Debate Question

[[Me...]]

I hope that this is the right forum. I have a question which I've debated with co-workers. Is a CMM good for measuring runout at .001? I say no. I'd measure it another way. Am I right?

[[Ja...]]

This question is very open-ended. What is the accuracy of your CMM? Scanning? TTP? How repeatable is your feature/datum structure? Filter/outlier settings? How many points are you taking? Etc.....

[[Ca...]]

Inches? Metric?

[[Me...]]

We're just speaking generally. But I do have an example that I just came across. I small aluminum ring with the largest diameter at .750 inches and the part itself is only .350 high. It's so small that it has to be held down with 2 sided tape. It's my opinion that trying to measure .001" runout is asking for a lot.

[[Ca...]]

Depending on your hardware, checking 0.001" runout seems totally reasonable.

[[De...]]

If the call out is a realistic one then the CMM should be able to check 0.001 inches runout without a problem.

We have many programs where we are checking runout less than 0.001" - I am fairly certain we have at least one as low as 0.0002" and more than a few with 0.0003".

It cannot be stressed enough how much form error and outliers on your datum features can impact your results however. Even with very good form sometimes results can be skewed greatly based on the alignment of the datum features due to the effect of outliers on the outer tangential evaluation. Often a check of the datum feature for form error as well as a second check of runout using LSQ evaluation of the datum feature can give a 'sanity check' to verify if it is outliers that are adversely affecting your results.

When there is an issue that keeps it from being possible, it generally is caused by a combination of bad call outs on the blueprint and people that are used to checking things incorrectly on a surface plate.

In many cases, if you ask the engineer that specified runout of one cylinder to another cylinder to check it using manual inspection methods, they proceed to lay the part on a face on a surface plate push it to bumpers on the reference diameter and spin the part with an indicator. This aligns to a face and therefore is not a check of one diameter to another. In the same situation, the CMM will align the part to the reference diameter, where the check above aligns the part to the surface touching the surface plate.

Too often call outs are made that would not even be possible to check with a manual check - for example face runout of a 10 inch face to a diameter 0.250 long. This check is almost impossible on a plate, so people usually start by aligning the part against an angle plate prior to using some bumpers and never even realize that is not the check that was asked for by the print.

[[sh...]]

I've been able receive results using a CMM for Runout and V-block/indicator for Runout and usually get
them to repeat within .0001".


Then again it would also matter whether you are checking to threads, checking in the exact same location, or if you're creating
Cylinders or Circles.

[[Me...]]

Wow, thanks for all the info. I'll definitely trust it more now.

[[FE...]]

Hi Guys,
I´ve never measured runnout, so i have some doubts:
Do i need a Rotary Table to do this measurement? Or it can be done without a RT?

[[Ow...]]

Please sign in to view this quote.

Kudos for Derek.
I've always said knowing how to measuring things manually should be a prerequisite to becoming a CMM programmer.
If for no other reason than knowing how to double check their CMM program but, also to be able to articulately explain how a part may have been incorrectly checked manually. I don't have enough finger,toes or bones in my body to count how many times I've been challenged to prove my checks vs what they get on the shop floor over the years. If you can't, that's where a lot of companies lose faith in checking stuff on the CMM and I've seen quite a few of them.