O/T trying to understand cmm results

[dolphin1]

Im machining a 5.900 radius on a arc segment of 36 degrees with a boring head and heavy metal boring bar.the part is a piece of alum standing verticle in a vise 4.3 wide and 5.00 tall and holding onto .750 in hardjaws.the tol is +/-.002 on the radius.

 

the cmm report is coming back saying the radius is changing as the boring head cuts down the part.can someone explain to me how the cmm comes up with the dia/rad getting larger at the bottom of the part.starts at R5.899 at the top and goes to R5.904 at the bottom.

[chowquoe^]

In my experience the CMM is not the most reliable source of measurement for a radius or diameter; although one would think that .004" is well with in its capabilities.

 

How deep is your bore? Is there any possibility that the vise is being over tightened and causing the pre-drilled hole to collapse and then spring open when the pressure is relieved?

[dolphin1]

the partial arc is on the face of a solid part.like i said its a arc segment of 36 degs.it goes 4.00 down the face.

[peon]

Just for xxxxs and grins, have the inspector remove the part from the cmm and set it back up and check the results again. If your cmm dept is anything like my old employer, we got different results every time. I never got a good explanation from them.

[Columbo™]

How true that is with the cmm/inspection department. two different inspectors two VERY different results.

[Rick Henrickson The Boeing Co.]

or for a quick check, just run an indicator down the rad and see if it is tapered

[Kevangel]

Does your spindle have an adjustment like a Moore Jig-grinder to grind a slight cone shape as opposed to tilting head like a Bridgeport and making same radius while you bore just on a angle? Does not seem very likely but thought I would ask and do you take a light finish cut and same results? I have seen some crazy #@%! from CMM inspectors..

[TonyNanni@BNB]

I don't think it has anything to do with the accuracy of the CMM. Think about what you are asking for mathematically. You want to predict a radius based on 36degrees of arc length. The smallest of error even .0005 will result in up to .010 error in predicted radius. when dealing with this type of CMM measurement I find it more practicable to ask for profile instead of predicting a radius. use the same +/-.002 countour tolerance. Depending on your CMM this should be a no-brainer.

[gcode]

My guess is,

your part is not perfectly vertical in your vice

so the C/L of your bore is not parrallel to the back face of the part

 

the CMM driver is hitting the back face a couple of times and defining that as a vertical plane

He then inspects your bore assumming it is

parrallel to the back face.

Since its not, his report says your bore is tapered.

The bore is not tapered, it is not parrallel

to the back face.

 

Try laying your part, back face down on a surface plate and sweep in the low spot of your radius

in a couple of places. I'll bet you find

the bore and back face are not parrallel

[MLS]

It is possible there is taper, so if you can confirm the part is correct by conventionally inspecting it...then perhaps you can explore how to check it properly. Reverse inspection process engineering. Heh

 

+1 Tony It would be better to check the profile deviation as opposed to the radius.

 

Simply bad practice IMO to check small diameters or radii segments with a CMM.

[TonyNanni@BNB]

make sure the CMM programmer takes individual radial scans. top, mid, bottom and dimensions each scan asking for profile of form only, not form and location. If you have a good CMM you can get plots of each scan.

[Kevangel]

Another way to check part at the machine is to pick-up a face and zero spindle centerline to face , now move out 5.900 from face and with an indicator holder set a dial indicator to zero on the face you had just picked up so that indicator is set to zero as it revolves around in the spindle at 5.900 rad. Now move back to spindle position on machine where you actually bore rad. and sweep in radius top to bottom and see how close you are to actual size. (very handy to have a Moore #3 jig bore with a 50 millionths readout to quickly check parts like this (us small shops do things like this)

[dolphin1]

quote:

---

Just for xxxxs and grins, have the inspector remove the part from the cmm and set it back up and check the results again.

---

have done this and the results are pretty close.

 

quote:

---

or for a quick check, just run an indicator down the rad and see if it is tapered

---

have done this and it doesnt show but a few tenths.

 

quote:

---

Does your spindle have an adjustment like a Moore Jig-grinder to grind a slight cone shape as opposed to tilting head like a Bridgeport and making same radius while you bore just on a angle?

---

im running a variaxis 3+2 5axis.spindle doesnt have a adj but i can move the trunion.

 

quote:

---

I don't think it has anything to do with the accuracy of the CMM. Think about what you are asking for mathematically. You want to predict a radius based on 36degrees of arc length. The smallest of error even .0005 will result in up to .010 error in predicted radius. when dealing with this type of CMM measurement I find it more practicable to ask for profile instead of predicting a radius. use the same +/-.002 countour tolerance. Depending on your CMM this should be a no-brainer.

---

thanks for this comment as i was able to get some help with this type.i agree it should be a no brainer but that was pushing the issue for now but tomorrow is another day.

yes i understand about the partial arc being a prob for cmm's but have developed work arounds but that didnt help with the inspect dept as they werent open to them.

 

quote:

---

My guess is,

your part is not perfectly vertical in your vice

so the C/L of your bore is not parrallel to the back face of the part

---

well the part is being machined in one setup so the vise isnt the prob.

 

quote:

---

It is possible there is taper, so if you can confirm the part is correct by conventionally inspecting it

---

we went there and it didnt show up.

 

quote:

---

make sure the CMM programmer takes individual radial scans. top, mid, bottom and dimensions each scan asking for profile of form only, not form and location.

---

it was brought up today at a meeting but the time constraints were too long according to them.

 

quote:

---

Another way to check part at the machine is to pick-up a face and zero spindle centerline to face , now move out 5.900 from face and with an indicator holder set a dial indicator to zero on the face you had just picked up so that indicator is set to zero as it revolves around in the spindle at 5.900 rad. Now move back to spindle position on machine where you actually bore rad. and sweep in radius top to bottom and see how close you are to actual size.

---

i might try this tomorrow as well as some other checks.

 

i want to say thanks for all the help.its been determined that the ways on the z axis are out and it appears to be more than one direction on the z axis.it looks like the head is changing direction in more than one way as it goes down hence the change in diameter to the cmm.