CMM runout reading different from hard gaging

[[Du...]]

currently checking this part on my cmm and getting a measurement of .015 for the runout of the highlighted feature to datum A, the machinist checked it in the lathe and is getting .004 at most, could anyone explain how we are getting such different results?

if it helps my base alignment is a 3d line made up of 2 circles on either side of the part for X and Y and the plane on the left side of the part as Z.

[[Ma...]]

How machinist Is measuring this One ? The piece Is worked blocked on center opposite hole ?

However you should consider that surface oscillaton afflict base allignment. And tò me the best allignment for measuring this One is tò take axis on external diameter 1.25 and After make comparison beetween "C" and "A".

 

If Pieces Is worked by center holes, you should Better to try a 3d line for allignment taken on 2 self centering point ...so you can replicate working axis by machinist.

Edited yesterday at 07:03 AM

[[Ma...]]

You can not trust lathe in this - you can not know if piece is clamped perpendicular to machine axis,

Here is simply short datum - this should be measured as A-C and ignored alone A or alone C

[[Cl...]]

As far as the CMM result. It's most likely a larger error due to projection error from Datum -A- which is too short and too far from the evaluated feature.

Edited 20 hours ago

[[To...]]

Please sign in to view this quote.

In theory, you are correct.  However, as inspectors, we don't have the option to interpret a bad drawing any way we want. 

[[Du...]]

Please sign in to view this quote.

would it be valid to check this while constraining the normal vector of the datum in the evaluation? 

[[To...]]

Please sign in to view this quote.

This is probably one of the most common "issues" when comparing results from the CMM to the lathe.  The problem is that any deviation, however small, when establishing the datum feature A, get magnified as the axis extends out to the considered feature.  For example, .0001" (0° 0' 46") axis deviation over the 1/2" length of datum feature A, causes the axis to be skewed .0026" when extended 13" out.

Martin is correct that the drawing should set up the datum runout reference frame as A-C.  Plus, have two runout specs   A to A-C and C to A-C.

Edited 20 hours ago

[[Ma...]]

Please sign in to view this quote.

Sorry for misunderstanding - in my answer i was targeting what should be changed on drawing - not how we should measure it. My mistake :-)

We are making forged parts, so customer sends drawings, engineers are making some changes due forging technology and they are in tact with us for corections on callouts.
For tools we have ( as many others i believe ) almost always something to correct on drawings just because engineers have a little knowledge about callouts evaluation.

[[To...]]

Please sign in to view this quote.

no worries.  I realize that sometimes there is some confusion between people speaking differently languages...

Respectfully,

Tom Oakes

Multi-Lingual Illiterate (cannot read or write in many languages) 🙂

 

[[Cl...]]

Please sign in to view this quote.

I'm not sure that would work, due to the size and distance between them. I think the drawing is wrong. We ran into this exact problem and the part would not pass a Type-2 GRR. We had to petition the customer to change the datum to something more reasonable.

Edited 19 hours ago

[[Ky...]]

If you constrain the normal vector, you will effectively be using datum A cylinder for the center, but your base alignment would be controlling the direction of the axis. This is usually not what is desired.

 

This is a very common question on the phone support line, and there is unfortunately no real good answer. Short and wide cylinders just do not make great datums (and honestly, I have my doubts about something that thin realistically controlling orientation in the finalized part). I've made this comparison before, and I will probably again, but this is like trying to throw a penny into a 3-jawed chuck and hoping that it will have the same orientation each time.

 

As stated, there is no "solution". The best course of action honestly is to discuss with the designer/customer and try to get them to change the datum to something more reasonable.

[[Da...]]

I am fully expecting that the machinist is staging that part on both the A and C datums, and is actually checking local runout. The callout specifies it is only to datum A, which is a 1.42" long cylinder about 10" away, which I'm nearly certain is NOT the way he measured it. It sounds nearly impossible to measure this without a CMM. What he is measuring is what would be correct if it was written |.002|A-C|. With measuring it only to datum A on the CMM, you have one cylinder in space and another cylinder in space, then the runout captures both the local runout PLUS the offset between the 2 cylinder axis. So, his .004 local runout may be correct, but add that to say, an .011 axis offset, and you get .015 pretty quickly.

[[Ma...]]

It seems that all of us are taught that if a datum can control a degree of freedom then it must. There seems however to be no rules about what to do when a datum theoretically can control a degree of freedom but in reality it can't.

Two of the most common examples are the case presented here, and the case of a short cylinder as |A| and a relatively large plane as |B| with callouts to |A|B|. 

In neither case can the primary datum actually control the orientation of the part. When you encounter this what do you do?

• Try and get the drawing changed
• Evaluate according to the drawing
• Evaluate the way it will function
• Evaluate it according to the drawing and according to function and let someone else decide what to do about it?

I've done all of these multiple times over my career.