It Don't Add Up

[DaddyBluto]

I have one that is leaving me stymied. I have an o-ring groove that needs the TSC diameter measured on it. I construct a plane at the bottom of the groove from 4 points and 4 3D lines of 3 points each on the side walls at 90° from each other. Then I pierce the plane with the lines and use those points to construct the circle. No problem really. Today I had an out of tolerance part during a source inspection which means I need to run a 100% sort for that dimension. I took the original program, saved a copy and deleted out anything not involved in the setup or measurement of the circle and ran the same part. I did not touch the part other than with the stylus and I got .0033" smaller diameter (which is in tolerance). I figured I deleted out some necessary code, so I reran the full program stopping it after the setup and jumping to the circle section and got a third value in between the two!!! I've tried to account for every variable I can think of, but I should not have this much of a delta.

B&S One 7-7-5
PC-DMIS CAD 3.7 MR3
TP20
same probe
no re-cal in between
no crash
same speed
same code
same part
clamped in Rayco the whole time
climate controlled lab ± 3°
probe not loose

Have any of you run into something similar? My boss told me to ask my friends, so I came to y'all (can't shake that from AL).

TIA

Jaime

[bob mappes]

Mold it and use a comparator...it's alot easier

This is something I haven't attacked yet, partly because our o-ring grroves are usually small...... about 1/8" at the TSC and a diameter at the same time...would make life interesting if I didn't already have too much other stuff to check

[cmmguy]

Are measuring the lines along the "cylinder" of the o-ring groove? If so, how are you sure that you are exactly on centerline and compensating the probe radius correctly?

[Tested]

One way you may be able to verify your readings is to measure a cone on the surface where you are measuring the lines. Then construct the circle. This should remove any probing error and reduce the Reliance on the alignment at the same time.
As far as the readings changing I'm not sure about that. What is the difference between your good run and your bad run? My guess would be as cmmguy said.

[#2#]

Well, when I run into something like this, I usually try to start from scratch ignoring what my current program does and come up with how I would measure it today. Then I compare that result and method with my previous program and decide which is the better way to measure. I also try to check it with another method (non-cmm) to verify my results.

[rangerboat72]

I have one that is leaving me stymied. I have an o-ring groove that needs the TSC diameter measured on it. I construct a plane at the bottom of the groove from 4 points and 4 3D lines of 3 points each on the side walls at 90° from each other. Then I pierce the plane with the lines and use those points to construct the circle. No problem really. Today I had an out of tolerance part during a source inspection which means I need to run a 100% sort for that dimension. I took the original program, saved a copy and deleted out anything not involved in the setup or measurement of the circle and ran the same part. I did not touch the part other than with the stylus and I got .0033" smaller diameter (which is in tolerance). I figured I deleted out some necessary code, so I reran the full program stopping it after the setup and jumping to the circle section and got a third value in between the two!!! I've tried to account for every variable I can think of, but I should not have this much of a delta.

B&S One 7-7-5
PC-DMIS CAD 3.7 MR3
TP20
same probe
no re-cal in between
no crash
same speed
same code
same part
clamped in Rayco the whole time
climate controlled lab ± 3°
probe not loose

Have any of you run into something similar? My boss told me to ask my friends, so I came to y'all (can't shake that from AL).

TIA

Jaime

Have you found a solution???

[DaddyBluto]

I have no solution yet. Apparently I did not make myself very clear. When I run the full program multiple times removing the part or not I repeat within .0001". When I run the sorting program I repeat within .0001". The two programs to each other differ by .0033" for the same part. I didn't even cut and paste, I just cut out the unused code so all I had left was the setup and features used to make that dimension. I am trying to figure out why the two programs don't match.

[#2#]

I have no solution yet. Apparently I did not make myself very clear. When I run the full program multiple times removing the part or not I repeat within .0001". When I run the sorting program I repeat within .0001". The two programs to each other differ by .0033" for the same part. I didn't even cut and paste, I just cut out the unused code so all I had left was the setup and features used to make that dimension. I am trying to figure out why the two programs don't match.

Did you remove any alignments? Possibly if you removed an alignment, the circle isn't measuring the same as it was. It could maybe be measuring at a different height or maybe everything isn't measuring as 'square' as it was.

[Matthew D. Hoedeman]

Did you remove any alignments? Possibly if you removed an alignment, the circle isn't measuring the same as it was. It could maybe be measuring at a different height or maybe everything isn't measuring as 'square' as it was.

OR, have you lost the location of an external alignment perhaps?

[DaddyBluto]

Nope. All the alignments associated with the groove and the features that make that alignment are there.

[Tested]

Not trying to be smart but is it trying to recall a alignment i.e.
A1 =ALIGNMENT/START,RECALL:A0,LIST=YES
ALIGNMENT/TRANS,XAXIS,CIR1
ALIGNMENT/TRANS,YAXIS,CIR1
ALIGNMENT/TRANS,ZAXIS,CIR1
ALIGNMENT/END
That might have gotten removed?

[Wes Cisco]

I have no solution yet. Apparently I did not make myself very clear. When I run the full program multiple times removing the part or not I repeat within .0001". When I run the sorting program I repeat within .0001". The two programs to each other differ by .0033" for the same part. I didn't even cut and paste, I just cut out the unused code so all I had left was the setup and features used to make that dimension. I am trying to figure out why the two programs don't match.

You have sinned in your heart. Most likely you have been drooling over Zeiss catalogs in the restroom. Repent and praise Hexagon whilst you still can!

In other words, I have no freaking clue, but hopefully I help you laugh a little.

[DaddyBluto]

No. The drool is on the Jessica Alba photos.

And thanks for the laugh.

[DaddyBluto]

Not trying to be smart but is it trying to recall a alignment i.e.
A1 =ALIGNMENT/START,RECALL:A0,LIST=YES
ALIGNMENT/TRANS,XAXIS,CIR1
ALIGNMENT/TRANS,YAXIS,CIR1
ALIGNMENT/TRANS,ZAXIS,CIR1
ALIGNMENT/END
That might have gotten removed?

Nope. Checked that, but don't be afraid to bring up simple stuff. It could easily be some mundane little detail.

[Tested]

Well here's another one... What about a working plane that got deleted?