GD&T issue...need help please

[bob mappes]

I have a positional callout that is extremely tight.

fairly large part (for us)
total order will be for around 400 parts
metric
overall size is 515 x 515 x 60
two features with a TP of 0.05 to A,B & C
one is a hole
the other is a slot
no MMC so size tolerance won't help

now, A is the main face, B & C are two side faces
B is called out Perp to A within 0.30
C is called out Perp to A & B within 0.30

It is my understanding that this TP is not valid because the Datums themselves have more possible variance than the TP itself allows. The controlling features have to have as tight or tighter tolerance than the called out feature tolerance.

We have already tried to ask the customer to relax their TP callout and were turned down. They said these were critical features.

We feel that if they are that critical, these features should be used as Datums themselves so everything else would relate to them, but they are not.

I need some better verbage to send to their engineers so that they understand our problem.

If the order was for 10 or 20 parts this would not be an issue, but doing a large quantity this is a problem.

Any help from those GDT gurus out there?

Thanks
bob mappes

[Matthew D. Hoedeman]

Tell them they are full of ****.
That hole/slot combination is the classic B-C locator and since they are the critical features, that doubles the fact that they should be the B-C datum pair.

Asof the perp being a larger tolerance than the TP, perp is NOT a form-location callout, it is a form only, the location of them can be anywhere from 0.01mm from nominal to 100mm from nominal, as long as they are perp, the location makes no never-mind. What I would do, if they will not allow you to change the datums to what they should be (IMO) is to measure 3 planes, top, and 2 sides. Then, intersect the top plane with both side planes, then intersect the 2 lines with each other. Use the top plane for level, the line from the top to b 'side' as the rotate, then the intersection of the 2 lines as the XY origin and teh top as the z origin.


gee, I didn't know that p o o p was a blocked word.

[ag162]

Hello,
It sounds as if the Engineer drew the part and misused the GD&T again. Do you know the application? Does the hole and slot actually control the position of the part? Or is the part held in place by another mating surface, like the edges being welded and the hole and slot holding a separate component? Is this a stamping, CNC Milling or a plastic formed operation? Either way,(in the alignment, after dat A, I would meas only one line, take a point as close to the corner as possible, then do the 3-2-1 (this will help eliviate deviation and error) and try to get the TP in tol rather than attempting to change the customers mind.

AG

[Wes Cisco]

Maybe make test pieces or CAD models of the what the parts could look like if they persist with the current tolarencing. Show them why having such loose controls on the datums makes having a feature controlled tightly to the datum silly. HTH

[bob mappes]

Tell them they are full of ****.
That hole/slot combination is the classic B-C locator and since they are the critical features, that doubles the fact that they should be the B-C datum pair.

snip.


gee, I didn't know that p o o p was a blocked word.

I have picked up the sides as planes but did not do the intersects to create "datum" lines....this I will try today

[ag162]

Why is it, that the design engineers who draw these parts can make a huge mistake, and we have to fight the world to try and get it corrected? Shouldn't the ones designing and assigning the datums and tol be more knowledgable than us?

[bob mappes]

Hello,
It sounds as if the Engineer drew the part and misused the GD&T again. Do you know the application? Does the hole and slot actually control the position of the part? Or is the part held in place by another mating surface, like the edges being welded and the hole and slot holding a separate component? Is this a stamping, CNC Milling or a plastic formed operation? Either way,(in the alignment, after dat A, I would meas only one line, take a point as close to the corner as possible, then do the 3-2-1 (this will help eliviate deviation and error) and try to get the TP in tol rather than attempting to change the customers mind.

AG

this is a solid aluminum block. not sure of the mountings

[Matthew D. Hoedeman]

I have picked up the sides as planes but did not do the intersects to create "datum" lines....this I will try today

The reason I said to use the intersections is because:

EXAMPLE:
If you bend a flat sheet into a box shape, the bend lines should be more stable and constant than the center point of the plane on the side wall, which is what Pcdmis works from if you use the plane itself.

[bob mappes]

Why is it, that the design engineers who draw these parts can make a huge mistake, and we have to fight the world to try and get it corrected? Shouldn't the ones designing and assigning the datums and tol be more knowledgable than us?

I can't tell ya who this is for.......and these guys ARE the leading edge of technology and GDT developement.

at least they were when I took drafting courses 25 years ago....my instructor was had close ties to an engineer who was on the GDT advisory board

hey! what do we know on the mfg end of it all? we just have to carve a block into jewery for em......LOL

[ag162]

Basically, just play with the alignment and you should find a way to get the TP in. This will be a lot better than disputing the customer. If you can set up stops and use them as part of the aligning process, this may help.

AG

[#2#]

Why is it, that the design engineers who draw these parts can make a huge mistake, and we have to fight the world to try and get it corrected? Shouldn't the ones designing and assigning the datums and tol be more knowledgable than us?

What I have found here is that often they design the parts to be easy to machine (cheaper to make). The problem is that when they do this they often times lose sight of which features need to be tightly controlled and which are unimportant. A classic example here is that we have a part which has a bore called out as Datum D. This bore is called out .005|A|B|C|. Around this bore are two dowel holes for aligning a mating part. They are also called out .005|A|B|C| because it is easier to machine them that way (one setup). In reality though, the relation of these holes to |A|B|C| is not critical. What is critical is that they are in the correct location relative to Datum D. In an ideal world, Datum D will be exactly where it is supposed to be and the dowel holes are where they are supposed to be. We know that rarely happens. What really happens is that Datum D may be shifted off to one side .002 in and then maybe the dowel is shifted .002 in the other way. In that case, both of them would meet the .005|A|B|C| requirement but maybe the mating part wouldn't fit. In the end, they (the engineers) make their job harder because they have more tolerance stack up to analyze. They also make our jobs harder because we are left scratching our heads about callouts that seem to make no sense.

[ag162]

And,,,,,the forum is here for help.

ag

[Jan d.]

I have a positional callout that is extremely tight.

fairly large part (for us)
total order will be for around 400 parts
metric
overall size is 515 x 515 x 60
two features with a TP of 0.05 to A,B & C
one is a hole
the other is a slot
no MMC so size tolerance won't help

now, A is the main face, B & C are two side faces
B is called out Perp to A within 0.30
C is called out Perp to A & B within 0.30

It is my understanding that this TP is not valid because the Datums themselves have more possible variance than the TP itself allows. The controlling features have to have as tight or tighter tolerance than the called out feature tolerance.

We have already tried to ask the customer to relax their TP callout and were turned down. They said these were critical features.

We feel that if they are that critical, these features should be used as Datums themselves so everything else would relate to them, but they are not.

I need some better verbage to send to their engineers so that they understand our problem.

If the order was for 10 or 20 parts this would not be an issue, but doing a large quantity this is a problem.

Any help from those GDT gurus out there?

Thanks
bob mappes

I do not see anything wrong with this.

What the engineer thought was that in some sort of OP10, you will establish datums A, B and C. You make some kind of a fixture and machine those 3 features (plus all other unimportant features you can reach from that side). Within the tolerance specified, they do not care where these datums are.

THEN, you move to OP20. Now you make a special fixture that allows the part to sit against A, B, C in accordance with whatever standard they tell you to follow (is it ASME or ISO???). Then you machine the hole and the slot to these datums.

By the way: If you can do all these features in 1 setup, chances are this part will be in spec, no problem. And you will make B and C a lot better than required.

Make sure your machining department processes the part per the standard. The print calls out the plane as a datum, so you will have to properly push the part against a flat. For B, they must make sure they establsih contact against the 2 highest points on B. For C: THE highest point on C. And you MUST measure it also that way: measure the whole B plane and find the highest 2 points. Measure the whole C plane and find the highest point.

If you do NOT process the part per the spec, or do NOT measure it per the spec, you are going to have lots of fun!

By the way, I have seen lots of these types of parts. Again, make sure they machine them according to the standard. And you measure them to the standard. Then all will be easy.


Jan.

[than_357]

Would guess that they're relying on the positional tolerance to control both location and squareness to datum A. Since position is tied into the sides, and this sounds like a block, I'm not sure that the intersection of the sides to the face is going to show exactly where the location of the bores are. I'd go with the sides as planes, datum A as the origin and shoot two TP callouts near the datum A side and near the opposite end of the bores. Clumsy, but you'll cover any wandering of the bores as far as squareness goes.

In this case- you'd have to rely that the faces are actually neat and square to get the location to work out.

Tom H