Common practice in casting design

I'm looking to gather information on what some of the casting
designers in the group do when they approach a casting design problem.

I've done a few aluminum castings this year, but I am beginning to
fell like I may not be going about it in the most efficient manner.
What I have been doing is to design the part in its machined state
(all of my castings have had significant machining done as finishing
operations). Designing in it machined state allows me to readil
control the position of the machined interfaces realtive to other
parts in the assembly. When I have finished the design of the
assembly and parts, I create a configuration or separate part that
will become the as-cast part. I add machine stock to this to this
part and fix draft problems that I may have left out in the finished
model.

I have found that I do not like the configuration route because I run
into problems with feature suppression and finding parent child
relationships that I didn't expect. Perhaps this is really just
exposing flaws in my modeling techniques...I'm not sure. I do not
consider myself a casting expert, and am definitely learning as I go.
Since there is no one at my company to show me a better way, I am
turning to the collective intelligence of this group. I then plan to
document what I find into a best practices instruction sheet for my
fellow designers to use in the future.

Any input or experiences that can be shared would be appreciated.

MHill

 

The way I like to work is to make the part as cast first, and then add
features to remove material to make the as machined. Using this method, I
never have problems with the configurations. You need to have a scaled up
config anyway to account for the shrink. Making cast or molded parts is
always a game of rock, paper, scissors anyway, between the fillets, draft
and shell. Complicating it by doing the machined features first is too
much for me to mess with.

That way makes more sense to me since it follows the manufacturing method.

I can see where you would have problems going the other way, because
creating drafted faces from the machined detail could be a real bear.

matt

 

Two responses so far and both are saying the same thing. I have a
feeling this trend may continue.

I absolutely agree that removing material for the machined features
makes the most sense from a manufacturing standpoint. I haven't done
it that way because it was easier to design in context if I was
designing to the machined state. I will have to experiment with this
and see how it works out for me.

Am I hearing that using a configuration is better than a separate
part?

MHill

 

Check this thread out about exactly the issue you describe:
http://makeashorterlink.com/?K294129F5

IMO, from a design perspective it is best to design the finished part.
Then use methods (insert part, for instance) to build the cast part from
that. Document your design intent, not the (current) manufacturing
process).

This is the same methodology as everyone seems to recommend for sheet
metal: Model what you want and let the manufacturers figure out how to
build it (just keep communication open to ensure you don't design an
impossible part).

I understand how this can be difficult to manage and I understand why
others do it 'the other way.'

 

Arlin:

I think I actually do it that way, but I build it in such a way that the
two can be separated. I'm thinking about how the mold will be made while
I'm designing the part, and I try not to model anything that I won't be
able to mold. If I've done a good job, when I'm done designing the
finished part, all I have to do is suppress and maybe reorder a couple of
features to get the cast part.

matt

 

I'm just saying that I'm accustomed to doing it that way. It is also
viable to use the as cast as a base part in another as machined part, but
if you do that, you still have to have scaled and unscaled configs of the
as cast (if you are building the molds as well). If you don't have to
worry about compensating for shrink, then that simplifies things.

matt.

 

Matt,

Unfinished first and machined later works for one component models.
However, if the process involves designing a complicated machine, the
components are usually modeled in their finished states. I don't know
how one would go about designing a complete machine as unfinished
components (castings, weldments, etc ...) and then circle back around
and create a finished machine assembly. That would be complicated.

Fundamental difference between product/mold and machine design.

Kman

 

Having done a few complex castings I'll throw in my 2 cents.

If the casting must become part of a complex multi part assembly then:

1. Create an assembly that defines the connectivity/bounding surfaces for
the components. The assembly will contain datum planes, axes, envelope
parts, etc. to define design intent. Envelope parts can be used to define
the mechanical connectivity before the casting is completed.

2. Hang the casting(s) on the assembly, i.e., top down approach with minimal
incontext features. Use the assembly to get the important interface
information and then work in the part for speed.

3. Keep the casting simple if it is complex till enough of the whole
assembly is firmed up, then refine the casting, but make sure to consider
draft and parting line issues early.

4. Add machining operations last. There are several very good reasons for
this.

a. If the casting is complex (and most will be because of draft and
fillet issues consider doing the machining operation on a derived part for speed.

b. Adding material to a model that is machined first can cause all ki
nds of issues with sliver edges, etc. It will probably be easier to cut aw
ay the casting.

c. Having the casting part first will speed up your cycle time as y
ou can let your castings drawings out while finishing the machined portion.
There is some risk to this, but some people have schedules that demand the risk.

4. Draft and possibly fillets will likely be your biggest time consumer
especially in conjunction with lofts and sweeps. Know how to use the Atomic
Bomb of Fillets.

 

(snipped good advice added to my tips folder)

Which is ???

 

MHill,

I have just finish creating a guideline for this. I will give you the
quick and dirty.

My approach is for long-term revision and understandability of the
design.

- FIY -
My method for deciding when to use a config and when to create a new
part/assy.
Use Configs for family of parts when only dimensions of existing
features are modified. When new features require to be managed with
configurations then try and opt to use base parts or new
parts/assemblies.
---------


1 - Create your as CAST design. This will be used as a BASE PART for
your machined casting. This becomes a base part you can use to create
any new-machined part. Sometimes one casting is machined to create
many different new parts...extremely dangerous to control with
configurations and risky.

2 - Create a new part, go to INSERT, BASE PART and select your
casting. All the features in this part are only for the creation of
the new part. (I created the requirement that all machining features
are to be colored RED.) It is clear to any user what has been done and
what needs to be machined. Revisions are clear and cannot affect
other machined part versions or the casting.

This method is currently being used with much success. Remember to
consider the reason why to use a configuration and what benefits you
will have.


Always try and use a MASTER set of layout sketches to capture the
design intent when creating the casting. This resolves the need to
create a machined version first and then adding material for the
casting.

I currently have a co-worker re-working an complex casting from the
configuration setup to the separate part setup. He has learned the
hard way the benefits of having a manageable design.

Jorge Medeiros
Mech Design Eng

 

You'll have to see Ed Eaton about that one. It is his baby and it works
wonders. I think it is part of his Curvy Stuff series which you can find
from the newsgroup and download.

Oh, alright I will try to explain it.

Most fillet failures are caused by a number of edges coming together in what
it euphemistically called a complex geometry failure. The idea is to
extrude a wall of material into the part centered on this failure point
then fillet up to the wall. After the fillets are in the intersection edge
of the extrusion and the part is picked and the fill surface command is
used to make a surface. That surface is then used to trim away the
extrusion and hopefully leave a smooth blended surface between the part and
the adjacent fillets.

I have sometimes had to use this four or five times to get a single fillet
to work.

 

Atomic bomb fillets-

I think Paul is referring to a section in the filleting tutorial that can be
found through the 'tutorials' link at www.dimontegroup.com

Atomic bomb filleting is a really terrific technique for working through
tough filleting situations (it just saved my rear this morning, ironically
enough) that came to my attention in Jason Pancoasts 'art of filleting'
presentation at SW World. The histrionic name is just how I refer to them -
I don't think that Jason had any name for the process.

 

Paul & Ed,

Thanks very much for the explanation. I downloaded "Curvy Stuff" some time
ago and must have missed or forgotten the technique.

In addition to having you two among the "A" list of experts on this NG, I'm
very fortunate to have Jason (and Keith Pedersen and Joe St.Cyr) answer my
VAR's support line. I just doesn't get any better.

Thanks again,

Art W.