Patternless Casting
a report by
Yehoram Uziel
Founder and Chief Executive Officer, Soligen Technologies
  Yehoram Uziel Yehoram Uziel is founder and Chief Executive Officer of Soligen Technologies, Inc., where he has developed a patternless metal casting process called Direct Shell Production Casting. He holds a number of patents related to this technology. From 1990 to 1993, he was President of ConceptLand Ltd, rapid prototyping service bureau, based in Israel. From 1989 to 1992, he served as Vice President of Engineering at 3D Systems, Inc., the developer of SLA machinery.

     Before 1989, Mr Uziel was co-founder of Opto-Mechanical Division, Optrotech, Inc., a manufacturer of inspection equipment for the practical circuit board industry. He is a frequent presenter at engine developmentand rapid prototyping conferences worldwide. He has a degree in Mechanical Engineering from Technion Institute of Technology.

     A designer has just completed a three-dimensional (3-D) computer-aided design (CAD) model of an exotic shape for a water jacket on a new cylinder head. Now it is time to test the real thing: so the designer clicks on a print icon, and minutes later the cylinder head materialises in the exact net shape of the CAD design.

     Advances in computing capabilities and Solid Freeform Fabrication technology (SFF)1 have made this designer’s ‘science fiction’ fantasy a reality for many part designers. SFF technologies use computer graphic representation and simple stock material to fabricate complex parts. Fused Deposition Modelling (FDM)2, Laminated-Object Manufacturing (LOM)3, Selective Laser Sintering (SLS)4 and 3-D printing5 can create functional parts in production materials ranging from plastic and nylon to ceramic and powdered metals. And with advanced computer numerical controlled (CNC) machining, metal parts with complex external shapes can be easily produced in limited quantities. But, for designers of metal parts with complex internal geometries like cylinder heads, manifolds or transmission housings, traditional foundry sand casting is still the method of choice for producing production-intent prototype parts for functional testing. This new desktop manufacturing paradigm remains a dream. To create their functional cylinder head, the designer will need to design a set of tools or patterns for conventional sand casting. They will need to design:

  • parting lines and add draft so the tool can be removed from the sand mould;

  • core boxes for the complex sand cores that will create the void for the water jacket; and

  • prints to aid the foundry worker in the assembly of the cores to the sand mould.

Of course, the design of each of these components must consider appropriate shrinkage factors for the particular alloy she has in mind. She might employ many of the rapid prototyping techniques above to speed up the fabrication of the sand-cast tooling. The completed tools would then be sent to a foundry, where an experienced mould designer would design gates and risers to supply molten metal to the mould and permit the metal to cool without unwanted shrinkage. Even for an experienced mould designer, it will take several iterations to get a solid casting that is free from any porosity that might cause leaks from the water jacket to the combustion chamber. If a single miscalculation or mistake occurs, it is necessary to go back to the beginning to try all over again. It is no wonder that engine developers allocate months for each design iteration of a new engine.

     However, designers are now discovering ways to apply SFF technology in the foundry to create even more complex metal castings in small quantities without costly and time-consuming tooling6. They have discovered that they can produce production-intent parts for functional tests in days instead of months by foregoing the design and fabrication of tooling until the part design is finalised. More design iterations are possible in less time. They can even test multiple concepts with different configurations or different alloys simultaneously. Furthermore, as the final part design is fully tested, mass production tooling is created only once and correctly the first time. Automated die cast tools for a typical automotive cylinder head can cost in excess of two million dollars. Every time a design changes, these tools must be modified or scrapped. By employing patternless casting, designers can better optimise their designs and eliminate hundreds of thousands of dollars in wasted retooling costs. 

     One technique for patternless casting, called SandForm, was pioneered by DTM Corporation. In this process, foundry sand mixed with an organic binder is formed into a mould using an SLS machine. The sand is deposited in a thin layer and the shape of the part is sintered with a laser beam. The mould, complete with integral cores, is built up one layer at a time. Molten metal is then poured directly into the mould. Once the metal is cooled, the mould is broken away revealing the near net shape part.

     Another technique for creating moulds directly from a 3-D CAD design is called Direct Shell Production Casting (DSPC)7 and was developed by Soligen Technologies, Inc. DSPC uses 3-D printing to create a ceramic mould with integral cores. A thin layer of ceramic powder is spread and then an ink jet printhead deposits a silicate binder onto the powder in the shape of the part. After the shell is built, complete with gates and risering, the loose powder is removed and molten metal poured into the remaining shell.

Table 1: Lead Time Comparison for Cylinder Head Development

Conventional Sand Casting Patternless Casting
Design changes to head 3 weeks 3 weeks
Design patterns and core boxes 2 weeks Not Needed
Fabricate patterns and core boxes 8 weeks Not Needed
Design gating system 1 week 1 day
Fabricate moulds 1 day 2 days
Cast Heads 2 days 2days
Machine, check, and assemble heads 2 weeks 2 weeks
Dynamometer testing 2 weeks 2 weeks
Total Time 18 weeks 8 weeks

     Patternless casting, using either of these techniques, offers a variety of benefits to the designer. Early design concepts can be tested and compared with no regard to the undercuts, parting lines and core prints usually required in sand casting. Since no cost-prohibitive tools are created, multiple part configurations can be produced and tested simultaneously. Design modifications are easily accommodated in CAD and require no retooling. The time between iterations can be reduced by over 50%, allowing for better design optimisation and reduced time-to-market. Following are three case studies highlighting the effective application of patternless casting for engine development.

Racing Intake Manifold

     A major racing engine developer had just completed the design of a major new cylinder head upgrade. However, initial test-drives proved disappointing. When coupled with the existing intake manifold, the engine lacked critical acceleration. A new intake manifold design was needed to realise the full potential of the new cylinder head. With only weeks remaining before the year’s biggest race, a non-conventional approach to design and testing had to be considered. The designer created three potential designs to solve the airflow problem. Using patternless casting techniques, all three configurations were cast in aluminium and delivered to the test track in two weeks. The optimum design was identified and several more castings created in time for race day. As a result, drivers using the new engine and manifold came in first, second and fourth.

Motorcycle Cylinder Head

     A major motorcycle manufacturer had some radical ideas for a new engine. Starting with a clean sheet of paper, they created five different iterations of the engine, optimising performance and manufactur-ability in less than a year. The initial engines produced via patternless casting validated some new patented concepts. The engines were so ugly, with undercuts and complex cores, that they could never be produced with conventional casting techniques, but they proved effective in demonstrating the validity of the new concepts. Later iterations optimised the design for permanent mould manufacturing, resulting in a design that could be mass produced at a cost similar to their older engine designs, even with the new patented features. Using patternless casting, this manufacturer began delivering new motorcycles to the market two years sooner than they could have using conventional techniques.
 

Automotive Cylinder Head

     A designer at a major automotive manufacturer had four months to complete the testing of a new combustion concept in a single cylinder head before his team needed to begin design of a complete multi-cylinder head. Sixteen weeks was just enough time to produce tools for conventional sand-casting and bench test one set of single cylinder heads. If anything went wrong, he would be out of time to recover and complete his tests. Using patternless casting, he produced his first pair of cylinder heads for tests in four weeks. After two weeks of tests, he had some ideas for further improvements. The design was modified and another set of single cylinder heads was produced and tested. The result was a combustion design that greatly exceeded all of the programmer's initial goals.

Conclusion

     By definition, metal casting is well suited to large-volume production. Complex systems can be cast as a single unit at a fraction of the cost of fabrication or machining. But, for the low-volume runs needed during the development phase, the time and the cost of creating part-specific tooling has made it difficult for designers of castings to test production-intent functional parts. There are too few parts produced to adequately amortise the cost of production tooling. In some cases, the designer is forced to compromise on material for testing. And when production-intent material testing is critical, the designer is forced to wait for the design and fabrication of tools to produce a few prototype parts.

     Now, geometric information about a part’s shape can be sent to an SFF machine to manufacture a mould for casting metal directly from an STL CAD file8. This desktop manufacturing is analogous to desktop printing. Offset printing is dedicated for high-volume mass production using speciality machines. Desktop printing is used for instant turnaround of short runs.

     While solid freeform fabrication techniques will probably never replace conventional manufacturing for mass production, it has proven that it can greatly reduce the cost and time associated with new product development. Finally, designers of complex metal castings can benefit from this technology as well. By delaying the creation of mass-production-type tooling until after a part is completely designed and tested, a designer can eliminate the time and expense of tooling and re-tooling as a part’s design is optimised. Mass production tools are created only once and correctly the first time potentially saving hundreds of thousands of dollars.
 

1. SFF: Solid Freeform Fabrication: a group of technologies that render a solid replica of a 3-D CAD design. It is sometimes referred to as ‘desktop manufacturing’.
2. FDM: Fused Deposition Modelling: a process in which a thermoplastic material is heated and extruded on an XYZ table to produce a 3-D part.
3. LOM: Layered-Object Manufacturing: a process in which one sheet of a material after another are made to adhere to each other forming the layers of a part. Areas of a layer not included in the part contour are cross hatched with a laser for eventual removal.
4. SLS: Selective Laser Sintering: a process in which powder is deposited in a thin layer and scanned with a laser so that powder particles adhere to one another in selected areas. The layering process is complete until the part is complete.
5. 3-D Printing: a process in which layers of powder are scanned by a printhead that deposits binder in selected areas. After successive layers are built, loose powder is removed to reveal the part.
6. Tooling: The jigs, fixtures, dies and moulds used to create metal castings in a foundry. Tools must be specially designed for each part and can take months to produce.
7. DSPC: Direct Shell Production Casting: a patternless casting process that uses 3-D printing to create ceramic moulds for casting metal directly from a 3-D CAD file.
8. STL file: a data format that approximates the desired part shape through a collection of triangular planar shapes. It is supported by all of the major CAD vendors and is considered the standard interface for SFF machines.