Using a complex blade surface as datum feature 1

[Yuyu28]

Hello all,

Imagine a carbon-fiber rotor blade that is manufactured by laminating the layers on an inner mold, and cured inside an outer mold that will determine the outer, complex aerodynamic surface. The blade is then trimmed at the root and at the tip, forming two planes that, theoretically, are perpendicular to the span direction. Then a pattern of holes is drilled; each hole's axis is normal to the outer aerodynamic surface, so each one has a different direction.

What would be a correct way acc. to ASME Y.14-5 (2009 or 2018) to determine the position and the form of the trimming planes and the hole pattern? In principle, the 3D CAD model is leading, but my company still wants tolerancing on 2D drawings.

I thought of defining the whole outer surface (excluding the root and tip planes) as a datum feature A. According to Fig. 4-3g (2009), that should restrain the 6 DOF; am I correct?

Then the root planes could be defined with a profile tolerance without any reference, their theoretically exact positions given by the 3D CAD model.

But the pattern of holes is more problematic. Their axes are not parallel to the root/tip planes. The only way to unequivocally define their position is in the 3D CAD model. Can I just give their positional tolerance referenced to datum A (meaning that the holes are normal to the surface), without any TED? Would I need to define a coordinate system and use it to position the holes with TEDs, as in Fig. 4-28 (2009)?

By the way, in Fig. 4-28, what does the lower profile tolerance in detail A point to? It looks like a very thin edge, but I have no idea what it is exactly.

 

[Dean Watts]

3DDave,
Of course the part is being restrained as specified in Fig 4-42. The point being made, that your responses are not addressing, is how to best identify the degrees of freedom being constrained by each datum feature. The least clear, but still valid, way to specify the datum targets for this figure would be to specify the targets as areas A1-A9, so there would only be one datum feature, but with the areas as specified, and with the restraint procedure all would work OK. A step towards a more clear picture of which datum targets restrain what degrees of freedom is as the figure is currently shown, with B1 and B2 specified on the lest side of the hood, and C1 specified at the front. Since The A1-A6 targets are not applied to a planar surface, but the desire is clearly that they restrain just 2 rotations and 1 translation, as if datum feature A were a planar surface, something is being left unsaid in the current figure. A more clear description of which datum features are constraining what degrees of freedom would be to use a customized datum reference frame, explicitly specifying that A constrains rotations about X and Y, and translation in Z (even though it appears to be capable of also constraining translation in X), and B constrains rotation about Z and translation in Y, and C constrains translation in X.

Drawoh and 3DDave.
If the part doesn't include much more than an aero surface and mounting features, one will be controlled relative to the other, so the best choice is to choose the feature that provides the more repeatable datum reference frame alignment. That will definitely be the much larger aero surface, so this would be my choice, especially if datum target points such as those shown in Y14.5-2009 Fig 4-54 or Y15.5-2018 Fig 7-53 can be use to model function adequately. Profile of a surface would still be applied to the aero surface.

chez311,
While the 2018 version introduces a new notation shown in Fig 7-64 when datum feature symbols are applied along with datum targets, I don't think the past practice was so bad. So I agree with Burunduk about the practice in general, in fact I go beyond to the point of disagreeing with him. I think it's fine to include the datum feature symbol even it is attached to a feature control frame. I'm not aware of any requirement that the datum feature symbol be applied in a particular way if datum targets area also specified.

Yuyu28 and Burunduk,
Yes, I agree that Fig 4-28/7-29 appears to be cylindrical rather than spherical. I don't think this really matters though, since I believe the intent, whether it is cylindrical or spherical or any other shape is to have datum feature A will behave like a planar datum feature.

Yuyu28,
If your company uses ISO standards, I agree that a customized datum reference frame is not an option. You can always do as in Y14.5 Fig 4-28/7-29 and Fig 4-42/7-49 (& 7-50), and wave your hands to make the complex surface constrain the same degrees of freedom as a planar surface. Of course better than waving of hands would be a note to specify this .

Dean

http://www.validate-3d.com

 

[3DDave]

Using the aero surface to set the mounting surface to set the orientation of the aero surface means that any error is doubled; a one degree twist error in setting up the inspection of the mounting surface is a two degree installed error; a tiny deflection at the root will produce a large deflection at the tip. That is a really bad idea. If the mounting provision is so poor that it cannot repeatably be used to guide the rest of the blade then maybe making it easier for the QC people is not the biggest problem. to solve. I'm guessing you have never been involved in aircraft propellers, but they are always defined relative to the mounting surfaces to prevent error accumulation that is being developed here.

You use VSA - crank out an example. It will look much like the discussion of Fig 2-5, '2009.

 

[Dean Watts]

3DDave,

What you are describing is not anything I proposed. Using a more repeatable option, when there are only two sets of features on the part, is only sensible. One set of features is going to be controlled relative to the other. Using the more capable, and therefore repeatable, set of features to set up the datum reference frame, then controlling the others relative to that datum reference frame will work fine. If those more capable and repeatable datum features are too complex though, as they might be in this case, the other set of features may be a better choice.

Dean

http://www.validate-3d.com

 

[Yuyu28]

chez 311,

Thank you for you clear explanations. I will discuss them with the reviewer when the time comes.

For now, giving up on using the aerodynamic surface as a datum feature, suppose I define three datum targets A1, A2, A3 on that surface. They would determine a datum plane A. Datum plane B would be defined by the trimmed root plane. How could I define the theoretically exact position of the pattern of holes in that case?

All the axes of the holes are parallel to datum plane B, so they could be positioned with a TED to it, but each hole has a different angle w.r.t. datum plane A. Even if I somehow indicated the angle of each hole to datum plane A, I would still need to constrain another DOF for each hole, but I don't have another datum feature that I can use for that.

To clarify the "in theory" part about the standards: in the drawings, ISO 1101 is specified. However, one of the problems with GD&T using ISO is that the standard is not "self-contained"; it references several other ISO standards like ISO 5458 for details on positional tolerancing or ISO 5459 for datum definition. I would ask them to buy them, but I suspect the financial situation will not allow it. Thus, for now I have to work with what I have access to, i.e., ASME Y14.5 and ISO 1101.

 

[Dean Watts]

Yuyu28,

Is the trimmed root planar surface large enough to constrain the final rotational degree of freedom in a reliable/repeatable way? You may need to use two widely spaced datum targets as B1 and B2, maybe on the aero surface.

Dean

http://www.validate-3d.com

 

[3DDave]

One cannot borrow bits and pieces. Either get the entire rule book that applies or just write up a procedure to follow to make the part you wish. The only reason standards exist is to provide a common basis - if you don't have that basis, you cannot use the standard.

Curiously - you and your company seem willing to spend large amounts of money on carbon fiber that could be entirely wasted because of an unwillingness to spend a few dollars on a standard? I'd quit that place.

 

[chez311]

For now, giving up on using the aerodynamic surface as a datum feature, suppose I define three datum targets A1, A2, A3 on that surface. They would determine a datum plane A. Datum plane B would be defined by the trimmed root plane. How could I define the theoretically exact position of the pattern of holes in that case?

Well that doesn't sound like giving up on it**, but its not typically clear what kind of datums are derived from several out of plane datum targets A1, A2, A3 (though I can say it likely isn't just a single datum plane). Its not even really typically that important, especially with datum targets - you define your theoretically exact location/orientation of the feature(s) of interest with basic dimensions to each other and to your datum features (and your targets relative to each other if utilizing targets) and ensure your DOF are constrained properly by your specification of datum features/datum targets. These datum features do not have to be (and in many cases - should not be) the surface(s) (or datum targets coincident with the surface(s)) from which the pattern of holes originates.

Edit: I should add that I fully agree with 3DDave, ask them what the cost is of a thousand or so dollars in standard purchases vs. potentially hundreds of thousands (or millions - depending on size and quantity) in improperly specified or improperly produced/measured parts?

**Edit2: I just realized by "that surface" you might have been referring to one of the mounting surfaces I suggested, I had assumed you were still talking about the aerodynamic surface as you mentioned the trimmed root plane, could you clarify?

 

[Yuyu28]

I am just seeing the posts that were written while I was typing my latest one. I will probably have to come back tomorrow to answer all of you.

Yuyu28, you mentioned that the 2D drawing has secondary significance in your company and that the model is leading for product definition. Does the model include tolerances?

Unfortunately, no, they haven't implemented that yet.

Is the trimmed root planar surface large enough to constrain the final rotational degree of freedom in a reliable/repeatable way? You may need to use two widely spaced datum targets as B1 and B2, maybe on the aero surface.

Which rotational DOF do you mean, for the aero surface or for the holes?

Regarding purchasing the standards, I agree with 3DDave and chez311, I think they should have purchased them since the beginning. Maybe they actually will if I ask, but I'm surprised no other engineer has raised this up before.

 

[Burunduk]

Yes, I agree that Fig 4-28/7-29 appears to be cylindrical rather than spherical. I don't think this really matters though, since I believe the intent, whether it is cylindrical or spherical or any other shape is to have datum feature A will behave like a planar datum feature.

Then I no longer understand your previous statement on this:

In 4-28/7-29 datum feature A could be spherical, so a customized DRF may not be needed.

I may be missing something here, but if it could be spherical or cylindrical and the intent is that it acts as a planar primary datum feature, shouldn't the degrees of freedom constrained by it be clarified by a customized datum reference frame, here too?

 

[Dean Watts]

Yuyu28,

Your primary datum, established using datum targets A1, A2, and A3, will constrain 2 rotations and one translation. The trimmed root planar surface is assumed to be nominally perpendicular to the datum plane established using A1-A3. It will constrain the final (3rd) degree of rotational freedom, and also one translation. My concern is that it may not be a sufficiently large surface to constrain that rotation in a repeatable way.

Dean

http://www.validate-3d.com

 

[Dean Watts]

Burunduk,

Quote (Dean)
Yes, I agree that Fig 4-28/7-29 appears to be cylindrical rather than spherical. I don't think this really matters though, since I believe the intent, whether it is cylindrical or spherical or any other shape is to have datum feature A will behave like a planar datum feature.

Then I no longer understand your previous statement on this:

Quote (Dean, 7 Jul 20 14:09 )
In 4-28/7-29 datum feature A could be spherical, so a customized DRF may not be needed.

I may be missing something here, but if it could be spherical or cylindrical and the intent is that it acts as a planar primary datum feature, shouldn't the degrees of freedom constrained by it be clarified by a customized datum reference frame, here too?

Yes, I should have said that differently. It does not matter what the shape of the non-planar datum feature A is. It clearly appears that the intent in both of the figures is that A provide the same constraints as a planar datum feature, so a customized DRF would make sense for both figures.

Dean

http://www.validate-3d.com

 

[Burunduk]

Dean,
Thanks. That clarifies a lot. I thought that you were seeing some essential difference (in terms of DOF constraints by the primary non-planar datum feature) between this figure and the hood figure, as well as the aerodynamic surface datum feature case.

Yuyu28, so what you are basically saying is that the "leading" CAD model is not MBD and it doesn't contain any information on the allowed variation of features. Your 2D drawings do contain this information yet they are anyway considered less significant than the CAD model and are only being made because the "company still wants tolerancing on 2D drawings" as you mentioned. In addition, you are not provided with the complete set of standards you need to properly specify those allowed variations.

Truth is, it doesn't sound like the 3D definition "leads" over 2D, it sounds more like the work method in your company is simply being careless towards tolerances and just count on the accuracy of the manufacturing process to produce the parts close enough to whatever is in the CAD model regardless any hierarchy between the features. It may work for a while but it is probably a bad approach in the long run. 3DDave says he'd quit this place. Another option is to talk to the relevant people and ask them to buy the documents which you need for your work. Normally these kinds of initiatives are welcomed. You don't have to wait for the other engineers to raise this up and it's actually not that surprising that they didn't so far.

 

[3DDave]

Were I to build a tower I would probably say "When the tower is mounted on the base the tower should be straight up and not leaning." Rather than "Trim the base and hope for the best, like in Pisa." Calling for the largest feature as the primary datum is a mantra that gets parts through inspection that perform poorly on installation when that large feature is not used to locate and orient the part as the primary control.

Burunduk - he already said that money was tight and they were "using" ISO standards but chose not to buy them. This isn't an engineer who notices that standards are a good thing to use in a company that had no idea. This is a company driven by willful ignorance and cost savings.

Additionally, most companies depend on the manufacturing to be far better than called for tolerances. The rare times this has not been true it's been found by a huge number of calls from manufacturing that parts don't fit together.

It's not as if they found some tolerance problem and hired a D&T professional to work through it - they have a guy with little experience and no documents not working with his manufacturing group to resolve a difficult problem.

That tells me he needs to go to a company that knows what they are doing where he can get the support and experience he needs.

 

[Burunduk]

3DDave,

Maybe they actually will if I ask, but I'm surprised no other engineer has raised this up before.

Sounds like it's worth a shot.

 

[3DDave]

That's because the other engineers were turned down and quit.

Notice this: It didn't bother this guy, the other engineers, the manufacturing staff, the inspectors, or the Quality control department. That's between 4 and 40 people in a typical company having no access to the commonly used standards normally relied on. So, sure. Worth a shot. So is picking numbers in a state lottery.

 

[Yuyu28]

Your primary datum, established using datum targets A1, A2, and A3, will constrain 2 rotations and one translation. The trimmed root planar surface is assumed to be nominally perpendicular to the datum plane established using A1-A3. It will constrain the final (3rd) degree of rotational freedom, and also one translation. My concern is that it may not be a sufficiently large surface to constrain that rotation in a repeatable way.

I understand. I think it should be large enough for that purpose, but I'll keep it mind, thank you.

**Edit2: I just realized by "that surface" you might have been referring to one of the mounting surfaces I suggested, I had assumed you were still talking about the aerodynamic surface as you mentioned the trimmed root plane, could you clarify?

I was still referring to the aerodynamic surface. With "giving up" on it I meant not identifying the aerodynamic surface with datum feature symbol A under the profile tolerance (as opposed to doing both that AND using datum target points on the aerodynamic surface, as I had asked about before). The rest of my post was based on my assumption that three datum target points would define a datum plane, but I think I need to study this in more depth. Thanks for your explanations.

I didn't want the conversation to derail too much on the standards topic, but I will raise it up at the company and post the result here if you guys are still interested in knowing. For more background, I'm a Junior Engineer (about a year's experience), self-taught on DG&T, and this company (a startup/scale-up) is not my direct employer; I'm a consultant doing some support work wherever they need more manpower. It's not an ideal situation, but it's what I have to work with (pun intended) for now.

 

[Burunduk]

Yuyu28,
Do update us on the results regarding the standards issue. I'm interested to know how it will develop. I think it will be great if your initiative brings a positive change to your company.