How many hits for a hole?

[Big John]

Guys,

I have a question. How many hits are needed to measure a hole?
At my company some people are starting a gage R & R test. I said that a minimum amount of hits are 3. With 3 hits it is not possoble to measure the roundness, so more hits are needed.
I then mentioned that I had heard somewhere that Prime numbers (ie 3,5,7,11,13,17 etc) are the best to use.
Is this so? If yes why?

Cheers

John

[John Kingston]

There have been numerous discussions on this. Here is one.

[#2#]

Well, I can't tell you what is best. I can only tell you what I use and why.

I almost always use 11 because I have found that on my machine with my probe, that gives me the diameter values which are closest to those measured by a bore mic. I don't do a lot of roundesses. I am also not in a business which requires high speed and throughput.

What I have heard about using prime numbers. On machined parts often times vibration will cause a regular patern of irregularities. By using a prime number for the number of hits you will ensure that the hits will not always fall on the same part of the irregularities.

Imagine a triangular part. 3 hits or any multiple of 3 hits will give a distinct pattern. With a square any multiple of 4 hits will give a distinct pattern. A prime number of hits will ensure that some hits will fall off of the patern.

[Underspec]

The more the better. I usually use 24 hits (Every 15 degrees).

[Matthew D. Hoedeman]

Just 4 hits, but I am automotive sheet metal, not NASA.

[craiger_ny]

I start out with 6 or 8 if there are no form requirements. Some processes and call-outs require more hits and/or strategic placement of the hits as Goodluck mentioned in regards to process induced form issues (usually, but not always workholding induced).

Craig

[JoeRZ]

Because of the algorithm used for diametral measurement, I tend to vary the amount of hits with the size of the bore / shaft, more surface - more hits.

I also use even numbers to simulate 2 opposing point method but I do run in situations where primes are more suitable.

Keep in mind that CMMs are not the best tool for diametral size / form measurement - but if you're looking for locations, it's pretty reliable.

[Underspec]

The more the better. I usually use 24 hits (Every 15 degrees).

Depending on what industry and what you are dimensioning. I'm in plastics and I dimension form alot. Yes the CMM is not very accurate for these types of things but the graphical analysis helps alot. For hole locations I agree with the 3-4 points.

A scanning CMM would be nicer if management would dish out the fund$. But, I have to work with what I got.

[rangerboat72]

According to the National Physical Laboratory (NPL) - Measurement Good Practice Guide No. 41 by David Flack, the recommended number of points for a circle is 7 (to detect up to six lobes).