hexagon logo
Register Sign in
  • Replies 19 replies
  • Subscribers 16 subscribers
  • Views 232 views
  • Users 0 members are here
Options
Related

Bogus Inspection Program

InspectorJester
Greetings!
So I started working at this company this year that had a CMM since 2013. They paid a chunk to get a couple programs out to us.
They used these programs (at least 3 of them, but there were around 7 in total)
I started working here with little/no experience in a shop. As in no CNC, no CMM, no quality, hell I was barely qualified to sweep the floor!

Well I love to program and literally the second I saw our Optiv 443 I wanted to see what it did.
It sat for almost 4 years as a glorified paper weight!

The individual who was responsible for programming it was our former Quality Manager (who has since moved to Materials manager or something)
She would run the inspection program until the part was good, then give it back to the machinist to say "Good job sir!"
---
I don't like this mentality. I like to program! And I like things to be right, not just "look good" on paper just to find out the customer rejected them again.
So after I got put into quality and got the ball rolling in that department (for about 6 months or more) the CEO of the company asks me if I like my job.

Well, let's say he appreciated my honesty and let me move out of quality and onto the CMM!

So I get some guy who was probably from Hexagon to come down for a day and run through the basics with me (which I already knew from genuine curiosity)
Then ol' Roger Conway comes down and calibrates it and I get more or less free reign from then on.

I can make or break programs, build or destroy!

Anyways, I was looking into the programs we purchased, and I thought they were garbage to be honest.

The vectors weren't perfect, was my biggest issue, along with extra nitpicky stuff.

So I rewrote them and now we have these really nice programs that are neat and concise.

To check a feature that we don't have the tools to verify....
------------
My question is, how can I know if I'm getting good data if there's no way I can verify the CMM is measuring this feature correctly?
It's not necessarily off topic, but we are able to accept these parts from the measurements we get so I moved it here
Parents
  • So today I had to send the parts back to the operator because the CMM measured them out a degree or so. We've shipped these parts hundreds of times, and if the CMM says notch from a set of 8 is bad, they're all probably good (as far as management is concerned) so we ship them. Haven't had them sent back. All I want is these parts to be good, obviously.
    So normally I'll get an issue where I run a part, and one notch is out by say .05°. There are 20 total, so I assume it's good, but run it again for good measure. Usually a different notch is out a little now, but was good at first, but the one that measured bad is now good.

    Well this is the first time I've personally dealt with a whole set being off by almost a whole degree (reminding that tolerance is ­± 0.17°, OutTol was around .85!)
    3 of 4 in a set were bad, and the part was measured twice to verify if they were indeed out. It still says yes.

    Well I don't know how to coordinate with the operator (a machinist? whats that? -.-) to see if the part is good, bad, ugly, what, because the only way we can measure that feature is via CMM, and I personally (due to my lack of in-depth understanding of CMM, metrology, CNC, lathes or mills or whatever he's doing to them) don't know how to either verify that the parts are bad/good, or what.

    Basically I don't know how to tell them the part is bad if I myself can't verify that the part is actually bad, I have to trust my skills as a CMM programmer (and with little formal training on the subject...)
    What would you recommend?
Reply
  • So today I had to send the parts back to the operator because the CMM measured them out a degree or so. We've shipped these parts hundreds of times, and if the CMM says notch from a set of 8 is bad, they're all probably good (as far as management is concerned) so we ship them. Haven't had them sent back. All I want is these parts to be good, obviously.
    So normally I'll get an issue where I run a part, and one notch is out by say .05°. There are 20 total, so I assume it's good, but run it again for good measure. Usually a different notch is out a little now, but was good at first, but the one that measured bad is now good.

    Well this is the first time I've personally dealt with a whole set being off by almost a whole degree (reminding that tolerance is ­± 0.17°, OutTol was around .85!)
    3 of 4 in a set were bad, and the part was measured twice to verify if they were indeed out. It still says yes.

    Well I don't know how to coordinate with the operator (a machinist? whats that? -.-) to see if the part is good, bad, ugly, what, because the only way we can measure that feature is via CMM, and I personally (due to my lack of in-depth understanding of CMM, metrology, CNC, lathes or mills or whatever he's doing to them) don't know how to either verify that the parts are bad/good, or what.

    Basically I don't know how to tell them the part is bad if I myself can't verify that the part is actually bad, I have to trust my skills as a CMM programmer (and with little formal training on the subject...)
    What would you recommend?
Children
No Data
© 2024 Hexagon AB and/or its subsidiaries. Privacy Policy | Cloud Services Agreement