Surface Integrity Cutting Fluids Machining and Monitoring Strategies_1 pdf

9.1.1 HSM Machine Tool Design Considerations Prior to a discussion concerning machine tool design factors that must be addressed, before to fully-imple-menting this HSM technology, one

9

Machining and Monitoring Strategies

‘Inque brevi spatio mutander saecla animantum

Et quasi cursores vitai lampada tradunt.’

TRANSLATION:

‘The generations of living things pass in a short time and like runners hand on the torch of life.’

Titus LUCRETIUS Carus

(94 – 55 BC) [In: On the Nature of the Universe, II]

9.1 High Speed Machining

(HSM)

Background to HSM

Possibly the first work of note on HSM was that of

Sa-lomon who ran a series of applied experiments from

the period 1924 to 1931, when a German Patent was

granted for this work The Patent was founded upon a

series of cutting speed curves plotted against

machin-ing temperatures for a range of materials (Fig 214) In

these tests Salomon achieved peripheral cutter speeds

of 16,500 m min–1, utilising either fly-cutters – for chip

morphology data, or helical milling cutters – notably

when cutting aluminium Salomon contended that

the cutting temperature ‘peaked’ at a specific cutting

speed, which he termed the ‘Supercritical speed’1,

fur-ther, when the speed was increased still faster the

tem-perature he noted, decreased Moreover, either side of

this ‘supercritical speed’ zone, it was suggested that

‘unworkable regimes’ occurred where the HSS

cut-ters could not withstand the severe forces and

tem-peratures generated As mentioned, when the cutting

speed increased beyond this ‘zone’ the temperature

reduced to those expected by ‘normal data’ cutting

conditions, permitting practical cutting operations to

be carried out The problem being with this early HSM

work is that any theoretical rationale is not available

and the experimental procedures are somewhat

un-clear, but Salomon can be considered the founder of

ultra-high speed machining (UHSM) – taking cutting

 ‘Supercritical speed’ mentioned by Salomon when HSM, has

never been truly substantiated The practical data was based

upon chip morphology* experiments with ‘fly-cut‘

climb-mill-ing of: non-ferrous alloys; ‘soft’ aluminium; red cast brass; the

latter being ‘un-machinable’ with HSS cutters between 60 to

330 m min–1 (i.e see Fig 214) As the machining parameters

and experimental details only exist as a partial fragment of the

original German work This fragmented machining

informa-tion, is because during the Second World War, unfortunately,

the vital details were lost, moreover, none of the participants

in this work also surviving to explain how this applied

re-search data was collated.

*Chip morphology was achieved by situating a heavily

wax-coated board, this being strategically positioned to allow the

fly-cut chips to stick to the board – during the peripheral

climb-milling experimental operation, ready for future

analy-ses.

data beyond that considered in the so-called ‘Taylor-equations’

Effectively it can be established that there were four

distinct periods of advancement in the field of UHSM, with the first one being from the early 1920’s to the late 1950’s, with each period during this time and there-after, being separated by a significant event Obviously during the first period, the work instigated by Salomon (i.e in the 1920’s – alluded to above), was followed by the originally-commissioned United States Air Force (USAF) major research contract, being from 1958

to 1961 Previous to this contract award, little in the way of UHSM work in the preceding decades had oc-curred, apart from that in the States by Vaughn (1958) Vaughn, shortly became aware of the Salomon Patent, acquiring limited information on this work through the United States Consul in Berlin Vaughn’s (Lock-heed) group were also familiar with the many tech-nical references concerning the ‘art’ of oil well tube

perforation utilising perforated cutters, these ‘cutting

actions’ being employed to perforate oil well casings

at explosive speeds Such background work, now meant

that Vaughn had ‘set the scene’ for the second period

of UHSM development

The USAF Materials Laboratory commissioned study – mentioned above, which was awarded to Lock-heed (i.e Vaughn’s group), with the objective of evalu-ating the ‘machining response’ for a selected range of high-strength materials to that of (surface) cutting speeds of up to 152,400 m min–1 The ‘principal aim’

of this ‘USAF-commissioned work’ was to increase

producibility, while improving both the quality and

efficiency of the fabrication of aircraft/missile com-ponents Vaughn’s experimental apparatus included the use of modern-day ex-military cannons, which were positioned on ‘railed-sleds’ to minimise subse-quent recoil upon firing, while obtaining the desired ballistic exit velocity of the ‘material-slug’ for the cut-ting speeds Unfortunately, these ‘machining results’

 ‘Ballistic cutting speeds’ , were obtained as the cannon fired

the projectile (i.e a specific material-slug) at ultra-high speed out of the cannon At the projectile’s exit from the cannon’s barrel, a very robust cutting tool arrangement was situated and held in an accurate position to take a linear cut along the exiting slug’s periphery The evidence of this ballistic cutting action was then recorded by very high-speed photographic equipment – this being both electronically-timed and stra-tegically positioned, for visual dynamic recording and future reference and analysis.

Figure 214 Graphical relationship of high speed machining of metals – according to Salomon’s machining trials [Source:

Salomon, circa 1920’s–30’s]

.

did not indicate how such ballistic speeds might be

in-corporated into a production application, also, an

ana-lytical model of this high-rate cutting phenomenon

was not developed

In the 1960’s and early 1970’s some consolidation

of important machining data occurred, with notable

work on UHSM cutting mechanisms, etc., that

oc-curred in various countries being led principally by

the USA in work from: Coldwell and Quackenbush

(1962), plus Recht (1964); Okushima et al (1966) and

Tanaka et al (1967) in Japan; Fenton and Oxley (1967)

in the United Kingdom; and Arndt (1972) in Australia

However, although there was a general increase global

research activity during this period, it was somewhat

disparate and mainly of a fundamental, rather than

ap-plied nature

The third period of HSM development was

insti-gated in the mid-to-late 1970’s by the United States

Navy, in conjunction with the Lockheed Missiles and

Space Company, Inc., who contracted a series of

ma-chinability studies on marine propellers Here, the

Lockheed group headed-up by King, were mainly

concerned with the feasibility of utilising HSM in a

production environment, primarily for aerospace

alu-minium alloys, then later, working on

nickel-alumin-ium-bronze King’s team at Lockheed, demonstrated

that it was economically feasible to introduce

‘high-speed-machining’ procedures into the production

en-vironment, thereby realising the significant

improve-ments in productivity with this HSM application Such

applied research work, promoted significant activity

and interest in this HSM field and, it soon became

clear that attention needed to be focussed for all of the

subsequent small and diverse research programmes

 ‘Marine propeller manufacture’ , whether from a

wrought-solid material, or finishing-off a casting, is probably the most

difficult and complex multi-axes milling operation that can

be undertaken – due to the fact that the propeller surfaces

to be machined are continuously changing their geometry as

they are essentially parabolically-curved Invariably, the part

geometry is typified by say, the NACA/NASA Standard 16-021

‘aerofoil cross-sectional profile’ , which for high-performance

propellers are exacerbated by normally having both

consider-able rake and skew to each blade – creating an exceedingly

complex geometry and fillet where the boss and blades

inter-sect*.

* See: Smith and Booth (1993) paper – in the references,

which goes some way to explain the propeller manufacturing

subject, and for more detailed multi-axes machining and for

subsequent machined propeller measurement information.

Finally, in these formative years of experimental work into HSM, the fourth period of development be-gan in 1979, when the USAF awarded a contract to the General Electric Company, in this instance, to provide

a scientific basis for faster metal removal via HSM and Laser-assisted machining A further contract by the USAF in 1980, was also awarded to ‘General Electric’– the group also being headed by Flom (1980) With this new HSM contract, also being granted to Flom’s group, with the objective to evaluate the production implica-tions of the previous contract Both of these contracts being supported by a consortium of industrial com-panies and selected universities in the USA – initiat-ing the fourth HSM period of development At around this time (i.e in the late 1970’s), the introduction of computer numerical control (CNC) systems occurred and as a result, they were immediately being fitted to

a new range of machine tools, significantly enhanc-ing both their usability and programmenhanc-ing capabilities, acting as a catalyst in the development of HSM strate-gies As these CNC controllers became more sophis-ticated and processing speeds substantially increased, this meant that the potential for HSM could now be fully realised In recent years, HSM machine tools are just about everywhere in machine shops around the world, where ever there is a need for highly produc-tive part production with very fast cycle times Obvi-ously, on HSM machines as the spindle speeds have increased in association with both tool and their re-spective workpiece materials (i.e see Fig 215), this has meant that there are now considerable design implica-tions on these machine tools, these topics will now be succinctly discussed

9.1.1 HSM Machine Tool Design

Considerations

Prior to a discussion concerning machine tool design factors that must be addressed, before to fully-imple-menting this HSM technology, one might ask the

question: ‘Why do we need to rotate cutters at such high

spindle speeds?’ There are a number of advantages that

can accrue from adopting such a progressive

produc- USAF contract to the General Electric Company in 1979, was:

F 33615-79-C-5119 – for a feasibility study into the fast metal removal operations by HSM and Laser-assisted machining.

tion machining strategy (i.e see Fig 216) and, they

can be succinctly summarised as follows:

• Direct benefits

– improved machining efficiency,

– reduction in cutting tool varieties,

– reduction in distortion of workpiece,

– effectiveness of swarf removal

• Indirect benefits

– quality of finish improved,

– increased cutter life,

– reduced changes in material properties,

– capability of machining thin walls/sections

These production improvements are by no means all

that occur, as invariably, due to the superior machined

surface texture, the final part surface may not need

to be deburred – a significant real saving on complex

component geometries Moreover, by employing an HSM strategy, more simple fixturing can be utilised, as the actual tool forces are significantly lessened

It is an established fact that with higher rotational cutter speeds the resulting cutting forces and tool push-off are considerably reduced In order to benefit from these improved cutting practices, the machine tool’s axes must have both faster acceleration and de-celeration – see Fig 221, more will be mentioned con-cerning these important dynamic aspects of a machine tool’s operational performance shortly Many of today’s conventional and HSM machine tools, are based upon

a modular design concepts (Fig 217a) This modular design philosophy, allows the machine tool builder

the opportunity to standardise certain features over

a range of machine tools Such practice benefits the manufacturer and consumer alike, by reducing design

Figure 215 The chronological development of cutting tool material introductions, which had an influence on high-speed

cut-ting trends [Courtesy of Yamazaki Mazak Corporation]

.

and development costs, while minimising the

custom-er’s purchasing costs, yet still allowing more attention

to be given to the detailed design of each ‘module’ in

the machine So, an identical column, or table may be

common to a variety of machines and this trend can

often be seen across a whole product range of

ma-chine tools Not only are the major castings, or large

fabrications manufactured by employing modular

concepts, but this design philosophy also allows any

other smaller components to be standardised and

fit-ted accordingly Such as: the recirculating ballscrews;

servo-motors; linear scales – if fitted; etc.; plus other standardised items to be built into the constructed machine tool ranges

In order to minimise ‘stick-slip’  in the slideway

mo-tions on heavy moving cast and fabricated members –

 ‘Stick-slip’ or ‘Stiction’ , is the jerky-motion between sliding

members due to the formation and destruction of junctions [due to localised pressure-effects] (Kalpakjian, 1984)

Figure 216 Diagram illustrating the main benefits to be gained from adopting a high-speed

ma-chining strategy [Source: Smith, CNC Mama-chining Technology, 1993]

.

thereby giving faster response to CNC commands,

‘Ty-choways’ (i.e see Fig 217b), or similar mechanical aids

are usually strategically situated at set positions along

each hardened way of the machine tool’s orthogonal

axes (i.e normally at the ‘Airy-points’ )

 ‘Airy-points’ , are associated with any ‘elastic body’ that is

subject to load, where will undergo elastic deformation The

magnitude of this deformation depends upon the extent of

the: load; contact area; plus the mechanical properties of the

contacting materials Hence, if the prospective body is not

correctly supported, it will be subject to elastic deformation

– under its own weight This problem was considered in 1856

by Sir G.B Airy (Astronomer Royal) – who was interested in

minimising the elastic deflections that arose when

attempt-ing to support and reduce the saggattempt-ing in long focal length

refracting telescopes Airy showed that by positioning these

two supports at prescribed distances, they could minimise

any potential error – when equated to the overall length of the

‘elastic member’ (i.e in the length of the telescope) He found

that two conditions occur, one being for ‘line-standards’ – to

bring the ends square for measurement and calibration, and

secondly, relating to machine tools and many metrological

applications, where the ‘Points of minimum deflection’ were

more appropriate, as follows:

Distance between each support = 0.554L

Using conventional recirculating ballsrews (Fig 217c) for HSM applications is possible up to 100 m min–1 – in certain applications After these linear velo-cities have been reached and it is necessary to reverse the axis direction, this can create ‘ballscrew wind-up’ problems This ballscrew twisting, is despite the fact that the ballscrew is very stiff of the order >2,000

N µm–1, so if greater kinematic and dynamic perfor-mance is required, then it might be necessary to utilise linear-motor drives A tabulated table for suitable comparisons of the various types of motional drive systems available today, is presented in Table 14

It has been widely-reported that either a high-qual-ity lead-, or ball-screw having rotary encoders, will have a unidirectional repeatability to within

6-to-Where: L = overall length of the ‘elastic member’ (Galyer and Shotbolt et al., 1990)

Example: If a machine tool’s structural moving member is 1,000 mm in length and it is situated on a much longer ‘bed-way’ , then the ‘Tychoways’ should (ideally) be symmerically-positioned (i.e fixed to the moving member) at a distance of

554 mm apart – in order to obtain the minimum of elastic dis-tortion as it moves backward and forward along the hardened bedway

Table 14 Comparison of different drive systems for machine tools

CONTRIBUTIONS: Leadscrew: Ballscrew: Belt-drives: Linear motor:

Noise Quiet Noisy Quiet Moderate

Back-driving Self-locking Easyback-drive

Backlash Increases with wear Constant Increases with belt wear Negligible

Repeatability ±0.005 mm ±0.005 mm ±0.004 mm Best (<2 µm)

Duty Cycle Max 60% Max 100%

Efficiency Bronze bushing 40% 90% 90% 90 to 95%

Life Short: high friction Longer Longer Longest

Shock-loads Higher Lower Low Highest

Smoothness Smooth: low speeds Smooth at all speeds Smoothest

Speeds Low High Higher Highest

Cost ££-Lowest £££-Moderate ££££-Highest

[Source: Johnson, 2001]

.

Figure 217 Typical ‘modular design’ and construction of machining centres, with ‘ballscrews’ and ‘tychoways’ [Courtesy of

Cin-cinnati Machines]

.

7 µm However, if linear encoders are fitted this

mini-mises any potential ‘ballscrew wind-up’ , although the

problem of an ‘Abbé -error’  is still present

Yet another source of machine tool-induced

prob-lems are more specifically termed ‘uncertainties’ ,

while ‘hysteresis’  can also contribute to the overall

‘er-ror budget’ Thus, hysteresis may result when the same

position has been commanded by the CNC controller,

 ‘Abbés Principle’ – was derived by Professor Ernest Abbé in

1890 (i.e having studied and graduating from the University

of Jena) and is still valid today The Abbé principle simply

states that: ‘The line of measurement and the measuring plane

should be coincident’ An example of this ‘Abbé Law’

well-known to engineers the world over, shows that a conventional

micrometer calliper ‘virtually-obeys’ the ‘Abbé principle (i.e

there is a small ‘cosine error’ present – which can usually be

ignored), whereas, the Vernier calliper has a much larger

off-set between the fixed and moving jaws – where the

compo-nent being measured is situated, to that of the beam – where

the scale’s reading for this measurement is taken Ideally, both

measurement and reading should be lined-up, without an

off-set [Sources: Busch, 1989; Whitehouse et al., 2002]

 ‘Uncertainty of machine tool positioning’ – the question often

asked in calibration-related tasks, is: ‘What is measurement

uncertainty?’ Uncertainty of measurement refers to the doubt

that exists about any measurement; there occurs a margin of

doubt for every measurement This expression of

measure-ment uncertainty raises other questions: ‘How large is the

mar-gin?’ and ‘ How bad the doubt?’ Hence, in order to quantify

uncertainty two numbers are required: (i) being the width of

the margin – its interval, (ii) plusits confidence level.

NB This latter value states how sure one is that the actual

value occurs within this margin.For example:On a CNC

ma-chine tool, the command may be to move the X-axis

slide-way 1000 mm plus, or minus 0.05 mm at the 95% confidence

level This uncertainty could be expressed, as follows: X-axis

slideway motion = 1000 mm ± 0.05 mm, at a level of

confi-dence of 95% In reality, what this statement is implying to

either the programmer, or calibration engineer is that they are

95% sure that the actual X-axis position now will lie between:

999.95 mm and 1000.05 mm Many factors can contribute to

the overall ‘error budget’ as it is sometimes known, but this is

beyond the scope of the present discussion – see references for

further information (Smith, 2002)

 ‘Hysteresis’ , can be defined as: ‘The difference in the indicated

value for any particular input when that input is approached in

an increasing input direction, versus when approached in a

de-creasing input direction.’ (Figliola and Beasley, et al 2000)For

example:Hysteresis usually arises because of strain energy

stored in the system [i.e in this case, within the machine

tool], by slack bearings, gears, ballscrews, etc (Collett and

Hope, 1979)

but from opposite directions, causing the motion sys-tem to creep by an amount larger than the backlash alone (i.e the hysteresis) This effect is the result of un-seen working clearances and compounded by the ma-chine’s elastic deformations, although by pre-loading the ballscrews it will minimise some of these effects

In HSM machining applications, all ballscrew and

indeed any screw-driven systems have some additional limitations, such as its ‘critical speed’ of rotation At the

critical speed, a ballscrew starts to resonate10 at its first

natural frequency (i.e termed ‘whipping’) Hence, the

critical speed is proportional to the ballscrew’s diam-eter and is inversely proportional to the distance be-tween the screw’s supports – squared For a very long and slender screw-driven machine tool application with wide supports, here, most recirculating ballscrews would have a critical speed of approximately 2,500 rev min–1 It should also be said, that for many ballscrews assemblies they can be rotated at higher rotational speeds than the 2,500 rev min–1 previously mentioned, before they reach their critical speed, but for very fast accelerations and decelerations, then they become in-creasing challenged In fact, on some HSM machine tool configurations, multi-start ballscrews have been employed to increase linear response, but here the

‘critical speed’ will probably be lessened – due to re-duced inherent ballscrew stiffness Even ‘matched’ twin ballscrews have been fitted to HSM machine tools – to minimise any potential ‘yawing motions’ at high lin-ear speed by the moving member along the machine’s bedway These ballscrew limitations are probably why linear-motional drives are becoming a realistic alter-native, but they are only fitted at present, to high capi-tal cost HSM machine tools

One of the bi-products of HSM’s greater stock re-moval rates, is the excess volume of hot swarf which must be speedily and efficiently removed from the vi-cinity of the machine – which is more readily achieved for horizontal machine tool configurations

Ther-mal effects in general on any machine tool become a

problem, particularly as many milling spindles utilise direct-drives, with the motor being mounted in-situ with the spindle Here, the spindle motor creates heat, the thermal effects of which can be analysed by either

0 ‘Resonant frequency’ , can be defined as: ‘The frequency at

which the magnitude ratio reaches a maximum value greater than unity.’ (Figliola and Beasley et al., 2000)

a three-, or five-probe spindle analyser, as depicted in

Fig 218 along with a typical case-study for a spindle’s

thermal drift graph Not only can such spindle analysis

equipment determine a spindle’s thermal growth, it can

also detect: thermal distortion; spindle error; machine

resonance; vibration; plus repeatability In the

case-study graphically-depicted in Fig 218b, the variation

in the machine tool’s temperature is the major cause

of positioning error This thermal drift graph offers-up

a number of significant questions that need to be

ad-dressed, such as: ‘How long does it take for the machine

tool to stabilise?’; ‘How much Z-axis growth does the

machine produce at full spindle speed?’; ‘How far has

the machine’s displacement become, because of

distor-tion in the machine tool structure?’ Spindle analysers

are able to operate at rotational speeds varying from

0 to 120,000 rev min–1, making them an ideal tool for

condition monitoring diagnostics on machine tool’s

equipped with HSM spindles for subsequent analyses

In Appendix 15, a trouble-shooting guide has been

produced to high-light: problems; causes; tests; etc.;

that can be obtained from analyses by employing

ma-chine tool spindle analysers

In Fig 219, the polar plots illustrated show how the

spindle analyser has the ability to evaluate both a new

and rebuilt machine tool spindle’s condition In the

spindle error indicated on the polar plot depicted in

Fig 219a, it illustrates here, that a very badly rebuilt

spindle is simply not acceptable in this state This poor

spindle performance, is in the main, largely the result

of significant radial variability (i.e total error: 12 µm

@ 4,005 rev min–1), which would severely

compro-mise any cutting tool’s machining performance

Con-versely, in Fig 219b, a well-worn spindle assembly is

shown prior to rebuild, having a total error of 4.6 µm

@ 1,702 rev min–1, after its ‘correct’ rebuilding (Fig

219c), the total error has been drastically reduced to

a total error of 1.9 µm @ 1,700 rev min–1 Visually, the

differences between these two polar plots is quite

as-tounding, in that both the asynchronous and average

errors present have virtually disappeared in the latter

case, making it ‘as-new’ and, ready to perform

signifi-cant machining service It is well-known fact by

ma-chinists familiar with their older and heavily-utilised

machine tools, that certain machine spindles will run

smoother and produce better machined roundness on

workpieces if they are run at the so-called ‘sweet-spot’ ,

equally, the quality of the parts produced will be

af-fected if the spindle is run at its ‘sour-spot’ By

utilis-ing a spindle analyser, significant information can be

gleaned from such rotational analyses, allowing speed

ranges to be selected which would currently optimise the present status of the spindle’s condition, prior to its potential rebuild – when called for at a due date in the future

For most machine tools today that are involved in HSM activities, in general the spindle cartridges are

of three distinct design configurations – which

inci-dentally do not normally include ball bearing spindle

types11, such as:

• Magnetic ‘active’ spindles (Fig 220a) – typically

might have a cartridge with a speed range from: 4,000 to 40,000 rev min–1, delivering 40 kW con-tinuous power, peaking at over 50 kW Such ‘active’ magnetic spindles can maintain 1 µm maximum run-out, by digital control of the series of specifi-cally-orientated magnetic currents – being initi-ated by radial and axial sensors, that continuously monitor the spindle’s rotational axis position 10,000 times second–1,

NB Further refinements to the spindle occur, with

these temperature-controlled milling spindle car-tridges maintaining dynamic balance, regardless of the milling cutting loads imposed, this latter fac-tor being an important criterion when attempting

to reduce cutting tool vibrational effects However, these ‘active’ spindles are not cheap to purchase and run, with another negative effect being that they are normally rated for only several thousand hours op-erational running time Such cartridges come with

a variety of rotational speed ranges and spindle power outputs

• Pneumatic spindles – have been available for many

years, with aerostatic bearings equalising the forces exerted while cutting and remaining centralised within the spindle’s housing, yet still achieving dy-namic rotational balance,

 ‘Ball bearing spindle designs’ , are not normally specified for

HSM operations, because at around 20,000 rev min–1 – this be-ing ‘mechanically-set’ as the upper rotational velocity of such spindles So, due to these high rotational speed effects and of the combination of centrifugal forces, it means that at approx-imately 80 m sec-1 rotational speed, the balls will lose contact with the journal walls As a result of this loss-of-contact the hardened balls and raceways will rapidly wear out (i.e the

re-sults of so-called: ‘Brinelling’ in the raceways, creating both

poor circumferential wear patterns, delaminations and associ-ated frictional effects – leading to major debilitating spindle roundness modifications.) (Smith et al., 1992)