GD&T On Drafted Plastic Parts

[steveapathy]

I'm looking for resources that will help with the application of GD&T on injection molded plastic parts. How does the application of GD&T callouts differ when draft is applied? Our software tools for application of GD&T and for analysis of GD&T do not seem to be built to easily define drafted surfaces.

I feel like the datum feature simulator that I want for a drafted rectangular slot is any different than that of an undrafted rectangular slot. The fact that none of our tools are capable of handling drafted surfaces, however, leads me to wonder if there are some nuances in the standard that I am missing. To put it in perspective, I often deal with very small plastic parts with linear tolerances of ±0.003"

Anybody with some experience in this?

Thanks!

 

[3DDave]

You can proceed as if draft is a known fabrication error. For example, there is nothing that prevents taper on machined surfaces as long as they are within other applied limits. Set size and other limits such that draft fits within those limits.

If draft is a noticeable contributor you might simulate that with profile or angularity, depending on the exact situation.

 

[MikeHalloran]

For the majority of the dozens of plastic parts I have designed or engineered, the actual draft, and even the draft direction, was not known at initial design time. ... so the Rev A drawings went out with draft mostly not specified, except for surfaces where a particular, or zero, draft was needed for design reasons.

In the few instances where we tried to anticipate what was needed, and bothered to specify a draft angle and direction, I'd estimate that in at least half the cases, the molder chose a completely different parting line than we had, so our drawings were 'wrong'. In the vast majority of those cases, the parts were satisfactory, so the molder was 'right', and we revised the drawings to reflect the parts as produced.

On the Rev B drawings, where the first parts have been measured, or in the rare case where the draft is a design feature to be controlled from day one, it's helpful to establish gage lines or gage planes, where the nominal feature dimensions are to be measured and controlled, with draft specified by an angle, or perhaps controlled by nominal dimensions at multiple gage planes.

IOW, attempting to estimate what draft will be necessary for molding purposes, or specifying a draft other than zero-ish on the (first) print, is largely a waste of everyone's time.

That reflects the actual moldmaking process, where cavities are usually sunk with near zero draft, and more draft is added iteratively until the molder can produce a satisfactory part.

The iteration between the molder and the moldmaker, and between the molder and the designer, is one reason why it's a good idea to develop a personal relationship with a molder, and to stick with as few molders as possible.


If your outfit has been poisoned by a Supply Chain Manager, who always takes the low bid, and tries to swap existing molds between molders to 'save' a few pennies, the iterations are _much_ more expensive in money and time, so your schedule should include a lot of 'water' for your time wasted finding 'better' molders, 'better' here being defined as someone who will put up with the Supply Chain Manager's crappy treatment and still produce decent parts when you need them. This will be a continuing drain on your time, since Supply Chain Managers keep pressuring molders for lower prices, shorter lead times, and smaller lots, and every molder you can find will eventually tire of the abuse, develop a relationship with your competitors, and be mysteriously unable to produce your parts when you need them.



Mike Halloran
Pembroke Pines, FL, USA

 

[steveapathy]

Dave,

The issue with taking it as fabrication error is that it allows for additional variation at the point we truly care about. For example, on a drafted rectangular slot, the smallest portion of the draft may be a critical dimension that must be molded to within ±.002". If they make the slot wider with no draft they could still meet the print while functionality is lost. I guess this is where angularity or strictly specifying draft angles would be required. I'm just concerned with cluttered or overconstrained drawings.

Mike,

We are extremely picky in some cases on where our parting lines are, and draft angles often have a functional requirement. What we tend to struggle with is indicating how a part should be measured in order to ensure it meets our functional requirements. This is further compounded when all of our features are both drafted AND rounded. I am not familiar with the gage line/plane terminology that you mentioned. Can you elaborate?

 

[pmarc]

steveapathy,
Take a look to Y14.8-2009. This is the standard dedicated for castings, forgings, and molded parts. Para. 3.6 shows different methods of draft angle specification. Unfortunately you won't find much info about application of GD&T callouts there, apart from quite extensive chapter on datum reference frame establishemnt from datum targets. But who knows, maybe you will find something interesting for you.

As for the gage line/plane terminology that MikeHalloran mentioned, I believe he meant something similar to what is shown in fig. 8-18 of Y14.5-2009. The plane located at basic 18 from datum plane B is the "gage plane", at which the size of the feature is measured. The remaining portion of the cone is controlled by the profile of surface callout.

 

[steveapathy]

I've skimmed through Y14.8-2009, and I found it to be just as you said. Good for additional rules that only apply to plastic parts, but not good for application of existing rules of Y14.5-2009 to plastic parts.

All of the methods I can think of using for completely describing drafted/rounded features will make a drawing confusing to all but the most experienced GD&T users, which our vendors are not. Heck, neither am I.

 

[pmarc]

Perhaps you could post a sketch showing at least simplified geometries you are dealing with, together with short description of functional/geometric requirements. Maybe we will be able to find a reasonable solution. Can't guarantee, but it is worth to try.

 

[3DDave]

By known fabrication error, I meant that you need to account for it in the tolerance description. It doesn't allow for anything extra.

In the example of the rectangular slot, apparently the need for control varies along the depth of the slot. You can apply a tolerance that varies with the depth. Ref figure 5-40, p138 of ASME Y14.5M-1994. In addition, there is nothing in the standard that precludes having a note or diagram on the drawing that shows exactly the limitations you want.

The standard provides a set of rules for typical situations to condense a large amount of description, related figures, and interpretive effort into a few symbols on a drawing. It does a fair job where parts are to locate, orient, or avoid each other, but it is less obviously applied for directly controlling secondary effects, such as flow rates that vary with orifice shape or friction loads that are due to an amount of interference combined with feature section properties.

Finally, if your design requirements exceed the ability of the suppliers to understand them, you either need new vendors or a new design based on the limited abilities of those vendors. Maybe you need both. It's possible your vendors will claim to understand the requirements, then ship garbage and argue endlessly, forcing you into court, bankruptcy, a new vendor, or all of the above.

 

[TommD]

We dimension typically to the MMC condition of the feature & tolerance accordingly. We almost always draft our parts (they're small, not tiny, usually under 1 x 1" with multiple functional features) & we also require the vendors to review their parting line, gating & tool geometry with us. After, all, they are making cavities from models we provide.

With GD&T features, we try & anticipate how they will get inspected since that after all is the purpose of putting it there. Part function dictates the design, but the dimensions are there for measurement.