|
January 2004
RAPID MANUFACTURING
Delivering on the promise of those 'impossible' aims
The ability to respond quickly, accurately and cheaply is
bringing rapid manufacturing techniques into the mainstream
production arena. Rick Dove looks at production options, deliverable
values and the concept of mass customisation…
Solid freeform fabrication or SFF techniques have been promising
a change in production options since the first stereolithography
deployments in the 1980s. Small wonder that the Star Trek
generation should see in this the birth of the beam-me-up-Scotty
machine - pure information transmitted to a distant location
and "poof!" a solid object appears.
Certain SFF technologies and their system configurations
have made recent advances that clearly cross the line from
prototyping to production economics. With this comes more
then just an alternative way to create a commercial object.
What comes of real importance is the ability to bring to market
products otherwise cost-prohibitive, products otherwise impossible
to make, and roducts that carry integrated service values:
one-off custom made, made at point of need, made on demand.
And so we enter a new era of production options and deliverable
values.
The phrase mass customisation has been applied to a variety
of product and service examples over the early years of its
introduction as a strategy. The concept of custom accommodation
offered by Reebok pump shoes and Gillette Sensor razors are
considerably different from the concept employed by Levi's
Custom Made Jeans. The former are all made identically, and
mould themselves at usage-time to the user's needs. The latter
are in fact custom-made at production-time to the user's specifications.
The shoes and razors are not unlike software that watches
your preferences of use and tailors its interface to you for
maximum efficiency. They are perhaps what you might call a
product of accommodation intelligence - but not, in my opinion,
a product of mass customisation.
In general mass customisation as a term has been applied
to soft products and services that lend themselves readily
to software/firmware customisation, menu-based option customisation,
or production lines with hands-on human-customisation intervention.
Until now, the heavy part of manufacturing - that dealing
with the fabrication of metal parts - has not had effective
applications and techniques for mass customisation.
Our intent is to explore some of the issues and current impact
that SFF technology has on the concept of mass customisation.
Before we can do that effectively, we must first establish
what we mean here by mass customisation.
Mass customisation, two words that seem at odds with each
other, was a phrase first coined by Stanley Davis in his 1987
book Future Perfect. Joe Pine's 1993 book Mass Customisation:
The New Frontier in Business Competition, and his 2001 follow-on
Markets of One with James Gillmore, examined contemporary
examples, projected future extensions, and generally explored
many facets of how that phrase might translate into substantive
goods and services.
The phrase has gained a life of its own beyond the attempts
of Pine and Gilmore to focus it. Nevertheless, it is generally
agreed to imply the mass production of individually customised
products and services. Pine and Gillmore go on to say that:
"Customising a good automatically turns it into a service,
and customising a service automatically turns it into an experience."
We will not address here the issues of translating a good
into a service through customisation.
As to customisation, I hold with Pine's belief that it means
demanded individuality rather than supplied variation. Mass
customisation produces goods to order, according to the customer's
specification. Providing an assembly from a large menu of
optional subsystems is not mass customisation, but rather
a sophisticated form of mass production.
Both intuitively and etymologically mass customisation takes
its meaning in counterpoint to mass production. With mass
production we make a lot of something in order to gain low
unit cost by taking advantage of volume. In the extreme, all
these somethings are absolutely identical in order to eliminate
costs necessary to accommodate variation.
Mass is common to both phrases, and, I'll suggest, drags
along with it, when applied to customisation, the expectation
that its purpose is to lower the cost of production. Not necessarily
as low as a mass-produced non-customised item, but close enough
that the cost of customisation does not preclude a market
that will support volume production.
The mass customisation concept could have been expressed,
without loss of meaning, as mass-produced customisation. I
make the point only to show that production is an inherent
part of the meanings of both mass production and mass customisation.
Thus, if we are to evaluate and examine mass customisation
as a strategic and operational alternative, we must by necessity
address the traditional metrics associated with production:
cost, cycle time, and quality.
Solid freeform fabrication is an additive creation technique
that "builds" solid objects from a pool of amorphous
materials according to shape-specifying digital information.
SFF got its commercial foothold by producing appearance models
and other forms of non-functional prototypes.
Production in Mass Customisation
Because mass customisation strives for low unit cost, we
rule out classic craft production techniques as a mass customisation
concept. Though craft production of rifles, say, can produce
a custom-built result every time, whether done by a single
person or a team with individual specialties, the cost does
not approach the mass production result. There is an active
market today for custom-made guns, but the cost premium of
current craft approaches limits the size to a few percentage
points of the total market. Plastic injection mould making,
on the other hand, has a very high annual unit volume, requires
a custom result by nature, employs a craft production technique,
and does not have a mass production option. Both rifles and
moulds, however, are candidates for mass customisation, if
a suitable approach can be found.
The most well known SFF mass customisation story is the one
about Align Technologies building plastic teeth aligners from
moulds made on eight high-end 3D Systems machines. A query
of their website reveals that they are currently treating
40 000 patients with custom made plastic aligners, each needing
a new aligner every two weeks for some period of time. Another,
a metal tooling example, is evident from the fact that General
Pattern employs four selective laser sintering machines to
produce plastic injection mould tools for customers, each
one a unique fabrication, and with alternative forms of production
also in-house, they clearly understand the tradeoffs (Figure
1).
 |
| |
SFF metal work typically produces near-net shape. In most
cases currently a final product needs additional machine work
to grind, polish, tap, or cut-to-tolerance. In other cases,
the SFF process can produce a final part as is, such as where
an otherwise cast part would be appropriate, or for certain
hand tools, production-line fixtures and patterns used for
constructing fixtures, piece parts like motor covers and air
vents, prototypes, architectural ornamentation, art sculpture,
and other such where an SFF surface finish is either unimportant
or a feature.
In the SFF approach lies certain inherent production cost
reducing potentials that can help mass customisation economics.
Some of these cost advantages are applicable whether customised
production or repetitive production is being done. The list
represents the set of possibilities for cost reduction, and
is not necessarily applicable to all forms of SFF:
• Waste Cost Elimination: In the case of metal, a subtractive
(machining) process generates waste in the form of chips and
unused stock, whereas the additive process uses little more
material then the object requires. Costs associated with wasted
material and with waste disposal or recycling are avoided
(higher costs for today's SFF materials are principally a
factor of relatively low-usage volume).
• Raw Material Inventory Cost Reduction: The generation
of objects of various sizes from a common pool of shapeless
(powdered) material rather than a diverse inventory of pre-sized
stock reduces inventory costs.
• Skilled Worker Cost Reduction: If the alternative
production method required a series of EDM electrodes to be
designed and machined, near-net SFF offers the possibility
of reducing this effort to a single electrode. Where complex
internal geometries are featured, such as with conformally
cooled moulds or other heat exchange applications, the skills
and time needed for multi-part design, machining, and assembly
are eliminated. Where a multi-part assembly of any kind can
be designed as a single SFF part, interface machining and
assembly is eliminated.
• Scrap Cost Reduction: When otherwise assembled parts
can be made as a single SFF object the opportunities for rejection
by QC are reduced. Also, as production is driven by digital
information, there is less opportunity for human error in
the production process.
• Tooling Cost Elimination: Customised generally means
unique or short run at best. Direct-to-metal SFF processes
can generate functional parts, tools, and goods without the
expense of production tools.
• Set-Up Cost Reduction: If a single material is used,
set-up costs are principally focused on CAD-file preparation
and verification, which are done off-line so they do not impact
machine production utilisation. Production interruption set-up
amounts to removal of a finished build box and immediate replacement
of a ready-to-go empty build box - typically a five-minute
operation is more than adequate, with times as low as 60 seconds
easily attainable if desired. Materials are replenished while
machine is in operation, so no lost production is incurred.
If multiple materials are permitted (eg, stainless 316L and
420), 30 minutes is more than adequate for simple cleaning
of the powder spreading mechanism, with a material-dedicated
feed box swap-out precluding the need to empty and clean the
feed box.
Whether making plastic or metal, many current SFF machines
employ technologies suitable for low-volume prototyping and
experimental environments, but not for rapid production. The
difference is principally one of cost: the cost of the technology,
the cost of the materials, and the cost of finished product
made from the two. According to Terry Wohlers in Wohlers Report
2002, the technology known as 3D Printing has an inherent
cost edge: "Owners of 3D printers pay about $25 to $40
for a part that fits in a 50 mm (2 inch) cube. This estimate
includes material, machine depreciation, system maintenance,
and labour. If the same part is built on an SLA 5000, for
example, the cost is in the $90 to $150 range". His cost
models at the time he developed them were for non-metallic
parts. Now, a year or so later, the very same numbers are
appropriate for metallic parts.
Inventory costs reduced
Production cost is only one half of the financial equation
for SFF, the other half is value - which is really just dislocated
cost. Being able to produce on demand parts and tools that
will be needed at a moment's notice means the inventory necessary
to satisfy immediate availability of seldom-used parts and
tools is eliminated, and also means the obsolesence of inventory
that is superceded by change orders or end-of-life is eliminated.
Inventory costs go well beyond the carrying cost of producing
a good and putting it on the shelf, it also includes the inventory
system of people, storage space, and computer systems necessary
to find that part and make it available on demand. And when
these systems don't work perfectly, inventory is stored that
is never found when needed, for which there have been plenty
of real life well-known military examples that undoubtedly
have their commercial equivalent in less visible corporate
environments. And of course, there is always outright inventory
loss beyond an inability to find where it is.
 |
Table 1 shows a cost model similar to the one referred to
above by Terry Wohlers. His data was for non-metallic production
perhaps 18 months ago. This table shows metal production at
virtually the same figure Wohlers showed for non-metal. Most
performance information readily available for SFF equipment
and processes is incomplete in specifying the dimensions and
geometries of the parts being made, and the quality and consistency
of the resultant material characteristics.
We are only interested here in production quality material
results, meaning characteristics that are predictable and
repeatable from part to part, and free from residual thermal
stresses that limit lifetime significantly from that available
from wrought materials.
Production costs are very much a factor of both production
quantity and part size for SFF technologies, and manifest
themselves with different processes in different ways. The
cost model in Table 1 does not care what size part(s) occupy
the build box, and only makes the assumption that 33 per cent
of the build volume for as high as the build goes (10 inch
max) will result in metal. A build volume of this size, that
can accommodate any number of multiple parts simultaneously,
puts the mass into mass customisation. Thirty dollars per
pound is not a production price that will replace the majority
of current metal production, but it is already lower than
lots of tooling costs, a good deal of otherwise-assembled
mechanisms, and many complex machined parts.
Many look at SFF technology as a curiosity that won't interest
them until it provides materials identical to those they currently
use with traditional fabrication techniques. They see it only
as an alternative way to do the same things they have been
doing. But some are starting to explore functional uses for
current SFF materials, such as Align Technologies making moulds
for custom teeth aligners, and some are starting to explore
what is possible with new design freedoms, such as DME's work
with MoldFusion™ conformal cooling for plastic injection
mould inserts and gate valves (right side Figure 2).
 |
Figure 2. Two Production Jobs Showing Customised
Variations. Two different single-build jobs run at
ProMetal's Irwin Rapid production Center. Although
mass customisation is exhibited by both jobs as is,
it should also be noted that other jobs run at other
times also made ball valves and gate valves configured
even more differently, making the point that mass
customisation in the context of these examples is
not isolated to a single build, but rather to the
functioning of the shop itself.
|
Quality and cost are issues
The cost will continue to drop as adoption increases and
materials become less expensive. Different technologies and
processes offer different advantages. Direct metal fusing
with lasers or electron beams, for instance, eliminates the
need for separate thermal processing, requiring less process
knowledge and fewer process steps to get a final part. In
contrast, with less control over the thermal process the material
characteristics have not matched those consistently obtainable
with properly controlled separate sintering, or with properly
controlled cast materials made from SFF-generated casting
shells and lost wax equivalents. Alternatively, some of these
direct metal-fusing technologies have announced a broader
range of materials than those currently offered by separate
sintering. Also, some of the directly fused techniques can
already demonstrate functional material gradients by controlled
mixing of multiple material feeds. Separately sintered approaches,
in current contrast, offer either infiltrated composites of
two metals, such as steel and bronze, or "full"-density
sintering of a single metal, such as steel or nickel. Though
separately sintered approaches recognise porosity overtly
as an integral process control variable, they do achieve higher
densities than the fused metal approaches. A subsequent HIPing
process-step can be employed to improve densities of either
approach. Preparation of electronic information to drive these
two different approaches is also different, with the fused-metals
approach requiring more attention and time to file preparation,
and in some cases restrictions on the range of part geometry.
Virtually all metal SFF technologies are being applied to
plastic injection mould tooling. SFF offers conformal cooling
advantages unavailable or at least very expensive any other
way. SFF can also offer faster cycle times on mould completion.
Mould makers generally fall into that category that wants
traditional materials rather than risk a job on a new process.
General Pattern has already shown there are good reasons to
make the move, but they, like most, apply these SFF techniques
principally to short run tools, because the material characteristics
have not yet achieved what tool makers want for production
tooling. Material quality is the issue.
Materials and process developments at ProMetal in the first
quarter of 2003 have produced stainless/bronze composites
with 35 Rockwell C hardness, and a fully dense stainless in
excess of 40 Rockwell C that will heat treat above 50 Rockwell
C. The separately controlled sintering process employed also
provides exceptional microstructure, moving SFF sintered material
into the production tooling arena.
Metal part and tool production economics have been achieved
with 3D Printing technology, enabling mass customisation for
a new and large class of consumer and industrial metal goods,
parts and tools. Mass customisation is a subset of agile manufacturing,
which achieves its response-able nature by employing design
principles known to enable highly adaptable systems.
These design principles are outlined in the author's book
Response Ability: The Language, Structure, and Culture of
The Agile Enterprise and are shown in Figure 3 applied to
the 3D Printing system that has broken the production economic
barrier for rapid metal parts.
 |
Figure 3. Agile Response Ability Model of
ProMetal System Architecture.
|
In order to provide mass customisation capability, the prime
production metrics must be satisfied, and these are cost,
cycle time, and quality. SFF by its very nature addresses
cycle time, though some are much faster than others. 3D Printing
is able to approach production economics by leveraging its
raster scan print head architecture, which can greatly increase
the amount of converted material by simply adding more print
jets, for relatively little increase in equipment cost. Finally,
recent advances in material quality for metals should remove
the final hesitancy for bringing SFF technology into the mainstream
manufacturing environment.
References
Align Technologies, "Introducing Invisalign®",
http://www.invisalign.com, 2000-2002
Davis, Stanley M., Future Perfect, Addison-Wesley, 1987
Dove, Rick, Response Ability: The Language, Structure and
Culture of The Agile Enterprise, John Wiley and Sons, 2001
DME, MoldFusion™ 3D Metal Printing, http://www.dme.net/dme/proserv/smfu.htm
General Pattern, "Additive methods Compared to Machining
or Subtractive methods", http://www.generalpattern
.com/set3.htm
Gilmore, James and Pine, B. Joseph II, The Experience Economy,
Harvard Business School Press, 1999
Pine, B. Joseph II, Mass Customization: The New Frontier in
Business Competition, Harvard Business School Press, 1993.
Wohlers, Terry, Wohlers Report 2002, Wohlers Associates, Inc.,
2002
The author
Rick Dove is the divisional director of ProMetal, a subsidiary
of Extrude Hone, one of the pioneering companies in rapid
manufacturing.
|
