The way you design a product should depend on the quantities of the product to be made. If you are to build one special machine or electronic system, it is less costly to use established engineering practice to make a product which will work right the first time, rather than to undertake extensive R&D, and to design for manufacturing processes which do not require large tooling investments (e.g., molds, special integrated circuits) even though it would be costly in quantity production. On the other hand, if you, as a design engineer, are designing a mass-production consumer product, it pays to do a great deal of R&D and to design for very low per-unit cost at the expense of large tooling investment. In accounting language, the total nonrecurring cost plus the total recurring cost for the complete production should be a minimum.
Business Considerations
This simple principle is modified in the real world by several other considerations which are matters of judgment rather than of calculation. Is there enough working capital available to pay for a large initial tooling investment? How sure is your company that it will really sell a large quantity? What is the probability that you will want to or have to change the design when you are only partway through the predicted production quantity?
Models
A perpetual problem is how to make single models and preproduction short runs of mass-production products without the full investment of both time and money required for production tooling. In electronic circuitry a special integrated circuit can be reliably replaced by standard components during R&D at the cost only of physical space. In mechanical manufacturing it is much harder without tooling to make parts and large assemblies which are equivalent to tool-made parts.
Many processes have been developed to solve portions of this problem, and it is worth the time to learn about them.
Eleven Short-Run Techniques:
- Single-cavity "soft" tooling can be made for plastic molding and die casting at a fraction of the cost of production tooling. The molds may require much manual work to assemble and disassemble, and the part may require some hand finishing, but the total cost is relatively low.
- There are companies which specialize in short-run stampings. The dies employ no die sets, and their use requires much handwork per part, but they are surprisingly inexpensive.
- Die-cast parts can be simulated by sand-cast or lost-wax-cast parts with machined finishing. Molded plastic parts can be simulated with cast plastic with machined finishing. Of course, almost any part can be "hogged out" of a solid block by machining. For small to medium quantities, NC machining may be surprisingly economical. There is, of course, the risk that for highly stressed parts the strength simulation will be poor, either too bad or too good, and material substitution may prevent correct simulation.
- Small quantities of lost-wax castings ("precision castings") can be made from expendable patterns machined from polystyrene components cemented into complex patterns.
- Flame-cutting and welding plate stock can simulate large iron castings as well as being a valuable production technique.
- Complex solids can be made by printing and etching thin sheet metal and soldering or cementing laminations into a solid.
- Simple parts such as lengths of tubing can be cemented, soldered, brazed, or welded into simulations of complex parts.
- There is a new technique in which complex plastic parts are made by progressively building up layers of liquid plastic hardened by a point of ultraviolet light which scans a container of the liquid under computer control.
- Remember the availability of some unconventional machining systems, usually employed for toolmaking, such as electrical-discharge machining (EDM) and laser machining.
- Kits are sold for making experimental printed circuits in your own laboratory in a very short time.
- There are breadboarding boards and terminals for simulating printed circuits.
