Materials: engineering, science, properties, and design (4th edition)

M.F. Ashby, H.R. Shercliff and D. Cebon

Exercises with Solutions for Chapter 18

Exercise E18.1 (a) Explain briefly why databases for initial process selection need to be subdivided into
generic process classes, while material selection can be conducted on a single database for all materials.
(b) After initial screening of unsuitable processes, further refinement of the process selection usually requires
detailed information about the materials being processed and features of the design. Explain why this is so,
giving examples from the domains of shaping, joining and surface treatment.

Answer.
(a) The different process families (i.e. shaping, joining and surface treatment) fulfill different functions and do
not therefore compete with one another. The attributes which describe the capabilities of the processes
(technical, quality and economic) differ significantly between process families, requiring different record
contents in the database. Material properties are broadly universal to all materials, so a single database
suffices.
(b) Initial screening handles most technical requirements and some aspects of quality. Many design
requirements are complex – the result of using a process is coupled to details of the material being processed
and/or details of the design. Examples:
Shaping:
- avoiding defects: the defects formed in shaping depend on the process class (e.g. casting, forging, powder),
vary from alloy to alloy, and depend on component details such as the aspect ratio of thin sections or changes
in section.
- product properties, such as strength: sensitive to alloy, particularly if a heat treatment forms an integral part
of the shaping process (e.g. extrusion).
Joining:
- joint properties: strength and fracture toughness of welds depend on alloy (e.g in steels, the hardenability),
and the thermal history, particularly cooling rate (which depends on design features such as thickness, and
process conditions such as power and speed).
Surface Treatment:
- wear resistance: depth and hardness of surface treatments for wear resistance are sensitive to process
conditions and alloy used, and the consequent wear resistance depends in a complex way on service loads,
environment etc.

Exercise E18.2 A manufacturing process is to be selected for an aluminum alloy piston. Its weight is
between 0.8 and 1 kg, and its minimum thickness is 4-6 mm. The design specifies a precision of 0.5 mm and a
surface finish in the range 2-5 µm. It is expected that the batch size will be around 1000 pistons. Use the
process attribute charts earlier in the Chapter to identify a subset of possible manufacturing routes, taking
account of these requirements. Would the selection change if the batch size increased to 10,000?

Answer.
Inputs used for aluminum alloy piston:
 Material: aluminum alloy
 Mass range: 0.8 – 1 kg
 Minimum thickness: 4-6 mm
 Shape: 3D solid (part hollow)
 Tolerance: < 0.5 mm
 Roughness: < 5 µm
 Economic batch size: 1000, or 10,000.

1
Material: all metals processes apply - sand, die, investment, and low pressure casting; forging, extrusion and
sheet forming; powder forming; electro-machining and conventional machining.
Mass range: sand casting and extrusion marginal.
Thickness: sand casting marginal.
Shape: extrusion and sheet forming excluded for 3D shapes.
Tolerance and Roughness: only sand casting excluded, but both could be rectified by machining.
Combining these screening stages, many options remain: sand casting (plus machining), die, investment, and
low pressure casting; forging; powder forming; machining.
Considering economic batch size, for the surviving processes:
- for 1000 units: sand casting (plus machining), investment casting; forging; machining.
- for 10,000 units: sand casting (plus machining), die and low pressure casting; forging; powder forming;
machining.
In this example, only shape and economic batch size provide some discrimination. Further refinement of the
selection requires a more detailed cost analysis. Experience indicates that die casting or forging are the main
contenders. Note also that the choice of casting vs. forging implies a different type of aluminum alloy in each
case (and hence properties).

Exercise E18.3 The choice of process for shaping a CD case was discussed in section 18.5, but what about
the CDs themselves? These are made of polycarbonate (another thermoplastic), but the principal difference
from the case is precision. Reading a CD involves tracking the reflections of a laser from microscopic pits at
the interface between two layers in the disk. These pits are typically 0.5 µm in size and 1.5 µm apart, and so
the tolerance (and roughness) must be rather better than 0.1 µm for the disk to work.
First estimate (or measure) the mass and thickness of a CD. Use the charts to find a short-list of processes that
can meet these requirements, and will be compatible with the material class. Can these processes routinely
achieve the tolerance and finish needed? See if you can find out how they are made in practice.

Answer.
Initial constraints:
Material: polycarbonate; mass of CD ≈ 15g; thickness ≈ 1.2mm.
From the charts (Figs. 18.3, 18.6, 18.7): Processes passing: injection molding, blow molding.
From the material-shape matrix (Fig. 18.9), blow molding is only applicable to dished sheet and 3D hollow.
CDs could be considered to be “flat sheet”, but the exclusion of injection molding in the matrix is associated
with larger working areas of flat material. Injection molding is clearly an option, via the compatibility with “3D
solid”.
From Figs. 18.9 and 18.10, injection molding can achieve a surface finish close to 0.1 µm, but the best
tolerance is usually around 100 µm – so much too imprecise for the features on the CD.
The problem of precision is overcome in practice by placing a pre-patterned metal stamper plate on one face
of the injection molding die, carrying the pattern of precise bumps needed to print the pits on the CD. There
are various ways of producing the metal stampers. For example, a glass master disc is coated with a light-
sensitive material (such as a photo-resist), treated by laser pulses, and etched to produce the pattern of pits.
Then the glass disk is metallised with nickel to a thickness of about 1/3 mm, to produce the metal stamper
with the correct pattern of raised bumps. The much greater stiffness and hardness of glass and nickel enable
the necessary precision and finish to be obtained. Injection molding of the polycarbonate discs is then
conducted under normal processing conditions, taking a few seconds.

Exercise E18.4 A process is required for the production of 10,000 medium carbon steel engine crankshafts.
The crankshaft is a complex three-dimensional shape about 0.3m in length with a minimum diameter of
0.06m. At the position of the bearings, the surface must be very hard, with a surface finish better than 2 µm
and a dimensional accuracy of + 0.1mm. Elsewhere the surface finish can be 100 µm, and the tolerance is
0.5mm. Use the process attribute charts and other information in the chapter to identify a suitable fabrication
route for the crankshaft.

2
Answer. See annotated process attribute charts below.
Shape: complex 3D solid – so need a casting process, or forging, or powder methods.
Material: medium carbon steel – some casting processes (and extrusion).
Economic batch size: sand casting, forging and power methods all appear economic for 10,000 units.
Mass: very approximately a long cylinder 30cm × diameter 6cm.
Volume = π × (0.03)2 × 0.3 = 8.5 × 10-4 m3, so for density of steel = 7.9 Mgm-3, mass ≈ 6.7 kg.
All processes viable at this mass.
Minimum Thickness: 6cm
Sand casting and forging, are the only options that also work with steel; too large for powder methods.
Tolerance: + 0.1mm on critical surfaces, + 0.5mm elsewhere.
Precision is marginal for even the best net-shape processes, but the higher quality tolerance target is not
achievable by forging or sand casting – hence require post-shaping machining process to obtain target
precision on the bearing surfaces.
For the rest of the component, forging can achieve the target precision – sand casting would require all-round
machining to meet the specification, adding significantly to costs.
Roughness: 2µm finish beyond the capability of forging, so machining also needed to give the target finish on
the bearing surface (possibly with a final grinding/polishing step). Elsewhere all processes achieve the
required finish.
Final selection: forging offers the best option, with machining of the bearing surfaces. Note that this gives the
additional benefit of a wrought microstructure (as opposed to a cast microstructure). Component is likely to
be hot forged (to reduce forming loads) and to have a further heat treatment (e,g, quench and temper) to give
the best combination of yield stress and fracture toughness. For the bearing surfaces, there is also the
possibility of a surface hardening treatment (provided that this does not change the dimensions or the surface
finish) – e.g. carburising, transformation hardening etc.

3
Exercise E18.5 A heat-treatable aluminium alloy automotive steering arm has a mass of 200g, a minimum
thickness of 6mm and a required tolerance of 0.25mm. Production runs over 50,000 are expected. Discuss
the implications for the manufacture of the component by forging, using the process attribute charts. Suggest
an alternative processing route, commenting on how this might influence the choice of Al alloy and final
properties.

Answer. See annotated charts above.

Shape: complex 3D solid – so need a casting process, or forging, or powder methods.
Material: aluminium alloys – no restriction on shaping process.
Economic batch size: casting (sand, die, low pressure), forging and power methods all appear economic for
50,000 units.
Mass: 0.2 kg. Too small for sand casting, other processes remain viable.
Minimum Thickness: 6mm. Marginal for sand casting, other processes OK.
Tolerance: + 0.25mm.
Sand casting unable to achieve target precision, other processes OK. For the complex 3D shape it is unlikely to
be viable to use machining over the whole surface to improve the precision.
Conclusion: forging passes all stages, as do die casting and low pressure casting. No surface finish target was
provided but, as for many moving parts in an automotive application, fatigue will be limiting so a good finish is
needed. The roughness chart shows that both casting options are as good as forging, with die casting the best.
Choosing either casting process would require a change to an Al casting alloy. The challenge will be in
matching the key material properties (strength, fatigue resistance and fracture toughness) – a heat-treatable
casting alloy will be required.

4
Exercise E18.6 A small polyethylene bucket is to be manufactured by injection moulding or rotational
moulding. The designer wishes to estimate the manufacturing cost for various batch size, using the cost
equation
C dedicated C capital + C overhead
Cost per part C = C material + +
n n
Define the meaning of all of the parameters in the equation. Using the data in the Table, determine the
cheapest process for batch sizes of 1000 and 50,000.

C material C dedicated C capital + C overhead n

(£) (£) (£/hr) (parts/hr)
Process
Injection moulding 0.25 5000 20 120
Rotational moulding 0.20 1000 20 40

Answer.
C material is the cost of the material used per part (allowing for any losses)
C dedicated covers the cost of tooling (dies, moulds, fixtures) assigned wholly to this part
C capital is the capital cost of equipment, spread over a suitable write-off period to give a cost/unit time

C overhead covers all other background hourly operational costs (labour, energy and so on)
n is the batch size or production volume (the number of parts in the production run)
n is the production rate

C material C capital + C overhead n C dedicated Total cost per part

n n C
Process (£) (£) (£) (£)
Injection 0.25 0.17 1000 5 5.42
moulding 50000 0.1 0.52
Rotational 0.20 0.5 1000 1 1.7
moulding 50000 0.02 0.72

Hence rotational moulding is cheapest for 1000 units, but injection moulding cheapest for 50,000.

Exercise E18.7 In a cost analysis for casting a small aluminium alloy component, costs were assigned to tooling
and overheads (including capital) in the way shown in the table below. The costs are in units of the material
cost of one component. Use the simple cost model presented in equation (18.7) to identify the cheapest
process for a batch size of (i) 100 units and (ii) 106 units.

Process Sand Investment Pressure Gravity

Casting casting die die
Material, Cmaterial 1 1 1 1
Capital + Overhead, C capital + C overhead (hr-1) 500 500 500 500

Dedicated, Cdedicated 50 11,500 25,000 7,500

Rate n (hr )
-1
20 10 100 40

5
Answer.
material cost dedicated cost capital cost overhead cost
↓ ↓ ↓ ↓
C dedicated  
C capital + C overhead
Cost per part C = C material + +
n n
↑ ↑
no. of parts production rate

500 50 50
Sand Casting: C = 1+ + = 26 + ,
20 n n
so for n = 100, C = 26.5; for n = 106, C = 26
500 11,500 11,500
Investment Casting: C = 1+ + = 51 + ,
10 n n
so for n = 100, C = 166; for n = 106, C = 50
500 25,000 25,000
Pressure Die Casting: C = 1+ + = 6+ ,
100 n n
so for n = 100, C = 256; for n = 106, C = 6
500 7,500 7,500
Gravity Die Casting: C = 1+ + = 13.5 + ,
40 n n
so for n = 100, C = 88.5; for n = 106, C = 13.5

(i) for n = 100 units choose sand casting

(ii) for n = 106 units choose pressure die casting.

Exercise E18.8 Look at the products around your house (or your garage, or garden shed), identifying as many
different types of joints as you can, and the process used. Select a few examples, and draw up a list of design
requirements that the joining process needed to satisfy.

Sample answer.
Bicycle: welded or brazed tubes; nut-and-bolt attachments (brakes, mudguards, accessories); elastic spring
clips (light fittings); rivets (mudguard stay brackets); adhesives (handlebar tape).
Examples of other types of joints: screws, adhesives, slotted press-fits (wooden furniture); screw threads (light
bulb fixtures); mechanical clinching and seam welds/spot welds (radiators).
Design requirements for bicycle joints:
- materials to be joined, and whether similar/dissimilar (e.g. alloy steel tubing; Al alloy crank onto carbon steel
axle)
- shape of parts (intersecting tubes, or press-fit of rectangle into a hole)
- joint geometry (inclined butt joint of welded tubes, or sleeve joint of one tube within another in brazed
tubes)
- material thickness
- need to disassemble (permanent for welds, reversible for bolted fixtures/accessories – a particular issue for
dissimilar alloys, which may seize due to corrosion)
- weld quality (effect on properties, e.g. welds can influence fatigue strength)

Exercise E18.9 Look at the products around your house (or your garage, or garden shed), identifying as many
different types of surface treatments as you can. Select a few examples, and draw up a list of design
requirements that the surface treatment process needed to satisfy, in particular the function of the surface
treatment.

6
Sample answer.
Anything painted – woodwork, vehicles, bicycles….. (both decorative, and for corrosion resistance).
Polymer powder coating on metals: hi-fi and computer units, filing cabinets, kitchen racks and shelf brackets.
Chrome plate (and other metal finishes) on steel: bath and kitchen fixtures, handles, automotive trim.
More difficult to identify are surface hardening treatments – e.g. cutting edges on steel tools, clamps on vices
– but most of these are ‘hidden’ (sliding surfaces in gears, bearings, pistons etc).
Design requirements for the edge of a steel cutting tool:
- function of treatment: hardening (for wear resistance)
- material compatibility: depends on detail of which steel, and process mechanism (e.g. whether hardenable or
not)
- retention of bulk properties (in particular, avoidance of embrittlement)
- part geometry and thickness, and proportion of surface to be treated
- dimensional precision/finish: does treatment change these?

Exercise E18.10 A capstan is a cylindrical device for winding in ropes and lines on a yacht, using friction to
reduce the force needed to hold or take in the rope. An aluminum capstan is to be manufactured by sand
casting. The figure shows two suggested cross-sections.

(a) For either geometry, where would you need to place the parting plane for this casting?

(b) Identify which of the two designs is preferable, giving three reasons for your choice.

(c) The central hole could be manufactured using a core, embedded in the mold and extracted from the
casting after the mold has been removed. Sketch how the core could be supported with respect to the
parting plane.

Sample answer.
(a) The parting plane would cut the capstan through the centre on its vertical axis – in practice, it would be
cast on its side, with a horizontal parting plane.
(b) The LH design is preferable due to:
- the taper angle on the flanges makes it easier to remove the pattern from the sand mold, without
damaging the mold;
- the radius on the corners improves metal flow, avoiding turbulence and mold damage;
- the stress concentration at the corners is lower, reducing
the risk of hot tearing during cooling.
(c) The core lies on the parting plane, so is easily fixed in the
mold: a rod longer than the capstan would be embedded in
the and at the top and bottom of the cavity, as shown in
the sketch.

Exercise E18.11 The capstan in Exercise E18.10 is 150mm in height, with an average outer diameter of
110mm. By approximating the casting to a solid cylinder with these dimensions, estimate the solidification
time ts for the component using Chvorinov’s rule:
2
V 
ts = C  
 A
where V and A are the volume and surface area respectively, and the constant C for sand casting of
aluminium has the value 1.0 s/mm2.

7
Answer.
Working in millimetres:
π D2 π (100)2
Volume: V = ×L = × 150 = 1178 × 10 3 mm 3
4 4
π D2 π (100)2
Surface area: A = 2 × + πDL = 2 × + π × 100 × 150 = 62.8 × 10 3 mm2
4 4
2 2
V   1178 
Hence solidification time is approximately: t s = C   = 1.0 ×   = 352 s, about 6 minutes
 A  62.8 

Exercise E18.12 For each of the following shapes, find an approximate expression for the ratio of the volume
to the surface area, V/A:
(i) a sphere of radius R;
(ii) a cylinder of radius R, and length L (where L >> R, so the end area may be neglected);
(iii) a rectangular plate of area L x B, and thickness T (where T << L and B, so the edges may be neglected).
Use these results to explain a simple physical basis for Chvorinov’s rule (see Exercise E18.11), in terms of heat
flow. How do you expect Chvorinov’s constant C to depend on the thermal diffusivity of a cast alloy?

Answer.
For sphere of radius R:
V = 4/3 πR3, A = 4πR2, so (V/A) = R/3
For cylinder of length L (>> R):
V = L πR2, A = L (2 πR), so [neglecting the ends]: (V/A) = R/3
For rectangular plate of thickness T (<< L, B):
V = L B T, A = 2 L B, so [neglecting the edges]: (V/A) = T/2

In each case the ratio (V/A) is equal to a fraction (1/3 or 1/2) of the shortest dimension of the part. This
reflects the dominant heat conduction distance from the centre of the part to the surface. Note that the three
shapes cover 3D, 2D and 1D heat flow respectively.

Transient heat flow theory (Chapter 12) shows that cooling time ∝ (heat flow distance)2, i.e. from x ≈ at
where a is the thermal diffusivity of the material. Comparison with Chvorinov’s rule shows that the constant
C scales with the reciprocal of the thermal diffusivity.

Exercise E18.13 A hot forged aluminum hub for a bicycle wheel is shown in cross
section in the figure, before the central hole for the axle is drilled through. Give
two reasons why this cannot be made in one forging step with a pair of closed dies,
starting from a solid cylinder. Sketch a two-step forging process using multiple dies
acting vertically and horizontally. Search online for ‘forging bicycle hubs’ to see
how it’s done in practice.

Answer. The hub has re-entrant features, so there isn’t a single ‘parting plane’ that allows the dies to
separate. A second problem is the non-uniform cross-section, which requires a first step to re-shape the
preform into a ‘dumb-bell’ shape with the material redistributed
towards the ends.
So the solution is first to forge the cylinder into a ‘dumb-bell’ shape,
then forge the mid-section and flanges (with the dies marked 1 in
the figure), and then upset the two ends simultaneously using dies
moving horizontally (2 in the figure). By profiling the upsetting dies,
the re-entrant features can be pressed in., while increasing the
pressure to ensure the flange cavities are fully filled.

8
Exercise E18.14 In open die forging of a long rectangular slab, of height 2h and width 2w, with parallel flat
dies, the pressure distribution on the tool-workpiece interface is given by
  w − x 
p (x) = σ y  1 + m  
  h 
  
where σy is the material yield stress, and m is a friction factor (= 0 for well-lubricated, up to 1/2 for sticking
friction). (See Fig. 18.25 in the text).
(a) For a friction factor m = 0.2, find the maximum pressure pmax at x = 0, for values of w/h = 1, 2, and 5, as a
multiple of the yield stress, σy .
(b) For each case in part (a), find the % increase in the total forging load, compared to the frictionless case,
when m = 0 (consider unit length of the slab in each case).

Answer.
(a) For m = 0.2, substituting x = 0 and values for w/h gives:
w/h = 1, pmax = 1.2 σy
w/h = 2, pmax = 1.4 σy
w/h = 5, pmax = 2.0 σy
(b) The load (per unit depth) is the area under
the friction hill. For the frictionless case, the load
per unit depth is equal to 2w x σy , regardless of
the aspect ratio w/h. For a linear friction hill, the
per unit width is the average pressure over the
die times the width 2w. The limiting values are p
= σy at the edges and p = pmax at the centre. So
the average pressure pave = (σy + pmax)/2. Hence:
w/h = 1, pave = 1.1 σy
w/h = 2, pave = 1.2 σy
w/h = 5, pave = 1.5 σy
Since the load scales with 2w in each case we can simply compare these values with the (constant) pressure in
the friction case, σy. The % increases in load are therefore 10, 20 and 50%, for w/h = 1, 2, and 5 respectively.

Exercise E18.15 The stepped component shown in the figure is to be manufactured by uniaxial compaction of
a ceramic powder, using a two-part die set of the type shown on the right of Figure 18.26. After mold filling
the powder has a packing density of 50%. The finished component is
considered to be of acceptable quality by compacting to a final
porosity of 3%. Find the dimensions of the filled cavity prior to
compaction which will produce a uniform density throughout the
part.

Answer.
The vertical dimension of the part scales inversely with its relative density (as the width is fixed across the two
parts of the component). The final relative density is 97%. Hence the depth of powder in each half of the
mold needs to be 97/50 x final vertical dimension. The filled cavity depths are therefore 38.8 and 23.3 mm.

Exercise E18.16 Look closely at a number of polymer bottles with threaded caps. The pre-form is likely to
have been injection molded, and the bottle then blow molded. Look for a parting line showing the location of
the parting plane when the preform was made. How was the thread manufactured on the bottle? Could the
thread in the cap be made the same way, and if so, how was the mold removed? Can you identify geometric
features of the bottle and cap that add functionality, and have been produced directly by detailing the mold
shape? Can you find similar features on any glass bottles?

9
Answer.
Parting line show up as small ridges running vertically down the bottle, located diametrically on opposite sides
of the bottle. For a bottle with an integral handle (such as a polyethylene milk bottle), the handle lies in the
same plane. You may see that the partlng plane is prominent in the thread region, but has almost disappeared
in the main body of the bottle – blow molding stretches and thins the wall material.
The thread was formed directly on the injection molded preform, by machining carefully aligned spiral grooves
into the molds. This might not seem possible for the caps – how can you pull the internal mold out when the
grooves are full of polymer? The solution is simply to rotate the mold halves as they separate, unscrewing the
cap from the mold.
There are many geometric features produced directly by molding bottles and their caps:
- vertical ribs on the caps to improve grip
- tamper-evident closures, with a constrained ring attached to the cap by fine ligaments that can be broken by
twisting the cap
- the base of the bottle is profiled to give a hollowed out shape to enable the bottle to stand stably
(with the same being seen underneath any glass bottle)
- integral handles may be formed (mile bottles), or the bottle profiled to help with handling (e.g. detergent)
- characters and logos molded directly into the surface: routinely used to add one of the standard recycling
codes to polymer products (also used for the glass bottles of selected expensive wines, as form of branding).
- Braille characters, particularly as a safety feature for the visually impaired on hazardous liquids.

Exercise E18.17 A restorer of musical instruments needs to replace 5 keys on an antique clarinet. All five have
different shapes with complex curved surfaces, and these have been 3D-scanned into a CAD package from an
equivalent instrument. The keys are made of a Cu alloy and weigh between 10 and 20g. The minimum section
thickness is 3mm, and the dimensional tolerance is 0.2mm. The keys will be electroplated and polished to give
a shiny, silver finish, so the roughness after shaping is less important: 20µm is the suggested target.
The current capabilities of the additive manufacturing methods of powder bed fusion are summarised in Table
18.14. Assess the design requirements for the clarinet keys against the technical, quality and economic data
for powder bed fusion, and compare the suitability of the AM route with investment casting, using the
attribute charts in Chapter 18. Identify another shaping process that would be viable, and comment on the
competition with powder bed fusion.

Answer.
The table summarises the match between the design requirements and the attributes for powder bed fusion
and investment casting processes.
Attribute Powder bed fusion Investment casting
Material compatibility Suitable for Cu alloys Suitable for non-ferrous alloys
Shape compatibility Suitable for 3D shapes Suitable for 3D shapes
Mass range 10-20g right at the bottom 10-20g right at the bottom of the range
of the range (0.01 – 20 kg) (0.01 – 20 kg)
Minimum section 3mm within range 3mm within range
thickness (0.8 – 100 mm) (1.5 – 75 mm)
Tolerance 0.2mm just within range 0.2mm within range
(0.2 – 0.8mm) (0.1 – 0.4mm)
Roughness 20µm within range 20µm easily achieved
(8 – 125µm) (normal range 1.6 – 3.2µm)
Economic batch size OK, as only economic for There are 5 keys, but all are different, so
fewer than 10 batch size effectively 1 – marginal for
this process (charts show 2-1000)

Both processes can deliver all of the target attributes. In both cases it is helpful to have the required shape in
CAD, for programming the powder bed machine, or for manufacturing the mold and pattern for investment
casting. One-off shapes are not normally investment cast, so this is a promising application for powder bed
fusion.

10
The only other processes that are economic for one-off shapes are sand casting and machining (Figure 18.13),
but sand casting is not suited to small delicate shapes under 1kg. Machining from solid is therefore an option,
and can provide the target precision and finish. It is also beneficial in machining to start from the CAD file –
but the complex curved surfaces make this a challenge for machining, and the process will be slow as well as
wasting much of the initial stock material. So the additive manufacturing process may well compete
economically.

Exercise E18.18 A student wishes to produce 10 identical prototype polymer housings for an electronic
device, and has access to a photo-polymerisation system. The target shape is a hollow two-part shell about 3-
4mm thick. Before proceeding with the design, she wishes to check whether there are other polymer shaping
processes available that should be considered. Which other polymer shaping processes may be economic for a
batch size as low as 10? Are they also compatible with the overall shape and section thickness? Compare the
relative performance of the conventional and AM processes for tolerance and roughness, based on the current
capabilities of photo-polymerisation summarised in Table 18.14, and the attribute charts in Chapter 18. The
student wishes to incorporate shaped features into the housing, for mounting the device securely inside, and
to facilitate closure of the housing with screws. Would this be easier with AM or with the competing
conventional processes?

Answer.
Figure 18.13 suggests that thermoforming or polymer casting may be viable for a small batch size of 10.
Thermoforming makes hollow shapes in sheet material, with 3-4mm well within range; polymer casting is used
for 3D solids and hollow shapes, but generally in greater thicknesses, so it is marginal. The tolerance ranges
for the processes are very similar:
Photo-polymerisation: 0.1-2mm, thermoforming: 0.2-1mm, polymer casting: 0.8-2mm
The AM method appears to be inferior in terms of roughness however: 100-125µm, compared to 0.5-1.6µm
for thermoforming and polymer casting. The importance of surface finish should therefore be considered
carefully (e.g. for mating surfaces between the two parts of the housing, and for its appearance).
Internal and external features for mounting and joining locations are the kind of detail associated with
injection molding (which is not economic for only 10), but they should be possible with polymer casting. They
are more difficult in thermoforming (which starts from a flat sheet and draws it into a profiled mold). Additive
manufacturing will be well-suited however to adding this shape detail.

11