Help me understand Profile of a Surface

[Jeff]

We have some 7075 aluminum parts that call out "Profile of a Surface .5mm to A,B and C".
With those being the Top, Back and Side datums.
The feature I'm cutting is a straight wall in Y, that's it. 2.25" deep.
Our customer is telling us that it has a deviation of .2mm
How in the hell is this possible?
Using a brand new end mill. Finish cut is .01" then a free pass.

[neurosis]

Our customer is telling us that it has a deviation of .2mm

That's technically + or - .0078? 

[Jeff]

That's technically + or - .0078? 

Yes, their CMM printout is saying that the straight wall we cut has a .0078" deviation. Which completely baffles me as to how.

[Jeff]

CMM's can give all sorts of readings depending how you set the datum planes.
Things like this I like to go old school....run a DTI down the face or with this amount (8thou) of error, put the back end of a square on it and look for the daylight!

The end mill was cutting a .0015" taper top to bottom before we swapped it out.

[JParis]

 sounds more like the issue is on the customer end...my guess is they are either inspecting it wrong or using the wrong values....

[Jeff]

sounds more like the issue is on the customer end...my guess is they are either inspecting it wrong or using the wrong values....

That's what I'm thinking! I just can't see how my wall could possibly have that much deviation when all it's doing is cutting in Y.

[gcode]

Is this surface perpendicular to one of the datums?
Say Datum -A- is the bottom of the part and your rejected face is perpendicular to the bottom
of the part.
They'll hit the bottom of the part  3 or 4  times to establish a plane, then hit your surface
If the bottom of the part is large, or slightly bowed or tapered and they are calling it a
plane, it won't take much to throw your wall out.... even if it's dead flat

[Jeff]

Is this surface perpendicular to one of the datums?
Say Datum -A- is the bottom of the part and your rejected face is perpendicular to the bottom
of the part.
They'll hit the bottom of the part  3 or 4  times to establish a plane, then hit your surface
It the bottom of the part is large, or slightly bowed or tapered and they are calling it a
plane, it won't take much to throw your wall out.... even if it's dead flat

It's perpendicular to the top.
I'm finishing the Top, Back and Side and this wall all in one holding and bolted to the fixture plate.

Good point on that area not being flat, it's about a 29" square top face so that might be the issue if it springs when I unclamp it. I've tried taking every precaution to make sure it doesn't spring though.

Thanks!

[gcode]

It sounds like they've over toleranced this feature.
Despite how it's toleranced on the B/P, if that 29" square face is not flat within
.2mm, your wall feature is going to be out.

[Jeff]

It sounds like they've over toleranced this feature.
Despite how it's toleranced on the B/P, if that 29" square face is not flat within
.2mm, your wall feature is going to be out.

I just found out that this was on one of the first parts that we ran a month and a half ago, so it's possible that it bowed on us. Just checked a part from yesterday and it's flat within .002" so we should be good.
I wish they would have told us sooner that it was one of the first parts especially since the part has seen a major revision since then.

[Zoffen]

When you get a cmm you start to notice things that are weird. Like letting the parts sit overnight they change a bit. or letting an OP100 sit overnight vs running it right away, The part will be change a little bit.

Cutting metal relives alot of stress within the material, as well as within the humans running the parts.

[Here's Johnny!]

CMM = Can't Measure Metal!!!!!

[gcode]

^^^^ this ^^^^
Years ago I was involved in a roughing and finishing an large aluminum hogout
The part started as a sawed plate 8ft x 3 ft x 8" thick. I don't recall what the
finish part weighed but two men could easily pick it up.

There were 6 main operations 2 rough, two semi finish and 2 finish ops using  vacuum fixtures.
The first time we build the parts they sat, sometimes for weeks, between operations, waiting on the next NC file.

The parts came out really sweet, passing all CMM inspections with flying colors.

A year later we ran the same part again.
All programs were done and we ran the parts from operation to operation with no delay.
It fought us every inch of the way and we really struggled to maintain the profile tolerances.
The bosses had all sorts of reasons this happened.
My personal opinion is the roughed stock was stabilizing between operation due to the 2 to 4 week delays
waiting for the next NC files  on the first run.
On the second run there were no delays because all the files were proven and the boss wanted to ship the
parts.

[crazy^millman]

^^^^ this ^^^^
Years ago I was involved in a roughing and finishing an large aluminum hogout
The part started as a sawed plate 8ft x 3 ft x 8" thick. I don't recall what the
finish part weighed but two men could easily pick it up.

There were 6 main operations 2 rough, two semi finish and 2 finish ops using  vacuum fixtures.
The first time we build the parts they sat, sometimes for weeks, between operations, waiting on the next NC file.

The parts came out really sweet, passing all CMM inspections with flying colors.

A year later we ran the same part again.
All programs were done and we ran the parts from operation to operation with no delay.
It fought us every inch of the way and we really struggled to maintain the profile tolerances.
The bosses had all sorts of reasons this happened.
My personal opinion is the roughed stock was stabilizing between operation due to the 2 to 4 week delays
waiting for the next NC files  on the first run.
On the second run there were no delays because all the files were proven and the boss wanted to ship the
parts.

Yes that was exactly the problem. Had this on a turn key years ago. Customer was roughing Big Ti Forgings kind of the same way. 60 days to process the parts they way they were doing it. Wanted it done in 10 hours. I laughed when I was first brought it and said pretty much that exact same thing. I am the crazy person in the room what do I know? Fast Forward 2 years later and guess what every single part went to be assembled on planes and wouldn't fit. They were brought back in for reinspecton and every single one of them moved. They had some sitting on the shelf that were run 3 months prior that passed inspection they were also out of tolerance. I tried warning everyone parts had normalization with the 8 operations they were taking reducing them down to 3 might seem like the best way, but cannot say it is. Customer lost the contract so guess that answers who was the crazy one after all.

[CNCAppsJames]

Something else to consider is the size of the probe styli they are using. Say for example you have a 64 surface finish callout and you held it to a 32. Now if Thee CMM guy is using a 2mm stylus and you hit a peak and a valley, that could easily throw out out 5mm in a worse case scenario.

I worked on a part one time that had a .02mm profile tolerance and the surface finish callnout was a 32. The CMM inspector was using a .5mm ruby and refused to change it. In order for the part to CMM good I had to get the finish down to between 6 and 8. Even then I was using 90% of the tolerance band. Sucked. That inspector was garbage.