Modelling Cast parts

I know kind of a cool way to see this; I'm not sure how it would work out on a drawing. But, for what a casting or forging designer needs to see (the extra, the material "allowance"), this is ideal: create either your cast or machined part, then, either cut away the extra, as would be done in machining, or design your machined part and add the material allowance to arrive at the casting. The first is CASTING-01, the second is CASTING-02. Make an assembly and put the two together, superimposed, one on top of the other. Do the casting as translucent with a lighter color ('View>Color and Appearances') so you can see the machined part 'inside'. Maybe, in a drawing, with one on top of the other, the inner model will give you a hidden line outline. Also, could this be done with a instances of a family table? or with views based on rolling back the part to show the earlier condition?

Sorry, don't know anything about Pro/CASTING but it does sound suggestive, doesn't it!?! Wouldn't surprise me a bit if it encompassed exactly the functionality you need. Although, if you go online to the PTC website, Pro/CASTING is an illusion. You get instead, TOOL design, but no specific module. Once again, PTC thinks it's being clever by creating a mystery, but will likely (if we can tell by thier history) shoot themselves in the foot. Good Luck, PTC! Hey, go with what you know.

I'd suggest you make & detail separate casting & machining models. If you model the casting, then use Insert - shared data - merge from other model to place it's geometry in the machined model as a single feature, you can re-use the same casting in numerous machined final parts.

Be sure to match absolute accuracy between casting & machining. (Use

When creating machining features, only reference datums etc in the machining model, so you can chop & change the casting without the machining going wrong.

Where are you based? I may be able to help you out further.

.02 mm or equiv?

Another option for leveraging a casting part model's geometry in a machined 'final part' model is the Inheritance functionality. Inheritance is an excellent tool for this particular application.

Ron M.

Better yet, model the machined part first and then use it as an Inheritance and basis for the casting. Changing a feature in the machined part will automatically alter the cast shape.

Who can model the casting first anyway? When you think about, the machined part drives the casting, not the other way around. You need to know what the final part looks like before you can determine the casting geometry.

Regards

Apparently most everyone from what I've read. Check out :

Cheers,

Ron

More on this topic. Just this week I had a situation where I knew what the cast part was going to look like, and I also knew where several machined hole features were going to be placed afterwards. Therefore I just modeled the cast part first, then created a new part model and added an Inheritance feature that referenced the cast part. From there I add the machined hole features.

So it really boils down to the individual's(designer/engineer) thought process, company's processes, etc. how you want your work flow in Pro/E to go. Therefore you really have three main choices here for the modeling aspect of dealing with cast and machined parts:

2) Inheritance functionality

I really don't think that there is a clear cut right or wrong way to generate these types of models in Pro/E, but the application most definitely offers the user some choices for the task at hand.

Best of luck to the original poster!

Ron M.

Which came first; the chicken or the egg?

The few casting models I've done began with a customer's furnished part. I copied machined faces, created over lay quilts (cast condition) and merged. New family table instance cut the machined faces using the copied surfs as a basis.

I `think' I'd go thru the same process starting from scratch. It fits the way my head works and I'm not sure I'd always know if a casting will be used during initial design. If I did a lot of that kind of stuff I might re-think. Knowing I'm looking for a casting and modeling from the get go as such would be more efficient, or so one would think.

Thank you very much to everyone who has posted. I'll need to see how my company want to do things - cast first then machined or vice versa. I'll certainly try out all of your suggestions and see which suits best.

I had thought about using the inheritance feature but wasn't too sure what this would do. What is the difference between inheritance and the merge option from shared data?

As for the Pro/CASTING, I found that we do have it installed - its Pro/MOLD / Pro/CASTING extension, seems to be mainly for setting up the die but I have still to investigate this further. As David Janes pointed out above, there doesn't seem to be anything on PTC's website about this. I'll just have to go through the help files some more.

Thanks again for all your help.

The way we do it is to:

1. Create the casting.
2. Add machining features.
3. Group the machining features (call it "machining" or something like that)
4. Create one drawing for the casting and one for machining using the same model.
5. On the casting drawing just use VIEWS>REPRESENT>SIMPLIFY (and select your machining group which will now dissappear on the drawing.)

NOTE: The only way to get the views menu (it was shown in previous releases of Pro/E but hidden for some reason in Wildfire) is to create a mapkey for it:

Use a text editor to open your config.pro file and add the following: (then save it). mapkey Representation @MAPKEY_NAMEDrawing View Representation;\ mapkey(continued) @MAPKEY_LABELRepresent;#VIEWS;#REPRESENT; Or use syntax for a typed mapkey like: mapkey vrv #VIEWS; #REPRESENT;

Restart ProE Wildfire and open a drawing. You can customize your screen to include the mapkey as an icon on the tool bar, or use the keystrokes vrv as in the example above.

See this webpage for reference:

"Who can model the casting first anyway? When you think about, the machined part drives the casting, not the other way around. You need to know what the final part looks like before you can determine the casting geometry."

Bye and large the two evolve in step with each other through the design process. Keeping the models separate has the advantage of letting you take multiple finished parts from one casting without having lots of suppressed groups of machined features at the end of the model tree which you need to drive with a family table and your pdm system probably won't support properly.

If you're lucky enought to have an IT department who thought about family tables when they configured your plm, can I come & work for you?

John,

I think you're doubling your work by maintaining to two distinct and un-related models. I've never known enough at the start of a design process to determine the casting's geometry before that of the finished part - it is just counterintuitive.

With Inheritance, you create the cast model, insert the Inheritance feature, suppress the machined bits that you can (holes, etc). Then add offsets to the surfaces that will be machined.

You would not use this part as the basis for another machined part. Instead, you would either family-table the original machined part or make another Inheritance-based part for your machined iteration. That way, a change to the original (which is changing the casting), would propigate.

Regards

John Wade wrote:

I can't speak from personal experience of inheritance, all I can do is say I have a process which works for me. I'm not following the creation process you outline, do you have pointers to any documentation?

Well, we always hope we help. Often, it's a shot in the dark. This one stirred up some process discussion, not that it directly pertained to your technical questions on representing, old style, cast and machined features in a single drawing. Well, easy in ACAD, not so easy in Pro/e, that's about all I got out of this discussion.

That, and the fact that most people don't seem to appreciate what a casting is. Well, think brass bell (Liberty comes to mind), think cannon, think trivet or potbelly stove or automobile frame or lathe bed or axle. Now we're starting to get into cast parts that actually require machining. But, please, one and all, the charm, the wonder, the economy of casting is that 95-100% of the finished product is captured in the rough casting. Even today, when modern products demand greater accuracy, some type of casting will fit the bill:

David,

People usually make castings for cost and, most importantly, because it is next to impossible to machine some shapes. And you don't put rough castings or forgings in a finished product, so the machined part is hardly an 'afterthought' - it is the primary focus.

Forget the liberty bell for a moment and let me give you a 'real world' (not Old World) example:

You are working on an engine. You want a bracket to mount the alternator to it; odds are pretty good you already have the alternator AND the engine (who would make the engine around the bracket?). So, you open the engine assembly and insert the alternator into it. You position the alternator assy where you want it in space; now you start making a bracket that fits.

Do you mean to tell me that, at this particular point in your design process, you think you can start modelling a CASTING? Of course not, you are going to make a finished part. You're going to put in hard points, mounting holes, screws, etc. You're going to make sure you can get a socket wrench in to the screw heads. You're going to be finessing the location so that you have good wrap angles on the belt and the center distance works out to a standard belt size while the tensioner is mid-range in it's motion. There are dozens of little corrections to fit and usability, none of which you can make with a casting model. If you're worth your salt as a designer, you already know if you're going for sand, investment, lost foam, or die casting and you're considering constant wall thickness and draft angles and other peculiarities to each process.

Now you find yourself with a finished, machined part - not a rough casting. In order for you to finish your bracket project, you then - downstream - need a casting model. Your first 10 parts you're probably going to make from solid or SLA while you're waiting on your foundry's tooling and samples anyway.

Meanwhile, your buddy Tom has decided that where you put the alternator is just IDEAL for the power steering pump - and he wants to know if you'd mind sharing your bracket with his pump? C'mon, all it needs are three holes and an ear, you can do that right? 'Course you can. So do you add it to your casting to see if it too fits? Of course you don't

- you add the holes to your finished part - and the ear too. You want to see how the power steering pump looks alongside the alternator. Now you need to get a three cylindrical bosses and an ear on your casting too.

Meanwhile, Fred needs to do something about the fact that you've just put a mounting pad onto the side of his engine block for your bracket - and you weren't working with the raw casting either; were you -how could you have been?

That's why you just cannot know in advance what your casting is supposed to look like until you've modelled the finished part. You cannot design an assembly like that working with raw material any more than you would put a piece of 4340 round stock in place of shaft in a gearbox - which is essentially what you're advocating when you get right down to it. Finished parts come first.

John Wade:

I'm sorry if I was a bit unclear in my description of the use of Inheritance for castings, so here it is again:

-Model your finished part

- Make a new empty part - File>New

-Insert an Inheritance of the finished part as the first feature - Insert>Shared Data>Inheritance from Other Model. Assemble it to Default.

- Expand the model tree of the Inheritance feature so you see the model structure of the finished part

- Working from the bottom of the model tree and going up, highlight each feature that would NOT exist in the casting (tapped holes for example) and suppress it. This will NOT propigate back to the original part. Be careful at this point to make sure that parent/child relationships are not causing the uninteded suppression of features further down the tree. If they do, you may need to alter the original (machined) model's design intent or find another way to deal with the feature.

-Once you've suppressed the machined features that you can, add offsets (Edit>Offset) to the surfaces requiring cast stock for finish machining or fill in with Extrudes areas that you want material on for fixturing, etc. Obviously, you're probably not going to be adding any cuts.

Offsets are just one way of adding material - there are also Variable Dimensions and plain old Extrudes.

Now that you've made your casting and made it parametric to the finished part, if you decide later on that you need to add something - or remove something - from the finished part, all you need to do is add it and then open the casting model and RMB on the Inheritance feature and do Update Inheritance for your changes to propigate.

Regards Peter Brown

Peter, it's unfortunate that you didn't trouble yourself to quote my presentation on this subject of castings. It was historical and fairly broad reaching. But, I thought, when I wrote it that it captured the essentials of castings, the main reasons for founded parts, then as well as now. What has changed is only that more is expected of founded parts, possibly more than ought to be expected of them. You seem to jump over them as something dirty and inconsequential. The fact is, that people continue to design castings which contain all the main elements (stuctural, mechanical, esthetic) of the functional part that they need. You have difficulty with the idea that machining is an "afterthought". I understand because the machined part is also essential, one might even say, the whole point of the part, the thing which makes it actually useable. On machined castings, I'll agree. The last castings I worked on were hydraulic valve bodies. Probably 25-30% of their surface area was machined. Beyond that, one begins to wonder why bother with castings. True enough, coring (which hasn't been mentioned so far, in this discussion) pre-removes material so that there is less to remove with machining. In the case of valve bodies, cores, which are essential to casting design, make passages through the valve body to facilitate the machining of valves and ports. As to the design of valves and ports, the essential design criteria are the final, machined feature. Only this decides such things as minimal wall thickness and how to space/mount hydraulic or electrical throws on valve stems.

You live in denial (ignorance?): fully 95% of all castings require NO FURTHER MACHINING. (By number, not gross tonnage). Most of the rest require a little (one or two machined features) modification. But that's the essential point of castings: you design the casting: all the functional, mechanical and esthetic features are included. All that is sometimes required is to smooth its interface witht he rest of the world. So, machined features are required to construct this interface. Generally, if designed properly, the casting will undergo little or no change because of such modification. When you reach the point where you are simply taking a finished, machined part and adding a little 'extra' everywhere, you've pretty much lost the POINT of a casting: MOST features could be used, AS CAST ~ that was the TRUE cost savings. When 90-100% of surfaces are machined, the advantages of the casting are lost. Might as well take a saw cut billet and cut everything out of it, especially if it's a more free machining, less abrasive, stronger material.

I do the bell and trivet, purely cast, no machined features and you do the engine block, extreme opposite, nearly 100% machined, the place at which manufacturers start to envision other technology because the raw material price (casting, saw cut block, etc.) is far outweighed by the cost of machining ~ 10 to 1. The more machined features you put on a casting, the less reason there is to use one, the less savings, the smaller the machining advantage from coring (see previous point on grey iron abrasiveness, increase in tooling costs.)

Again, bear with us here. If you are modelling a finshed part, you are modelling a finished part. The casting is almost incidental. I repeat: 95% of casting design is NOT like this. Functional castings are designed every day, they are not machined, but cast, parts. Zero, or minimal, machining is required. And the machining is because of interface, not structural/mechanical, requirements. The structual/mechanical stuff is already taken care of in the basic, cast part. So is the basic geometry. So are the esthetics. Somebody wants the bell to rock more smoothly, not wobble around; they machine the hole and normal sides of the tang. It rocks smoother, the sound is cleaner. Small modification, big effect.

Thanks for this nice description and introduction to the use of an inheritance feature: with this, I might actually try one now.

One of the best discussions we've had in this NG. Thanks, Peter, for taking the time and sharing your experience.

David,

Forgetting the historical perspective, what I completely disagree with is that you can model your casting before your machined part. I've had a few engineer pull that stunt when moving parts from 2D to 3D ( where the geometry was predetermined) and it is a pain in the ass. Why? Because parts change over time (the old PTC lifecycle mantra). And when we decide that we need more thread engagement on a part and a wall or flange has to be thicker, do you know how tedious it is to try to infer from a casting which dimension you need to change to achieve the desired result, without causing other geometry to move? And - AGAIN - I don't care what is going on with the casting really, I want more threads - a machined feature. I should change it at the machined part level, and then the casting should catch up with it.

And again, I'm sorry - but I have to take exception to your statement that "95% of all castings require NO FURTHER MACHINING"..? You generally make well-reasoned and rational arguments, but don't tell me that I am in denial - or that I'm ignorant. How can you come up with such an unsupported figure? I've spent 21 years in industry and I've never seen a casting that we just turned around and shipped out again as finished. 95% bears no relationship to reality, unless you are discussing a tape measure case or similar die casting. All that stuff's going injection molded anyway. I mean, c'mon, who is making BELLS? Look at all the products PTC touts - Harley V-Rods, John Deere, Caterpillar......not a bellmaker among them. Probably not a finished cast part either.

To put your 95% claim to rest, I challenge you to go in your garage and find a SINGLE - just ONE - cast part on your car, lawnmower, chain saw, iron rake, or bicycle that has NO MACHINING. Sintering doesn't count, molding doesn't count. Open the hood and find one, get out your creeper and find one. Cranks, cylinder heads, intake and exhaust manifolds, u-joints, steering forks, mower decks, control arms, differential housings, throttle bodies, aluminum rims, brackets, housings, brake rotors for God's sake.........ANYTHING. When you find one, let me know.

Until then, castings are the downstream BYPRODUCT of machined parts. Remember, virtually any cast part can be machined from solid, with enough effort....and hardly any part as a casting can substitute for the machined one, I don't care how much you try.

Regards, The Ignorant Peter Brown

David Janes wrote: