Accuracy and repeatability

[Smallsvillanova]

Our CMM's are calibrated and certified by a third party once a year. Are there any "tests" or "checks" you guys run throughout the year to ensure you have not lost any accuracy or repeatability between certification or even after an unfortunate but serious crash?

[louisd]

Here's my $0.02, if your crash consisted of anything below the probe head wrist, the chances of the crash affecting your machines linear accuracy, squareness, or other factors affecting the volumetric calibration of the machine is super low.

If your crash involved the bridge itself (someone put something heavy, to block path on air bearing ways), the Z Quill, or the upper portion of the probe head (where it mounts to the Z quill), then yes I would assess the CMM, suggest unscheduled maintenance, or re-calibration. If you don't have a Koba bar/step gage, or any other resources to evaluate the machine's scales, I would recommend the latter two options.

Of course there are always unique situations... how fast the cmm was moving when it crashed, what it crashed into, visual signs of the extent of the crash (dents/scratches) if you weren't present... I mean if you crash in slow mode with the jogbox on 2... Or run into an empty cardboard box and it squishes against a wall or desk until drives stop... i'm pretty confident these types of situations wouldn't damage anything.

[louisd]

if its below the probe head wrist, your sensor, module, stylus, and possibly your head will be fubar before anything else. Nothing better than going to the boss and asking for $12k for a new probe head.

[RIDER]

Nadcap M&I requires that you have a procedure in place to have verification checks scheduled between calibrations and also after an event occurrence. Examples may include earthquake, machine crash, relocation of equipment, or fork lift crashing into the CMM. The process may include periodic checks of master parts, artifact or calibrated gage. Since management at the time did not want to fork out for a step bar, I chose a master part. I measured the part 10 times on each machine to establish a baseline. The part gets measured monthly or after an event occurred and the results are compared back to the baseline.

Also, if I get any questionable results I will always grab a gage block or ring gage and measure them on the machine as a sanity check. HTH

[JEFMAN]

A ball bar can give a lot of informations quickly, and doesn't need to be certified nor calibrated, only rigid !
You can construct it on your own, with 2 calibration spheres and a rigid stem, and measure it in different orientations, idf you measure too much defects, the cmm is out !

[Peter Fuller]

I would say that you should still be using certified spheres since the form of the sphere can directly affect measurement accuracy.

[JEFMAN]

Peter Fuller : I talk about the length...
Using calibration sphere should give enough accuracy, calibrated or not (In this case, you don't search about 1 µm...)

[Disal]

I keep a "master" part from my major programs. I run those every few months, i haven't gotten any differences more than a couple .0001". It's more just like a quick check for me. i do this when i replace a stylus also (after calibrating).

[None_the_wiser]

We have a master part with 5 spheres that we run every 6 months (10 times in succession to check the repeatability) and also have a certified 6in ring gauge we use for the scanning function that is also run every 6 months.

[Smallsvillanova]

Thank you everyone. Very Helpful as always.

[Wes Cisco]

How deep down the rabid hole do you want to go?

Somewhere on this site you ought to be able to climb up onto the shoulders of a retired giant, Hilton Roberts. He wrote a most excellent report on how and why to use a 'golden part' for interim CMM machine system checks.

HTH & ymmv

[Smallsvillanova]

oooh I am very intrigued...

Off on an adventure I go!

[Wes Cisco]

https://www.pcdmisforum.com/forum/pc...nd-am-learning

https://www.pcdmisforum.com/forum/pc...rocess-control


P.S. Those were what I was able to find with a quick search, but not what I was referring to. There was a much longer document he posted years ago . . . maybe on the predecessor to this forum. It had to do with how an interim check revealed there was an issue, but not exactly what the root cause was. It then walked through the troubleshooting steps needed to drill down to identifying the source and implementing a solution. I have it stored somewhere, I'll see if I can find it and repost it here over the weekend. If Ironman is reading this, he can probably put his digital finger right on it.

[JEFMAN]

Here also :
https://www.pcdmisforum.com/forum/pc...check-your-cmm

[Smallsvillanova]

Wes Cisco a rabbit hole indeed. A brilliant and well spoken man. I have a lot of reading to do. Thank you.

[Smallsvillanova]

JEFMAN a very good read. Anyone who has not read what these gentelman have suggested i strongly recommend it