Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations cowski on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
Tapered Features as Datum Plane 3
- Thread starter Tarator
- Start date
CheckerHater
Mechanical
Well, I see few possibilities:
1. Use of datum targets. Straightforward, fixture-oriented approach; you just show how exactly to grab the part.
2. Use multiple datum features. Your centerplane is derived from two flats; each of them if legitimate datum feature. Mark your slanted planes [A] and , and identify your datum as [A-B] – common axis/centerplane derived from 2 features.
3. Try to sneak in Y14.5-2009 approach, and use the idea of complex datum feature; after all new standard is merely “a revision” of 1994, it doesn’t contradict it in every way. Then your case will be pretty much similar to Chapter 4.4 paragraph (g) (also see Fig. 4-3 (g))
I started a thread myself long ago (before 2009) and we didn’t really come to the conclusion how to produce axis from square shaft:
thread1103-278412
Good luck!
1. Use of datum targets. Straightforward, fixture-oriented approach; you just show how exactly to grab the part.
2. Use multiple datum features. Your centerplane is derived from two flats; each of them if legitimate datum feature. Mark your slanted planes [A] and , and identify your datum as [A-B] – common axis/centerplane derived from 2 features.
3. Try to sneak in Y14.5-2009 approach, and use the idea of complex datum feature; after all new standard is merely “a revision” of 1994, it doesn’t contradict it in every way. Then your case will be pretty much similar to Chapter 4.4 paragraph (g) (also see Fig. 4-3 (g))
I started a thread myself long ago (before 2009) and we didn’t really come to the conclusion how to produce axis from square shaft:
thread1103-278412
Good luck!
- Thread starter
- #3
-
1
- #4
CheckerHater
Mechanical
Just make sure your intent is clear from the drawing.
- Thread starter
- #7
CheckerHater
Mechanical
Using Profile you can turn practically anything into “complex datum feature”. This is described in finer detail in 2009, but isn’t getting enough attention in 1994.
So the question really is, how strong is your commitment to stick with older definitions?
So the question really is, how strong is your commitment to stick with older definitions?
-
1
- #9
pmarc
Mechanical
- Sep 2, 2008
- 3,242
CH,
If the angle between datum features A nd B is not basic, there is no way to properly simulate angle between datum feature simulators A and B during inspection. I do not think there is any difference between '94 and '09 in this aspect, even though this seems not to be explicitly stated in '94 std.
Tarator,
Yes, Profile of a Surface is good choice. To clarify one thing, I did not recommend Angularity, I just said that angular relationship between surfaces A and B must be controlled. In fact, I should have probably said that not only angular, but also locational relationship between surfaces A and B, must be controlled. And for that purpose, Profile of a Surface is good solution.
As for your sketch, I would probably try to avoid specifying profile to A-B on datum features, although this is geometrically interpretable. Simple profile with no datum feature references, with additonal notation '2 SURFACES' beneath profile FCF, would probably do the job better. The underlying principle behind this is quite similar to what is shown in fig. 6-21 of Y14.5M-1994. It is just that in 6-21 the surfaces A and B are nominally coplanar (basic 0 linear and angular relationship between them, thus the basic dimensions are not shown) - in your example you need to show some basic dimensions due to more complex configuration of surfaces.
If the angle between datum features A nd B is not basic, there is no way to properly simulate angle between datum feature simulators A and B during inspection. I do not think there is any difference between '94 and '09 in this aspect, even though this seems not to be explicitly stated in '94 std.
Tarator,
Yes, Profile of a Surface is good choice. To clarify one thing, I did not recommend Angularity, I just said that angular relationship between surfaces A and B must be controlled. In fact, I should have probably said that not only angular, but also locational relationship between surfaces A and B, must be controlled. And for that purpose, Profile of a Surface is good solution.
As for your sketch, I would probably try to avoid specifying profile to A-B on datum features, although this is geometrically interpretable. Simple profile with no datum feature references, with additonal notation '2 SURFACES' beneath profile FCF, would probably do the job better. The underlying principle behind this is quite similar to what is shown in fig. 6-21 of Y14.5M-1994. It is just that in 6-21 the surfaces A and B are nominally coplanar (basic 0 linear and angular relationship between them, thus the basic dimensions are not shown) - in your example you need to show some basic dimensions due to more complex configuration of surfaces.
CheckerHater
Mechanical
pmarc,
middle plane can be derived from any 2 flat surfaces regardless of angle between them. Did I miss the memo?
middle plane can be derived from any 2 flat surfaces regardless of angle between them. Did I miss the memo?
-
1
- #11
pmarc
Mechanical
- Sep 2, 2008
- 3,242
That is true.
But in that case I would rather go with something like this:
I would not use two separate datum feature letters, just like I would not do it for classic planar FOS (distance between two nominally parallel opposed surfaces).
Some open points would remain though:
1. No recipe in the standard on how to universally interpret this kind of callout (or the A-B callout without basic angle).
2. Is directly toleranced (non-basic) angle unambiguous way to control form of a wedge (or a cone) in the light of the standard?
But in that case I would rather go with something like this:
I would not use two separate datum feature letters, just like I would not do it for classic planar FOS (distance between two nominally parallel opposed surfaces).
Some open points would remain though:
1. No recipe in the standard on how to universally interpret this kind of callout (or the A-B callout without basic angle).
2. Is directly toleranced (non-basic) angle unambiguous way to control form of a wedge (or a cone) in the light of the standard?
CheckerHater
Mechanical
Makes sense.
It’ just OP is very vague about functional relationship.
It is not clear if wedge is used in the manner similar to countersink / machine center or it’s something more sinister
This is why I went with 2 separate features that will form combined datum “regardless of the angle”.
Now about interpretation.
What you can possibly derive from 2random planes? They will intersect forming one and only one line (axis if you wish). You can draw 2 mutually perpendicular mid-planes that can be used to form datum framework (leaving some degree of freedom). That’s it. I would really like to see other interpretation.
You have no problem deriving axis or plane from machine centers, which are far more complicated features than flats, have you?
It’ just OP is very vague about functional relationship.
It is not clear if wedge is used in the manner similar to countersink / machine center or it’s something more sinister
This is why I went with 2 separate features that will form combined datum “regardless of the angle”.
Now about interpretation.
What you can possibly derive from 2random planes? They will intersect forming one and only one line (axis if you wish). You can draw 2 mutually perpendicular mid-planes that can be used to form datum framework (leaving some degree of freedom). That’s it. I would really like to see other interpretation.
You have no problem deriving axis or plane from machine centers, which are far more complicated features than flats, have you?
pmarc
Mechanical
- Sep 2, 2008
- 3,242
CH said:
So here is some kind of other interpretation.
What if these 2 random planes derived from actual surfaces of the wedge are inclined to each other in more than just a single plane of coordinate system? This is quite possible, don't you think? This will result in a composite angle, so the intersecting line you mentioned will indeed be one and only, but if you draw 2 mutually perpendicular planes of datum reference framework intersecting at that line and start rotating them, you will never find a situation where at least one of these 2 mutually perpendicular planes perfectly coincides with any of the planes derived from the faces of the wedge. In some extreme cases of wedge form error you may even have all 6 degrees of freedom constrained - using this single datum feature only.
Now, think about repeatability of any measurements done from datum reference frame established that way. Take your part and Tarator's approach to tolerancing it (with the exception that the basic angle between wedge surfaces has been replaced by direcly toleranced angle and the datum feature symbol, say A, was placed like B or C in Tec-Ease tip). Can you say that locations of all remaining profile tolerance zones (now referencing to A only) are unambiguously defined (fixed) in space relative to the datum reference frame regardless of which as-produced part is currently mounted in a datum feature simulator?
Machine centers as datum features... Hmm... I will start thinking about it as soon as I find a functional need to use them as datum features ;-)CH said:
CheckerHater
Mechanical
I will not "rotate" planes. There is one and only one mid-plane for any actual angle.
About machine centers - so you DO have problem with them. Believe me, in real world they are used far more often than in imaginary one.
I was always curious, how they measure total runout on features being completely embraced by chuck / collet like on Fig. 4-25 in 2009. Seriously - can someone provide a picture?
About machine centers - so you DO have problem with them. Believe me, in real world they are used far more often than in imaginary one.
I was always curious, how they measure total runout on features being completely embraced by chuck / collet like on Fig. 4-25 in 2009. Seriously - can someone provide a picture?
pmarc
Mechanical
- Sep 2, 2008
- 3,242
Looks like my reply was not clear enough, so I will try differently.
I will again use the picture you drawn. If both actual surfaces of the wedge are perfecly flat AND PERFECTLY PERPENDICULAR TO THE PLANE OF THE SCREEN, the two datum planes A and B derived from these surfaces will be seen as two lines in the view shown. The mid-plane will also be seen as a line. And when you HYPOTHETICALLY rotate this midplane in any direction around intersection point, the mid-plane will eventually coincide with one of the datum planes.
But what if both datum feature surfaces ARE NOT PERFECTLY PERPENDICULAR TO THE PLANE OF THE SCREEN and these "perpendicularity" deviations are not equal? How will the two datum planes A and B look in this view then? How will the mid-plane look like? For sure these three planes won't be seen as lines in the view shown, right? In this case if you HYPOTHETICALLY rotate the mid-plane around an intersection line, it will never coincide with datum plane A or B. (This is what I have been trying to say by bringing this whole "rotation" idea up)
So depending on the actual geometry of the wedge the mid-plane, and in consequence the origin of measurement for other profile tolerances, changes. In my opinion this shall not happen if we want to obtain repeatable measurements - the tolerance zones shall be fixed in orientation and location relative to datum reference frame regardless of actual geometry of inspected part.
This is how I see the A-B scenario (without basic angle). The Tec-Ease approach probably makes things easier in terms of datum center plane (mid-plane) establishment.
As for machine centers, specifying them as datum features on the drawing is something totally different than using them as substitute datum features for inspection. Of course I agree - machine centers are used in real world (total runout on features being completely embraced by chuck/collet may be one of examples), and I really have no problems with that as long as design and inspection is fully aware that verification relative to substitute datum(s) and verification relative to drawing specified datums is not the same thing, and that it introduces some additional uncertainty to the measurement system, which sometimes may, but does not have to, have influence on correctness of assessment whether part is functional or not.
I will again use the picture you drawn. If both actual surfaces of the wedge are perfecly flat AND PERFECTLY PERPENDICULAR TO THE PLANE OF THE SCREEN, the two datum planes A and B derived from these surfaces will be seen as two lines in the view shown. The mid-plane will also be seen as a line. And when you HYPOTHETICALLY rotate this midplane in any direction around intersection point, the mid-plane will eventually coincide with one of the datum planes.
But what if both datum feature surfaces ARE NOT PERFECTLY PERPENDICULAR TO THE PLANE OF THE SCREEN and these "perpendicularity" deviations are not equal? How will the two datum planes A and B look in this view then? How will the mid-plane look like? For sure these three planes won't be seen as lines in the view shown, right? In this case if you HYPOTHETICALLY rotate the mid-plane around an intersection line, it will never coincide with datum plane A or B. (This is what I have been trying to say by bringing this whole "rotation" idea up)
So depending on the actual geometry of the wedge the mid-plane, and in consequence the origin of measurement for other profile tolerances, changes. In my opinion this shall not happen if we want to obtain repeatable measurements - the tolerance zones shall be fixed in orientation and location relative to datum reference frame regardless of actual geometry of inspected part.
This is how I see the A-B scenario (without basic angle). The Tec-Ease approach probably makes things easier in terms of datum center plane (mid-plane) establishment.
As for machine centers, specifying them as datum features on the drawing is something totally different than using them as substitute datum features for inspection. Of course I agree - machine centers are used in real world (total runout on features being completely embraced by chuck/collet may be one of examples), and I really have no problems with that as long as design and inspection is fully aware that verification relative to substitute datum(s) and verification relative to drawing specified datums is not the same thing, and that it introduces some additional uncertainty to the measurement system, which sometimes may, but does not have to, have influence on correctness of assessment whether part is functional or not.
CheckerHater
Mechanical
I still cannot grasp a concept of "mid-plane" being somewhere else except "middle". Looks like time to wrap it up.
To me it looks like matter of fixturing: basic angle will imply solid "functional" gage while toleranced angle will require adjustable self-centering devise similar to what's used to establish feature RFS.
Since OP's picture is "intentionally incomplete" and we don't know the real function of the part, the argument can go indefinitely.
This is why my first choice was using datum targets - just show clearly, how you want to grab the part.
On the matter of machine centers - I wouldn't say they are limited to substitute datums role.
Imagine that you want to control every feature of a shaft in relation to one single common axis, say, for the sake of balancing. Balance is functional requirement, isn't it? How do you establish the axis for the entire shaft? Surely not by using the whole shaft as a datum. Machine centers will create axis going from one end to the other, so all the features can be controlled. Yes, they will dictate how you hold the part, but I always said the function should go "hand-in-hand" with machining and QC.
And pmarc, I have to say that arguing with you usually helps me to see different point of view on the same problem, even if we rarely come to complete agreement.
To me it looks like matter of fixturing: basic angle will imply solid "functional" gage while toleranced angle will require adjustable self-centering devise similar to what's used to establish feature RFS.
Since OP's picture is "intentionally incomplete" and we don't know the real function of the part, the argument can go indefinitely.
This is why my first choice was using datum targets - just show clearly, how you want to grab the part.
On the matter of machine centers - I wouldn't say they are limited to substitute datums role.
Imagine that you want to control every feature of a shaft in relation to one single common axis, say, for the sake of balancing. Balance is functional requirement, isn't it? How do you establish the axis for the entire shaft? Surely not by using the whole shaft as a datum. Machine centers will create axis going from one end to the other, so all the features can be controlled. Yes, they will dictate how you hold the part, but I always said the function should go "hand-in-hand" with machining and QC.
And pmarc, I have to say that arguing with you usually helps me to see different point of view on the same problem, even if we rarely come to complete agreement.
pmarc
Mechanical
- Sep 2, 2008
- 3,242
Okay, so your first choice was using datum targets. Wouldn't it be much more reasonable then to apply basic dimension(s) between wedge surfaces to clearly communicate the relationship between equipment simulating datum targets?
And I did not say machine centers were limited to substitute datum features role. Using your shaft balance example, if these machine centers are truly functional datum features (in other words, if they are the features establishing axis of rotation in real assembly), go ahead and use them. This is what I call "verification of part based on how it functions". But if it is otherwise, for example if axis of rotation is in reality established by two cylinders on both ends of the shaft, verification relative to the axis established by machine centers won't give you 100% clear picture of real shaft unbalance (unless each machine center is perfectly centered relative to the axis of its corresponding cylinder, which is quite unlikely). If you can live with this uncertainty, stay with machine centers. But if there is any trace of suspicion that this additional "noise" may lead to serious functionality loss, using true datum features during balance check is the only solution, in my opinion.
And I did not say machine centers were limited to substitute datum features role. Using your shaft balance example, if these machine centers are truly functional datum features (in other words, if they are the features establishing axis of rotation in real assembly), go ahead and use them. This is what I call "verification of part based on how it functions". But if it is otherwise, for example if axis of rotation is in reality established by two cylinders on both ends of the shaft, verification relative to the axis established by machine centers won't give you 100% clear picture of real shaft unbalance (unless each machine center is perfectly centered relative to the axis of its corresponding cylinder, which is quite unlikely). If you can live with this uncertainty, stay with machine centers. But if there is any trace of suspicion that this additional "noise" may lead to serious functionality loss, using true datum features during balance check is the only solution, in my opinion.
- Status
Similar threads
- Question
- Locked
- Solved