Metal Machining Episode 6 pot

is the split-tool method Figure 2.21, although even this is limited – by tool failure – tostudying not-too-hard work materials cut by not-too-brittle tools.Figure 5.12 shows a practical

The need to generate detectable strain imposes a maximum allowable stiffness on adynamometer This, in turn, with the mass of the dynamometer depending on its size or onthe mass supported on it, imposes a maximum natural frequency Simple beam

Fig 5.8 Octagonal ring and parallel beam dynamometer designs: (a) Octagonal ring type tool dynanometer; (b)

paral-lel beam type tool dynanometer

dynamometers, suitable for measuring forces in turning from 10 N to 10 kN, can bedesigned with natural frequencies of a few kHz The ring and the strut types of dynamome-ter tend to have lower values, of several hundred Hz (Shaw, 1984, Chapter 7) These

frequencies can be increased tenfold if semiconductor strain gauges (Ksfrom 100 to 200)are used instead of wire gauges However, semiconductor gauges have much larger driftproblems than wire gauges They are used only in very special cases (an example will begiven in Section 5.2.2) An alternative is to use piezoelectric force sensors

Piezoelectric dynamometers

For certain materials, such as single crystals of quartz, Rochelle salt and barium titanate,

a separation of charge takes place when they are subjected to mechanical force This is thepiezoelectric effect Figure 5.10 shows the principle of how it is used to create a three-axisforce dynamometer Each force component is detected by a separate crystal oriented rela-tive to the force in its piezoelectric sensitive direction Quartz is usually chosen as thepiezoelectric material because of its good dynamic (low loss) mechanical properties Itspiezoelectric constant is only ≈ 2 × 10–12 coulombs per Newton A charge amplifier istherefore necessary to create a useful output Because the electrical impedance of quartz ishigh, the amplifier must itself have high input impedance: 105MW is not unusual.Figure 5.11 shows the piezoelectric equivalent of the dynamometers of Figure 5.8 Thestiffness is basically that of the crystals themselves Commercial machining dynamome-ters are available with natural frequencies from 2 kHz to 5 kHz, depending on size

5.2.2 Rake face stress distributions

In addition to overall force measurements, the stresses acting on cutting tools are tant, as has been indicated in earlier chapters Too large stresses cause tool failure, and fric-tion stresses strongly influence chip formation The possibility of using photoelasticstudies as well as split-tool methods to determine tool stresses has already been introduced

impor-in Chapter 2 (Section 2.4) The maimpor-in method for measurimpor-ing the chip/tool contact stresses

Fig 5.9 The loading of a ring by radial and tangential forces

is the split-tool method (Figure 2.21), although even this is limited – by tool failure – tostudying not-too-hard work materials cut by not-too-brittle tools.

Figure 5.12 shows a practical arrangement of a strain-gauged split-tool dynamometer.The part B of the tool (tool 1 in Figure 2.21) has its contact length varied by grinding awayits rake face It is necessary to measure the forces on both parts B and A, to check that the

Fig 5.10 The principle of piezoelectric dynamometry

Fig 5.11 A piezoelectric tool dynamometer

sum of the forces is no different from machining with an unsplit tool It is found that if

extrusion into the gap between the two tool elements (g, in Figure 2.21) is to be prevented,

with the surfaces of tools A and B (1 and 2 in Figure 2.21) at the same level, the gap should

be less than 5 mm wide (although other designs have used values up to 20 mm and a

down-ward step from ‘tool 1’ to ‘tool 2’) The greatest dynamometer stiffness is required This

is an instance when semiconductor strain gauges are used Piezoelectric designs also exist.Split-tool dynamometry is one of the most difficult machining experiments to attemptand should not be entered into lightly The limitation of the method – tool failure, whichprevents measurements in many practical conditions that could be used to verify finiteelement predicted contact stresses and also to measure friction stresses directly – leaves amajor gap in machining experimental methods

Fig 5.12 A split-tool dynamometer arrangement

5.3 Temperatures in machining

There are two goals of temperature measurement in machining The more ambitious isquantitatively to measure the temperature distribution throughout the cutting region.However, it is very difficult, because of the high temperature, commonly over 700˚C evenfor cutting a plain carbon steel at cutting speeds of 100 m/min, and the small volume overwhich the temperature is high The less ambitious goal is to measure the average temper-ature at the chip/tool contact Thermocouple methods can be used for both (the nextsection concentrates on these); but thermal radiation detection methods can also be used(Section 5.3.2 summarizes these) (It is possible in special cases to deduce temperaturefields from the microstructural changes they cause in tools – see Trent, 1991 – but this willnot be covered here.)

5.3.1 Thermocouple methods

Figure 5.13 shows an elementary thermocouple circuit Two materials A and B are

connected at two junctions at different temperatures T1and T2 The electro-motive force(EMF) generated in the circuit depends on A and B and the difference in the temperatures

T1and T2 A third material, C, inserted at one of the junctions in such a way that there is

no temperature difference across it, does not alter the EMF (this is the law of intermediatemetals)

In common thermocouple instrument applications, A and B are standard materials, with

a well characterized EMF dependence on temperature difference One junction, usually thecolder one, is held at a known temperature and the other is placed in a region where the

Fig 5.13 An elementary thermocouple circuit (above) with an intermediate metal variant (below)

temperature is to be deduced from measurement of the EMF generated Standard materialcombinations are copper-constantan (60%Cu–40%Ni), chromel (10%Cr–90%Ni)–alumel(2%Al–90%Ni-Si-Mn) and platinum–rhodium In metal machining applications, it ispossible to embed such a standard thermocouple combination in a tool but it is difficult tomake it small enough not to disturb the temperature distribution to be measured One alter-native is to embed a single standard material, such as a wire, in the tool, to make a junc-tion with the tool material or with the chip material at the tool/chip interface By movingthe junction from place to place, a view of the temperature distribution can be built up.Another alternative is to use the tool and work materials as A and B, with their junction atthe chip/tool interface By this means, the average contact temperature can be deduced.This application is considered first, with its difficulties stemming from the presence ofintermediate metals across which there may be some temperature drop.

Tool–work thermocouple measurements

Figure 5.14 shows a tool–work thermocouple circuit for the turning process The hot tion is the chip/tool interface To make a complete circuit, also including an EMF recorder,requires wires to be connected between the recorder and the tool and the recorder and thework In the latter case, because the work is rotating, the wire must pass through some slip-ring device If the junctions A, B and C, between the work and slip ring, the slip ring andrecorder wire and the tool and the recorder wire, are all at the same (cold junction) temper-ature, the circuit from A to C is all intermediate and has no effect on the EMF But this isoften not the case

junc-Fig 5.14 A tool–work thermocouple circuit

Dry slip rings, with their rubbing interface, frequently create an EMF The solution is

to use a liquid mercury contact If an indexable insert is used as the cutting edge, thedistance from the hot junction to the cold junction C may be only 10 mm In this case, toeliminate error due to C heating up, either the measurement time must be kept very short,

or the insert must be extended in some way – for example by making the connection at Cfrom the same material as the insert (but this is often not practical) – or the heating must

be compensated Figure 5.15 shows a cold junction compensation circuit and its ple The single wire connection at C is replaced by a standard thermocouple pair of wireswhich are terminated across a potentiometer in a region where the temperature is notaffected by the cutting The connection to the EMF recorder is then taken from the poten-tiometer slider The thermocouple wire materials are chosen so that the tool material has

princi-an intermediate EMF potential between them, relative to some third material, for ple platinum The slider is set at the point of interior division of the potentiometer, at the

exam-same ratio a/b as the tool material potential is intermediate between the two

thermocou-ple materials Copper and constantan are found suitable to span the potentials of most toolmaterials

Tool–work thermocouple calibration

The EMF measured in cutting must be converted to temperature Generally, the EMF–temperature relation for tool–work thermocouples is non-linear It can even vary betweennominally the same tool and work materials It is essential to calibrate the tool–work ther-mocouple using the same materials as in the cutting test Figure 5.16 shows one calibra-tion arrangement and Figure 5.17 shows its associated measurement circuit In thisarrangement, the tool–work thermocouple EMF is not measured directly Instead, the EMFbetween the tool and a chromel wire is measured at the same time as that of a

Fig 5.15 A circuit for compensating the cold junction C

chromel–alumel thermocouple at the same temperature Thus, the tool–chromel EMFversus temperature characteristic is calibrated against the chromel–alumel standard This

is repeated for the work–chromel combination The tool–work EMF versus temperaturerelation is the difference between the tool–chromel and work–chromel relations

Figure 5.16 shows an overview of the tool or work in contact with the chromel–alumelthermocouple (detail in Figure 5.17) The contact is made at one focus of an infrared heat-ing furnace with reflecting walls, shaped as an ellipsoid of revolution, with a 1 kW halo-gen lamp at the other focus The chromel–alumel thermocouple is fixed to the furnacebody and the tool or work is pressed on to it by a spring It is necessary to prepare the tooland work materials as rods in this method, but it is possible to heat the hot junction to1000˚C in about 10 s: the lengths of the rods, to avoid the need for cold junction compen-sation circuitry, need only be sufficient to be beyond the heat diffusion distance over thistime Example results, for a P10 carbide tool and a 0.45% plain carbon steel work, aregiven in Figure 5.18 Even at 1000˚C the EMF is only 10 mV, so a high sensitivity recorder

is needed

Inserted thermocouple measurements

Figure 5.19 shows two further possibilities of tool temperature measurement In Figure5.19(a), a small diameter hole has been bored in the tool and a fine standard thermocouple

Fig 5.16 A tool–work thermocouple calibration set-up

has been inserted in it It has the advantage that a precise measurement of temperature atthe bottom of the hole can be made, relying on the standard thermocouple, but a disad-vantage that the hole may disturb the temperature gradients in the tool If it is desired tomeasure the temperature distribution in the tool, while only boring one hole, the rake andclearance faces of the tool may be progressively ground away, to vary the position of thehole relative to the cutting edge.

A finer hole may be bored if only one wire is to be inserted in it Figure 5.19(b) shows

a single wire, for example chromel, or in this case platinum, making contact with the work

at the chip–tool interface In this way, the temperature at a specified point can be measured,

Fig 5.17 A detail of the hot junction and the associated measurement circuit

Fig 5.18 Calibration test results for P10 carbide and a 0.45% plain carbon steel

but it is necessary to calibrate the thermocouple, as was done with the tool–work couple.

thermo-5.3.2 Radiation methods

Inserted thermocouple methods require special modifications to the cutting tools Thetool–work thermocouple method only determines average contact temperatures; andcannot be used if the tool is an insulator Thermal imaging methods, measuring the radiation from a surface, have a number of attractions, if surface temperatures are ofinterest

Fig.5.19 (a) Inserted thermocouple or (b) thermocouple wire

The laws of electromagnetic energy radiation from a black body are well known The

power radiated per unit area per unit wavelength W ldepends on the absolute temperature

T and wavelength l according to Planck’s law:

Equation (5.5) can be differentiated to find at what wavelength lmaxthe peak power is

radiated (or absorbed), or integrated to find the total power W Wien’s displacement law

and the Stefan–Boltzmann law result:

lmaxT = 2897.8 mm K

(5.6)

W[W m–2] = 5.67 × 10-8T4Figure 5.20 shows the characteristic radiation in accordance with these laws.Temperatures measured in industry are usually 2000 K or less Most energy is radiated in

the infrared range (0.75 mm to 50 mm) Therefore, infrared measurement techniques are

needed Much care, however, must be taken, as real materials like cutting tools and work

materials are not black bodies The radiation from these materials is some fraction a of the black body value a varies with surface roughness, state of oxidation and other factors.

Calibration under the same conditions as cutting is necessary

One of the earliest measurements of radiation from a cutting process was by Schwerd(1933) Since then, methods have followed the development of new infrared sensors Pointmeasurements, using collimated beams illuminating a PbS cell sensor, have been used tomeasure temperatures on the primary shear plane (Reichenbach, 1958), on the tool flank

Fig 5.20 Radiation from a black body

(Chao et al., 1961) and on the chip surface (Friedman and Lenz, 1970) With the

develop-ment of infrared sensitive photographic film, temperature fields on the side face of a chipand tool have been recorded (Boothroyd, 1961) and television-type infrared sensitive video

equipment has been used by Harris et al (1980).

Infrared sensors have continued to develop, based on both heat sensing and ductor quantum absorption principles The sensitivity of the second of these is greater than

semicon-the first, and its time constant is quite small too – in semicon-the range of ms to ms Figure 5.21

shows a recent example of the use of the second type Two sensors, an InSb type sensitive

in the 1 mm to 5 mm wavelength range and a HgCdTe type, sensitive from 6 mm to 13 mm,

were used: more sensitive temperature measurements may be made by comparing thesignals from two different detectors

Most investigations of temperature in metal cutting have been carried out to stand the process better In principle, temperature measurement might be used for condi-tion monitoring, for example to warn if tool flank wear is leading to too hot cuttingconditions However, particularly for radiant energy measurements and in productionconditions, calibration issues and the difficulty of ensuring the radiant energy path fromthe cutting zone to the detector is not interrupted, make temperature measurement forsuch a purpose not reliable enough Monitoring the acoustic emissions from cutting is

under-Fig 5.21 Experimental set-up for measuring the temperature of a chip’s back surface at the cutting point, using a

diamond tool and infrared light, after Ueda et al (1998)

another way, albeit an indirect method, to study the state of the process, and this is ered next.

0.6)

Figure 5.22 shows the structure of a sensor An acoustic wave transmitted into the

sensor causes a direct stress E(DL/L) where E is the sensor’s Young’s modulus, L is it length and DL is its change in length The stress creates an electric field

equation (5.7b) lead to the possibility of detecting length changes DL as small as 7 × 10–15

m: for a sensor with L = 10 mm, that is equivalent to a minimum strain of 7 × 10–13 AE

Fig 5.22 Structure of an AE sensor

strain sensing is much more sensitive than using wire strain gauges, for which the mum detectable strain is around 10–6.

mini-The electrical signal from an AE sensor is processed in two stages It is first passedthrough a low noise pre-amplifier and a band-pass filter (≈100 kHz to 1 MHz) The result-ing signal typically has a complicated form, based on events, such as in Figure 5.23 In thesecond stage of processing, the main features of the signal are extracted, such as thenumber of events, the frequency of pulses with a voltage exceeding some threshold value

VL, the maximum voltage VT, or the signal energy.

The use of acoustic emission for condition monitoring has a number of advantages Asmall number of sensors, strategically placed, can survey the whole of a mechanicalsystem The source of an emission can be located from the different times the emissiontakes to reach different sensors Its high sensitivity has already been mentioned It is alsoeasy to record; and acoustic emission measuring instruments are lightweight and small.However, it also has some disadvantages The sensors must be attached directly to thesystem being monitored: this leads to long term reliability problems In noisy conditions itcan become impossible to isolate events Acoustic emission is easily influenced by thestate of the material being monitored, its heat treatment, pre-strain and temperature Inaddition, because it is not obvious what is the relationship between the characteristics ofacoustic emission events and the state of the system being monitored, there is even moreneed to calibrate or train the measuring system than there is with thermal radiationmeasurements

In machining, the main sources of AE signals are the primary shear zone, the chip–tooland tool–work contact areas, the breaking and collision of chips, and the chipping andfracture of the tool AE signals of large power are generally observed in the range 100 kHz

to 300 kHz Investigations of their basic properties and uses in detecting tool wear andchipping have been the subject of numerous investigations, for example Iwata andMoriwaki (1977), Kakino (1984) and Diei and Dornfeld (1987) The potential of using AE

is seen in Figure 5.24 It shows a relation between flank wear VB and the amplitude level

Fig 5.23 An example of an AE signal and signal processing

of an AE signal in turning a 0.45% plain carbon steel (Miwa, 1981) The larger the flankwear, the larger the AE signal, while the rate of change of signal with wear changes withthe cutting conditions, such as cutting speed.

References

Boothroyd, G (1961) Photographic technique for the determination of metal cutting temperatures.

British J Appl Phys.12, 238–242.

Chao, B T., Li, H L and Trigger, K J (1961) An experimental investigation of temperature

distri-bution at tool flank surface Trans ASME J Eng Ind 83, 496–503.

Diei, E N and Dornfeld, D A (1987) Acoustic emission from the face milling process – the effects

of process variables Trans ASME J Eng Ind 109, 92–99.

Friedman, M Y and Lenz, E (1970) Determination of temperature field on upper chip face Annals

CIRP 19(1), 395–398.

Fig 5.24 Relation between flank wear VB and amplitude of AE signal, after Miwa et al (1981)