LSQ evaluation and roundness

[[Am...]]

Hello everyone, I am looking for some insight on how roundness can affect an LSQ evaluation. Would you expect to see a big variation in diameter change if you fixtured a part in such away that it hold the part more round? I would think that using LSQ the result should not change too much because all of the points are equally weighted.

I have a part that has some crazy roundness variations, the material has a lot of stress and depending on billet position the roundness is going to vary. I ran a Gage RR and found the results very repeatable 1.5% GRR even with parts varying .015" they all consistently measured the same. I made a fixture to kind of disprove to my self that the roundness would be changing the results much from the LSQ evaluation but it did give me different results and more preferable results. Now I'm trying to figure out if im stretching the parts with this fixture and it is not a valid way to fixture them.

[[Da...]]

Unless you have a good reason to do so, LSQ is not the correct algorithm for reported characteristics.

Its great for Alignments, and a few other things, but not really what you want in general. It is consistently, repeatably, wrong. See the attached document for more information.

If you want to see the actual shape of the part, use the CAD->Evaluation feature on the roundness. That will show you the shape.

Algorithms_for_CZ_Workshop.pdf

[[Je...]]

.

That's good advice, David.


The CAD Evaluation in Calypso is a powerful tool in communicating feature attributes. Here's a few examples. I bet you can spot the image that helped convince a machinist that his 3-jaw chuck was exerting too much clamping force, which was tri-lobing parts.


Jeff Frodermann
Windings, Inc.
New Ulm, Minnesota

[[Ja...]]

I am not sure how to answer the original post. If the part is flexible then that is a problem. The cookbook suggest using minimum zone for roundness. I wouldn't say LSQ is never the right thing to use. When evaluating position you should evaluate the feature using LSQ.

[[Ow...]]

Some parts are nearly impossible to manufacture without inducing deformation or seeing how innate material stresses affect the form/part after machining and temperature acclimation.

Evaluation is very important however, keep in mind that it may be totally irrelevant once the part is assembled to mating parts and any deformations they may have.

If it's possible, there is nothing wrong with creating a fixture to check a part that simulates exactly how the part conforms after it's assembled to its mating components because, that can often be cheaper than changing manufacturing processes or material.

Other than CMM programming errors, it's probably the No#1 reason machinist/assemblers/engineers quickly disregard CMM results. 🫣

[[Am...]]

The reason for LSQ is this is a check for process control and what the engineers want to see. I was just surprised to see the big change when making the part round. Just logically in my head with LSQ being all equally weighted points I would think that making the points round would make the part closer to where the LSQ diameter is.

[[Cl...]]

• Provides a consistent,
stable result (for what?)

(Why is it an option then?).
• Consistently provides the
– Wrong Size
– Wrong Location
– Wrong Form

[[Ke...]]

Please sign in to view this quote.

I believe it is recommended in the Zeiss cookbook, for process control studies(?)

[[Je...]]

Please sign in to view this quote.

Thanks for bringing this up, Owen. Very true.

I guess industry and product play a roll in the valuation of CMM results. In the industries my company serves (aerospace, military, oil & gas, performance racing, precision automation), the machinist/assemblers/engineers do not have an option of disregarding the CMM results. Doing so could result in... highly unfavorable consequences.

Instead, there is often subsequent evaluation requested to help minimize the uncertainty of the CMM results and, as you mentioned, consideration of part functionality. Parts either get rejected or dispositioned to "use-as-is" by the customer, which is always documented.

Another way to look at it is low-volume prototyping vs. high-volume manufacturing.

Jeff Frodermann
Windings, Inc.
New Ulm, Minnesota

[[Ow...]]

Jeff Frodermann,
Yeah, I was kind of speaking general to a maybe a small percentage of people who have worked with others who dispute CMM result, it happens.
I don't let anybody walk away feeling uncertain about anything I've checked on a CMM. I've proved a lot of people wrong over the years and have even been wrong myself a couple of times in the last 30+ years. 🤣
Clark,
• Provides a consistent,
stable result (for what?).
CMM Repeatability.