true position datums 2

[Bob Pierce]

This is related to a previous post which I am not sure I received responses in total agreement.
When a drawing dimensions to feature control frames, does the the probing of the features need to be made in the alignment defined in the FCF?

Matthew ?? CMM Guy ?? Can I get you into the discussion?

Help
and thanks

[Wes Cisco]

This is related to a previous post which I am not sure I received responses in total agreement.
When a drawing dimensions to feature control frames, does the the probing of the features need to be made in the alignment defined in the FCF?

Matthew ?? CMM Guy ?? Can I get you into the discussion?

Help
and thanks

I am not Matt or J, but here is my .02. You do not have to be aligned to FCF to probe the feature, but you need to be aligned to the FCF when you Dimension (Trueposition) the feature. HTH

[JamesMannes]

I am not Matt or J, but here is my .02. You do not have to be aligned to FCF to probe the feature, but you need to be aligned to the FCF when you Dimension (Trueposition) the feature. HTH

I am not Matt, J, or Wes, but I would agree with Wes.

[Matthew D. Hoedeman]

I'm not me, but I also agree.

The dimension HAS to be to the FCF, you can measure it in any datum scheme you can.

[#2#]

I'm not me, but I also agree.

The dimension HAS to be to the FCF, you can measure it in any datum scheme you can.

One quick addendum to Matt's post although I am none of the people who posted above...

You can measure any feature in any alignment you like. You just need to be carefull if you are measuring 2d features. Imagine you are measuring a circle in a bore that isn't quite square to a plane. Measure the bore as a cylinder and then align to that cylinder. Now, measure a circle with a vector parallel to the cylinder (0,0,1). The circle will look like a circle and may have very good roundness. Now, align to the plane and measure the circle with the circle's vector perp. to the plane (0,0,1). Your circle now looks like an ellipse and the roundness won't be very good.

[cmmguy]

I believe that you will get better part/measurement repeatability if you measure the features AFTER you measure and establish the datum features. While you are "in" that datum scheme.

[craiger_ny]

I am Batman

[Bob Pierce]

I would like to thank all of the good folks who responded to the post.
I only called out names ( wont do that again !!! ) because on the previous post they had offered differing answers
They still do and I respect that.

Part alignment to each datum prior to probing the features seems challenging, especially when using CAD + Part.

Hmmm.

[Wes Cisco]

I would like to thank all of the good folks who responded to the post.
I only called out names ( wont do that again !!! ) because on the previous post they had offered differing answers
They still do and I respect that.

Part alignment to each datum prior to probing the features seems challenging, especially when using CAD + Part.

Hmmm.

This is how I do almost everything. Many of my parts are very complex; some have 20 Datums and hundreds of features. It is not uncommon for one program to have 30+ alignments, and that is not counting recalls. I measure the datums in the FCF, align, measure the features associated with that datum structure, dimension, then on to the next datum structure.
I do not get CAD often, but I have on occasion written entire programs from CAD this way before the parts ever get to me. This is probably a more time consuming method than what Matt uses,(measure all, then create alignments and dims at the end), but with the parts I have, it is the best way to keep it all straight. I agree with J, (cmmguy), that this is also probably the most repeatable way. HTH

[bob mappes]

I believe that you were trying to also ask about the "order" of the controlling features (datums)

say if the top plane is -A-
front edge is -B-
right edge ic -C-
and your callout is TP to A,B,C
then that is the order of the "controlling features"....A,B,C
but if your callout is B,A,C or something different, then the controlling features are weighted differently

unfortunately, the persons responsible for the drawings don't always know how to apply GDT.
So, you may have a hole on the top face that is called out to A,B,C....and then have a hole on the side of the same part still called out to A,B,C, when it probably should have been called out something like B,A,C

it really depends on the functionality of the features
HTH
bob

[#2#]

This is how I do almost everything. Many of my parts are very complex; some have 20 Datums and hundreds of features. It is not uncommon for one program to have 30+ alignments, and that is not counting recalls. I measure the datums in the FCF, align, measure the features associated with that datum structure, dimension, then on to the next datum structure.
I do not get CAD often, but I have on occasion written entire programs from CAD this way before the parts ever get to me. This is probably a more time consuming method than what Matt uses,(measure all, then create alignments and dims at the end), but with the parts I have, it is the best way to keep it all straight. I agree with J, (cmmguy), that this is also probably the most repeatable way. HTH

I don't know about repeatability but... I do it the other way. I don't think either way is necessarily right or wrong. I do it the other way because often times I have some datums on one side of the part and some datums on the other side of the part. If I were to measure all datums first and then align to them and measure the rest of the features, I would have to start out on one side of the part, measure a datum. Move to the other side of the part, rotate the head and measure another datum. Sometimes, I would then have to move to another side of the part, rotate the head and measure another datum. Now, I would align and then have to re-visit each side of the part to measure the features on each side. Lots of time wasted on moves around the part and if you ask me it increases the likelyhood of a crash.

[Cumofo]

Part alignment to each datum prior to probing the features seems challenging, especially when using CAD + Part.

Hmmm.

Once the the part is aligned you can do any alignment you want, because they all relate to the previous alignment. Measuring in the FCF Datum set is a good idea, but in version 4.2 you don't have to any more.

(At least that's what James Mannes said.)


(Just Kidding JAmes!)

G

[John Kingston]

I've done both. Depends on the complexity of the part and the results. If I feel that the CMM result is questionable, I double check that feature on a surface plate. I truly get a kick out of blueing and scribing parts. (OK, I'm OLD school). Usually that helps me to reinforce the "discussion" that ensues when I reject things.



JAMES MANNES I FART IN YOUR GENERAL DIRECTION!


I might have a new signature!

[JamesMannes]

I've done both. Depends on the complexity of the part and the results. If I feel that the CMM result is questionable, I double check that feature on a old mattress. I truly get a kick out of blueing and scribing parts. (OK, I'm OLD school). Usually that helps me to reinforce the "discussion" that ensues when I reject things.

I am now a "demi-guru", each post must have my name in it somewhere. Do not post again without my name in the post.

[Wes Cisco]

I am now a "demi-guru", each post must have my name in it somewhere. Do not post again without my name in the post.

Hahahahahahhahahahahahahahhahahhahahhahahahahaha!