instead of a learned feature. IF you know the vector and the XYZ nominal, that's the way to go, no worries about all the individual hits or any of that garbage.

This is what I do if I know the normal vector of the hole.

[dwade]

Good idea. But, translate to the hole(x,y, and z) and then rotate, then you should be at a perfect 0,0,0 and perfect 0,0,1 vector.

This is what I do also and then if you don't want the extra alingment in your program, write the code for the feature, then delete the alignment and say "YES" to most confusing question the DEMON will ask you.... Update Commands.....

[PowerJunkie]

Aye, trying this with an auto circle. I'm using 4.2 and the "probe toolbox" wont let me edit the vectors. Am I just not "getting it" ?

[Wingman]

Aye, trying this with an auto circle. I'm using 4.2 and the "probe toolbox" wont let me edit the vectors. Am I just not "getting it" ?

You don't need to edit the vectors of the hits. Just set the normal vector for the hole with I, J, and K.
For example a hole set in an anle of 30° off the X Axis might be defined as follows
X = X Location From Print
Y = Y Location from Print
Z = Z Location from Print

I = 0.5
J = 0.866
K = 0

45° would have IJK like this
I = .707
J = .707
K = 0

The IJK values are the SINE and COSINE of the Angle from Square.

It looks like the vector points you listed earlier were actual hit vectors. What you need is the normal vector for the hole. The viwepoint if you will. I am not sure you have that.

So, if you translate to where the hole is supposed to be in XYZ then rotate 5° so that the hole is normal you can measure it like James suggested with 0,0,0 location and 1,0,0 vector (or whatever plane you are in is the 1 vector).

Basically, do you know the nominal location of the hole? You should. Just translate to it in X,Y, Z (offset your alignment). Then determine the closest Axis to it (the one that is only 5° off). I will use the X Axis for demo here. It will be the same concept for Y or Z. If the hole faces 5° off from the X Axis the you want to ratate about the Z Axis 5° to make the hole Normal to the X.

[Wingman]

Aye, trying this with an auto circle. I'm using 4.2 and the "probe toolbox" wont let me edit the vectors. Am I just not "getting it" ?

I have posted a pic of what I explained in last post.

[JamesMannes]

I have posted a pic of what I explained in last post.

Do as Wingman says and you will be good to go.

[#2#]

If your drawing is lacking dimensions to be able to do it as described above or you are looking for a quick and dirty way to do it...

Measure it as a 'learned' cylinder with say 8 hits on 2 levels. Do not edit this feature. Next, align to this cylinder. Then, either put in another 'learned' cylinder or and auto-cylinder at 0,0,0. This way, you don't need to know the angle to rotate to, the vectors, or anything.

That is what I do when I don't have a print or the print does not have enough information to do it the 'better' ways.