# Contacts to Force the High Current, Columnar Arc into the Diffuse Mode

In Chapter 2, Section 2.7.2, in this volume, I discussed the effect of an axial magnetic field (AMF) on the high current columnar arc, which forms between opening contacts. A typical sequence is illustrated in Figure 2.82. After the rupture of the molten bridge, the bridge column forms. As I have discussed in Section 2.7.2 in this volume, in the presence of a high enough AMF, this column expands into a diffuse column and then into a fully diffuse arc. The cathode spots spreading out over the cathode’s surface characterizes this arc. Electrons from these cathode spots are confined by the magnetic flux lines in the inter-contacts region and, because of the associated creation of radial electric fields, the ions are also confined to the intercontact region. An example of a 50kA diffuse arc is shown in Figure 3.39. Under ideal conditions this high current, diffuse vacuum arc distributes the arc energy over the whole cathode and anode contact surfaces and thus greatly reduces the opportunity for gross erosion of the contacts. As I have also discussed in Section 2.7.2 in this volume, any practical contact design that develops an AMF for the purpose of forcing a high current diffuse vacuum arc to form, has to ensure that for a given arc current, the resulting axial magnetic field is higher than a critical value.

The earliest experiments using the AMF for current interruption used a coil wrapped around the outside of the vacuum interrupter that carries the circuit current and produces the AMF [123], see Figure 3.40. This arrangement has the advantages of:

• 1) The contacts can be a simple butt structure
• 2) Increasing the number of coil turns can increase the strength of the AMF [124]

• 1) The coil must be insulated because it has to be connected to one end of the vacuum interrupter (usually the fixed end), and will have its potential
• 2) At high currents the coil is under a high “loop” force, thus either the coil itself will need to be very strong or else it will require extra external support
• 3) The coil and its connection to the fixed terminal will add extra impedance to the whole vacuum interrupter structure (this disadvantage can be alleviated to some extent by clever design [125])

FIGURE 3.39 Examples of high current diffuse vacuum arcs between Cu-Cr(25wt.%) contacts in an axial magnetic field: (a) a current of 50kA for a contact diameter of 50mm and (b) a current of 101 kA for a contact diameter of 86.5mm.

FIGURE 3.40 A vacuum interrupter arrangement with an external magnetic field coil to provide the axial magnetic field.

• 4) The resulting coil and interrupter system tends to be bulky, which may affect the pole spacing in a three-phase switch or circuit breaker
• 5) When the fault current occurs, it takes time for the AMF to penetrate the contact gap

In spite of these disadvantages, there have been a few practical applications of an external coil. All of these have been with vacuum interrupter designs where the shield is attached to the fixed terminal of the interrupter (see Section 3.4.2) and for circuit voltages < 12kV [126, 127]. Figure 3.41 shows a photograph of a three-phase 7.2kV contactor with internal field coils around the fixed end of the vacuum interrupter.

Almost all vacuum interrupter manufacturers, who utilize the effect of the AMF to form the high current, diffuse arc, incorporate the AMF into the design of the structure supporting the contact itself. Figure 3.42(a) shows one example of a ‘/2-coil, two-arm, AMF contact structure [128]. In Figures 3.42(a) and (b) the current flows up the copper terminal and splits into a half coil before passing into the contact [129]. The passage of the current through the half coils on the back of each contact results in an AMF in the contact gap. The diameter of the contact’s surface through which the magnetic flux passes is the internal diameter of the coil. It is important to note that in these contact structures the area available for the cathode spots to spread can be significantly less than the contact’s total area. There have been many designs described in the technical and patent literature, e.g., [128-141]. The coils behind the contacts do increase the impedance of the contact structure. This impedance can be reduced by using the three- or four-segment coil designs showm in Figure 3.42(c) and (d). Wang et al. [130] have analyzed the effect of the number of coil segments, the contact diameter, the coil height, the coil width and the contact gap on the AMF strength and on the phase shift of the maximum AMF wnth respect to the maximum current. A summary of their findings is given in Table 3.11. Figure 3.43 shows an example by Yanabu et al. [131, 132]. Here, a

FIGURE 3.41 A three-phase vacuum contactor with external axils magnetic field coils around the vacuum interrupters (courtesy of Eaton Corporation).

FIGURE 3.42 (a) An example of a contact structure for a vacuum interrupter with the magnetic field coil behind the contact face [128], (b) a two-segment coil design or '/2-coil, two-arm design, (c) a three-segment coil design or '/з-coil, three-arm design (d) a four-segment coil design or ‘A-coil,four-arm design, and (e) a single coil design.

TABLE 3.11

The Effect of AMF Coil Parameters and Contact Gap on the AMF and the Phase Shift of the AMF with Respect to the ac Current [130]

 Design Parameter AMF strength Number of coil segments (2-4) Contact diameter (30mm-100mm) Coil height (4mm-18mm) Coil width (5mm-15mm) Contact gap (6mm-12mm)

T Increasing, l Decreasing, 0 Little effect

FIGURE 3.43 Four segment axial magnetic field contact structure proposed by Murano et al. [131] and Yanabu et al. 132].

four-segment coil is placed behind the contact. The slots in the contact faces are to reduce eddy currents, which result in reducing the effective AMF in the contact gap.

Another design by Kurasawa et al. [133] gives an interesting magnetic field structure that changes direction across the contact face (Figure 3.44). This type of design was first developed at the end of the 1970s and has been resurrected as the “new” development in AMF contact structures in the late 1990s [134]. Another structure that has also had practical success is shown in Figure 3.45. Here the “coil” is a cup with slots cut into it in the way shown [135-138]. The six-slot design shown in Figures 3.45(a) and (b) find the most common usage. Wang [139] has shown that a four-slot design gives an enhanced AMF that is useful at the longer contact gaps needed for single vacuum interrupters operating in 126kV circuits. While this design gives a lower AMF than that experienced in the coil designs shown in Figure 3.42, it also works well if slots are placed in the contact surface to disrupt the induced eddy currents. It also has a lower impedance than the coil designs. Figure 3.46 shows data from Xiu et al. [140] who have compared the AMF for the two-arm coil design, see Figure 3.45(b) and the six-slot cup design, Figure 3.45(a).

The major advantage of this class of contact structure is its ability to keep the arc diffuse during very high currents because the magnetic field increases as the current increases. It has been used to interrupt very high currents. Yanabu et al. [132] report interrupting currents as high as

FIGURE 3.44 A bipolar axial magnetic field contact structure proposed by Kurosawa et al. [133].

200kA in 12kV circuits and Voshall et al. [68] report currents up to 63kA in 72kV circuits. The major disadvantage is that the contact design increases the impedance between closed contacts. Thus, the vacuum interrupter must dissipate a higher energy when passing high steady state currents. Mayo [141] has shown that the increase in impedance can be mitigated to some extent by using a coil with a trapezoid cross section. Liu et al. [142] have proposed another interesting design, which is illustrated in Figure 3.47. This uses a central Cu post that is forced to make contact to the underside of the main contact when the vacuum interrupter closes and deforms the cup. This allows the bulk of the circuit current to bypass the AMF coil. When the contacts are parted the cup restores itself to its original dimensions (this may be aided by some contact welding). It is also aided by the Lorenz force between the adjacent segments in the cup. The central pillar breaks contact with the underside of the main contact and the current flows through the cup and the AMF is generated.

One other proven AMF design that separates the generation of the AMF from the current carrying path is shown in Figure 3.48 [143]. Here a series of steel “horseshoe” plates is placed behind a butt contact. The magnetic field that surrounds the conductor from the current flowing through the conductor is trapped in the steel horseshoes. By judiciously designing the gap in the horseshoes, it is possible for this magnetic flux to preferentially jump the gap between the contacts to a horseshoe behind the opposite contact rather than jump the gap to its own horseshoe. When this happens, a

FIGURE 3.45 The cup, magnetic field structure [135,136].

FIGURE 3.46 A comparison of the axial magnetic flux for a two-segment AMF coil design with a six-cut cup design, with a 100mm contact diameter and a current of 40 к A. [140].

bipolar, AMF results. The major disadvantage of this structure in the past has been that for contact gaps greater than about 12mm, the design of the horseshoe structure becomes more difficult. That being said, vacuum interrupters are being developed with long contact gaps [144] and for use in circuits with a system voltage of 126kV [145]. This structure has the major advantage that its total impedance can be lower than the coil or cup designs.

FIGURE 3.47 One proposed AMF cup contact design to give a low impedance contact structure when carrying rated continuous current and to give the AMF when interrupting fault current [142].

FIGURE 3.48 The “horse-shoe” contact structure that provides a bipolar axial magnetic flux [143].

FIGURE 3.49 A simple axial magnetic field example used to estimate the magnetic flux.

An easy way to estimate the value of the axial magnetic field is shown in Figure 3.49. Here two rings of radius “r” a distance “d” apart give the magnetic field BA at mid-gap:

Where //„ is the permeability in free space (4к x 10-7 T.m.A-1), “i” is the current in amperes and “r” and “d” are in meters. This simple structure can also be used to estimate the added contact force for closed contacts with a high current passing through the AMF structure behind the contact faces: see Section 4.4.2 in this volume. A plot of:

Is shown in Figure 3.50. Chen [146] gives for a 100mm diameter the ratio of:

From Figure 3.50 the ratio is also 0.78. Figure 3.50 shows that for a given contact gap, BA decrease as “r” increases and for a given “r,” and BA also decreases as the contact gap increases which agrees with the findings shown on Table 3.11. I have discussed in Section 2.7.2 in this volume that as the current to be interrupted increases, the AMF contact area and/or the BA must also increase to prevent the plasma from the cathode spots from constricting and overheating the anode. Figure 2.89 also shows that as the contact gap increases the plasma plumes from the cathode spots also tend to overlap. Eventually they will also form a region where the plasma is constricted which will overheat the anode. Therefore, as vacuum interrupters are developed to interrupt higher currents and are being developed for use at higher voltages, where larger gaps are required to withstand the higher BIL voltages, the AMF contact has to be designed to reduce the effects of the plasma plume overlap.

FIGURE 3.50 The calculated effect of the contact gap as a function of the coil radius.

Zhang et al. [147], Ryu et al. [148], Yao et al. [149] and Zhang et al. [150] have proposed the two- layer, three-arm coil design shown in Figure 3.51. The effect of the contact gap on the magnetic flux for this contact design is shown in Figure 3.52 [149]. Zhang et al. [151] propose a 3/4-coil design (see Figure 6.3), which also provides adequate field strengths at longer contact gaps.

As I have discussed in Section 2.4 in this volume, to be most effective the AMF must be greater than a minimum value. In an ac circuit the sinusoidal current flowing in the coil structure behind the

FIGURE 3.51 The two-layer, three-arm coil proposed for high voltage and high current vacuum interrupters [147-150].

FIGURE 3.52 The magnetic flux as a function of contact gap for 100mm diameter, two-layer, three-arm coil design [149].

contact face results in an AMF that also has a sinusoidal structure. This varying AMF will in turn develop eddy currents in the contact structure’s face and the base (for cup-shaped designs). These eddy currents produce a counter AMF that reduces the effect of the AMF from the coil structure behind the contact and hence the AMF in the contact gap itself. A reduced AMF can hinder the development of the fully diffuse, high-current, vacuum arc. In order to minimize the effects of eddy currents slots can be cut into the face of the contacts. Figure 3.53 gives some examples. The eddy currents also change

FIGURE 3.53 Examples of slots cut into the face of the vacuum interrupter’s contact to reduce the effects of eddy currents in the AMF contact structure.

the phase of the AMF with respect to the ac current by as much as 20° to 30°, see for example Figure 3.52. The maximum flux from the AMF does not occur at the maximum current through the contacts. This is also an important design consideration during the opening of the contacts and the development of the high current diffuse vacuum arc. This phase difference between the AMF and the ac current results in a significant AMF across the open contacts when the ac current reaches zero.

The development of the AMF type of contact structure has been greatly assisted by the advent of engineer friendly, three-dimension, finite element software and high-speed, high capability personal computers. Stoving et al. [152] give an excellent example of FEA analysis of a number of AMF contact designs. This type of analysis has proven extremely useful in analyzing the effect of current flow through the coil and contacts on the distribution of AMF across the face of the contact [153]. It has also shown the advantage of placing slots in the contact face to alleviate the effects of the eddy currents induced by the AMF; see, for example, Figure 3.54 [154]. FEA analysis, other magnetic structure analysis and computer modelling are now the design tools used by all engineers involved with advanced contact design for vacuum interrupters: see, for example [147-150]

The vacuum interrupter designer has limitations placed upon the contact design. First of all, there is a cost constraint. As larger diameter ceramics are more expensive than ones of smaller diameter, the contact diameter is usually restricted. That is, the contacts diameter must be optimized for a given short circuit current interruption performance. Thus, the need to limit the vacuum interrupter’s diameter will be satisfied. Another complication in the design is the spacing between the contact and the shield. In order to maintain the vacuum interrupter’s high voltage performance this spacing also has to be optimized. Thus, it is possible, even if the AMF is high and the cathode spots spread out across the cathode’s surface, the plumes above the cathode spots will begin to overlap if the current is high enough. Schulman et al. [155] describe three vacuum arc modes in an AMF after the bridge column stage of the opening contacts. These vacuum arc modes that depend upon the current and the contact’s diameter are:

1) The multi cathode spot vacuum arc: Here the cathode spots spread over the cathode surface with their associated plasma plumes. The space between the cathode spots is great enough that their associated plasma plumes do not overlap. This results in a low level

FIGURE 3.54 The effect of slots in the contact face and in the cup structure of the coil shown in Figure 3.42(a) on the magnetic flux in the center of the contact gap as a function of the radial distance from the center of the contact [154].

of light intensity from the inter-contact gap. As the cathode spots spread over the whole cathode surface, there is a more or less uniform erosion rate of the cathode, which can be measured as a (constant) gC_l. The anode remains a passive receiver of electrons.

• 2) The high current diffuse vacuum arc: Here the cathode spots still spread over the cathode’s surface, but now there is overlapping of the plasma plumes and visible plasma now fills the intercontact gap with a more or less uniform light intensity. The individual cathode spots can still be observed, but the individual plasma plumes can no longer be resolved.
• 3) The high current diffuse column vacuum arc: Here there is a central plasma core in the intercontact gap that joins cathode to anode. This plasma core is much brighter than the cathode spots that can no longer be resolved. It is also much brighter than the peripheral plasma that surrounds it. This diffuse column arc tends to constrict as it approaches the anode. The plasma constriction at the anode can then develop an anode spot. This anode spot reaches a high temperature which, in turn, releases metal vapor into the inter-contact gap: see Sections 2.7.2 and 2.8 in this volume.

The current level, the contact diameter, the contact gap and the strength and structure of the AMF generally determine these vacuum arc modes. Figure 3.55 shows a schematic of the local light intensity from the vacuum arc as the current increases. Chen et al. [156] present photographs of vacuum arcs in an AMF that illustrate these effects nicely. Schulman et al. [155] have used observations of such changes in the vacuum arc’s luminosity to develop the “Appearance Diagram” shown in Figure 3.56. Here the distinctive plasma core develops at 35kA. They also have developed an analysis based upon Gundlach’s research (see Figure 2.83 and Equations (2.44) and (2.45) that, if the arc voltage, Uarc, is greater than a value given by:

where Uarc as a function of time, t, is a function of ian(t) and the contact gap <7(7), a[d(t)] is a constant that is a function of d(t) and b[d(t)] is a constant that includes a term for the increase in Uarc with arc length d(t) ( h « Uan.[BA-> 0, iarc —> 0])), then the vacuum arc would be in the diffuse column mode. Matsui et al. [157] have explored the range of current and AMF where the diffuse column results in

FIGURE 3.55 Schematics (not to scale) of the intercontact plasma, showing the increasing luminosity and the constriction to a bright diffuse plasma column as the arc current increases; contact gap about 8mm [155].

FIGURE 3.56 The appearance diagram for a high current, vacuum arc in an axial magnetic field [155].

FIGURE 3.57 The threshold anode melting current as a function of axial magnetic field for 22mm diameter Cu-Cr contacts with a contact gap of 6mm [157].

anode melting. Figure 3.57 shows the range of currents above which anode melting occurs for 22mm diameter Cu-Cr contacts with a 6mm contact gap. They establish an equation:

where ith к A is the threshold current and BA mT is the magnetic field. Keidar et al. [158] have shown that once Uan. of the diffuse column vacuum arc exceeds a given value, it is difficult for a cathode spot to exist outside it. It is thus possible for the spread of the cathode spots from a high current vacuum arc to be restricted to an area less than that of the total contact’s area even in the presence of an AMF. At high currents, for a given contact diameter, the energy input into the cathode and the anode will be mostly confined to the limited area. This can result in melting of the contact surfaces and a potential limitation on the level of short circuit current that the vacuum interrupter containing this contact can handle. Watanabe et al. [159] show for a Cu-Cr (50 wt.%) contact the anode reaches a temperature of about 1750K at its current interruption limit. At this temperature the vapor density of evaporated contact material reaches about 3 x 1020 atoms/m3 or a pressure of about 8 Pa. They conclude that the density of metal vapor in the contact gap dominates the reignition of the vacuum arc after current zero. Niwa et al. [160] in a further investigation show that for Cu-Cr contacts with an AMF the current interruption limit is reached when the temperature of the anode contact reaches 2000C to 2500C. They also show' that as the Cu content of the Cu-Cr contacts increases, so the contact’s current interruption ability also increases. They attributed this to the increased thermal conductivity of the anode contact as its Cu content becomes greater. I will discuss this further in Section 4.2.3 in this volume. Fortunately, the melting of the contact surfaces does not necessarily result in an immediate failure to interrupt a circuit. Schellekens et al. [161] have studied contact surfaces after the interruption of the high current diffuse column arc. They have shown that the flow of molten metal over the contacts’ surfaces after arcing can help to distribute the heat energy more efficiently and help to cool the contact surfaces more quickly, thus helping the dielectric recovery of the contact gap. In spite of this, experienced vacuum interrupter designers know' that a given contact design wall always have a limit in the level of current it can interrupt.

Since the performance of an AMF contact structure depends upon the spreading of the cathode spots over that portion of the contact’s surface area that is equal to the internal diameter of the coil structure, one would expect that the maximum current that can be interrupted imax (rms) is given by:

where AB(Z) is the area with diameter DB(Z) of the contact surface through which the magnetic flux passes. Figure 3.58 presents a set of data for an AMF contact structure similar to that shown in Figure 3.42(a) (i.e., a ‘/2-coil, 2-arm design wdth slots cut in the contact faces). The dependence of imax (rms) is certainly not proportional to DB(Z), Figure 3.58(a). It does appear, however, that it does

FIGURE 3.58 An example of the interruption ability of an AMF contact structure with Cu-Cr contacts as a function of DB(Z) for a 38kV system voltage (a) Interruption current as f(contact diameter) (b) interruption data from (a) as f(contact diameter)2 (c) interruption data from (a) as f(contact diameter)1-4.

FIGURE 3.59 The contact surface area affected by the magnetic flux for a '/2-coil design without and with slots in the contact face [162].

satisfy Equation (3.31); i.e., there is dependence on Db(z), Figure 3.58(b). Del Rio et al. [162] show in Figure 3.59 that the BA does not cover the whole contact face. Henon et al. [163] suggest that the actual effective area on the contact’s surface is where B(z) > 4mT.kA-‘. They also show in their study that:

This dependence for the same set of data is shown in Figure 3.58(c). The scatter in the data is such that both Equations (3.31) and (3.32) seem to be satisfied. Again, the interruption data in Figure 3.58 should not be considered the absolute interruption limit for AMF contacts. The ultimate limit will depend upon many variables including; the exact composition of the contact material, the design of the AMF structure and the internal geometry of the vacuum interrupter. These data show that the contact diameter is an important design parameter. Its importance results from the greater area available for the expansion of the cathode spots from the high current diffuse vacuum arc. The limit on interruption comes when this area confines too high a density of cathode spots. When designing the AMF contact structure, the designer also needs to consider the required contact gap. The strength of the AMF for a given contact structure is a function of the contact gap. Also, the expanding plasma plumes above the cathode spots have a greater opportunity to overlap as the contact gap is increased. Now the AMF, Bz is a function of the contact gap d. Liu et al. [164] propose that the maximum rms current, imax, that can be interrupted by a given AMF contact design can be given by:

They show that over the range of 45mm < DB(Z) < 70mm and 3mm < d. < 15mm that:

where Bz is the maximum AMF given in mT.kA"1. Again, the vacuum interrupter designer should treat these data with care. Equations (3.33) and (3.34) seem to imply that for a given contact diameter the imax can be increased simply by the ever increasing Bz. It is certainly true that before the high current vacuum arc can be fully diffuse, the B. has to be greater than a critical value Bcr as I have already discussed in Section 2.7.2 in this volume. However, the density of cathode spots for a given contact diameter certainly limits the overall current interruption performance of the contact structure. Most AMF designs seem to show maximum values of B. in the range lOOmT to 400mT.

In fact, Chaly et al. [165] show that increasing the AMF only appears to be effective for B. values of less than 400mT.

Because of the formation of the diffuse column vacuum arc, much effort has been applied to develop contact geometries, which distribute or fragment the AMF. All of the practical contact designs produce an AMF, which is nonuniform across the contact diameter and in the contact gap: see, for example Wootten et al. [166]. Examples of the variation in magnetic field for different AMF contact structures are shown in Figure 3.60. These variations of AMF across the face of the contact in practical designs have an effect on the distribution of the cathode spots. Some researchers have attempted to modify the contact structure and the resulting AMF to move the peak of the AMF from the center of the contact to its periphery. One advantage of this is that the area A* of high B(z) is now moved to an annulus with an inner radius of r, and an area given by:

where r is the contact radius and w is the annulus width. Indeed, it has been demonstrated by Chaly et al. [167], using a “magnetic barrier” to shape the AMF distribution to that shown in Figure 3.61 does effectively distribute the cathode spots more uniformly across the cathode’s surface. They then show that the anode erosion resulting from an 1 lkA arc using a similar AMF with a 30mm diameter contact is much less pronounced than that obtained by using the profile shown in Figure 3.60(a). This work, however, uses only limited values of current, AMF and contact diameter. Homma et al. [168] propose such a design, which they call a SADE contact (self-arc diffusion by electrode). Their patent [169] presents the details of the desired AMF: it must have its lowest value in the center of the contact and have its highest value toward the contact’s edge. They use a value of BA in the center of the contact that is 75-90% of the BUlmjn) value given by Gundlach for a given current (see Figure 2.83). The BA then gradually increases to the BU(mjn) value at a distance from the contact’s center of between 20% and 40% of the contact’s radius. It then peaks to values between 1.4 to 2.4 times BU(min)

FIGURE 3.60 Examples of the variations of magnetic flux across the diameter of the contact gap, (a) the unipolar structure, Figure 3.42(a), (b) the horseshoe structure, Figure 3.48, (c) the bipolar structure, Figure 3.44 and (d) the structure shown in Figure 3.45(a).

FIGURE 3.61 Extreme distribution of the magnetic flux to the edges of the contact [167].

FIGURE 3.62 The distribution of the magnetic flux for the SADE, vacuum interrupter, contact structure [169].

at 70 to 100% of the contact’s radius. Figure 3.62 taken from [169] gives a representative example of the profile of the BA as a function of the contact’s radius. One way of achieving such an AMF profile is shown in Figure 3.63. Here horseshoe steel plates are placed around the support posts between the field coil and the contact itself. Current passing through the support posts induces a magnetic flux through the steel. If the contact gap is chosen judiciously, the magnetic flux can be forced to cross the contact gap, giving rise to an enhanced AMF at the contact’s edge. This is similar to the principle of the horseshoe contact described earlier, Figure 3.48. The major problem with this design is that in spite of the publications citing its virtues, there is very little real data on its practical design or of its actual performance in a power circuit. Thus, it is difficult to evaluate whether or not this contact does indeed perform in a manner superior to other AMF contacts. Experiments by Shkol’nik et al. [170] do seem to show that there is some advantage in using such AMF structures. Here the interruption data, however, should be judged warily, because they open their contacts with a low current and allow the cathode spots to spread over their contacts before applying the full current pulse. In doing so, their contacts do not experience the full effect of the formation of the high current diffuse

FIGURE 3.63 An example of a possible SADE contact structure [169].

vacuum arc while the contacts are opening. There have been other proposals to produce a more complex AMF pattern. One of these is an extension of the original horseshoe contact [171]. Here the horseshoe pieces themselves are cut in half to produce an AMF that crosses the contact faces four times; making it a quadrapole design. Fink et al. [172] have developed a similar design. They claim that their horseshoe quadrapole deign can interrupt 63kA. It is also possible to cut slots into the contact faces to direct the current flow in the contact itself, which can also be used to enhance AMF peaks; see Figure 3.53(c) and (f). Other designs have used magnetic materials inside the cup behind the contact face in order to shape the AMF (e.g. [173]). Examples are by Shi et al. [174, 175] are given in Figure 3.64. Liu et al. [176] show a variation of Figure 3.64(a) with rectangular plates

FIGURE 3.64 The use of magnetic materials inside a cup-shaped magnetic field structure to shape the AMF in the gap between the contacts [174, 175].

in place of the cylinders. They claim that the addition of the plates increases the Вл by as much as 125% over the design without plates for the contact diameter and contact gap they used. Zheng et al. [177] show the effect of placing an iron ring inside the cup AMF design: See Figure 3.65. At a 2mm contact gap the effect of the iron ring is much greater than at 10mm. At 2mm, however, the vacuum arc is developing into the fully diffuse form from the transition vacuum arc, see Section 2.7.2 in this volume. Therefore, the AMF structure formed with the iron ring would greatly assist the cathode spots to distribute to the outer periphery of the cathodes surface. Even though the outer edge of the AMF decreases as the contacts open, I would expect that the AMF would maintain the cathode spots in their widely dispersed pattern. Niwa et al. [178] also claim that a split iron ring placed inside the cup AMF design effectively produces a more uniform AMF across the contact gap. Liu at al. [179] have analyzed the iron plate structure shown in Figure 3.64(b). They show that the AMF is not greatly affected by the number of slots in the iron plate, its inner radius, its outer radius, its height and its thickness. Kulkani et al. [180] orientated the two contact structures with the ‘/2-coil, two-arm design by 90° with respect to each other: see Figure 3.66. They show that the 90° orientation gives a more uniform BA across the face of the contacts. This results in less severe erosion of the contacts after interrupting high currents.

It has been claimed that the bipolar designs and quadrapolar designs discussed above also fragment the magnetic field and split the plasma up across the contacts. Indeed, photographs of vacuum arcs between these contact structures [133, 134, 172] have shown multiple concentrated arc regions operating in parallel. Within each of these regions the vacuum arc mode is that of the diffuse

FIGURE 3.65 Comparison of the AMF for a six-cut, 48mm diameter, cup AMF contact with and without an iron ring inside the cup for a 2mm and a 10mm contact gap and a current of 20kA (rms) [177].

vacuum arc. If, however, the arc is restricted to one segment (i.e. at the position of the initial bridge column arc), then an intense diffuse column will develop in that quadrant with a resulting overheating, perhaps gross melting of the contacts and the potential for the vacuum interrupter to fail to interrupt the circuit current. The effect of limited arc structure is well illustrated in Figure 4 of reference [134]. Here the erosion pattern for a bipolar magnetic field contact clearly shows that the vacuum arc is confined to only one portion of the contact. There is also some indication that the lower AMF, for a given current, produced by the quadrapole structure results in more constriction of the plasma towards the anode and a lowering of its current interruption ability [181]. Both the coil and cup AMF designs are usually manufactured from OFHC copper. After manufacture these structures will have been annealed and can easily be deformed. It is therefore necessary to braze a support structure inside these designs. This structure will prevent the AMF structures from collapsing or from expanding if the contact faces weld. Li et al. [182] show that a stainless-steel support does not affect the distribution of the original magnetic flux. A novel interrupter design first described by Reese [183] and developed by Alferov [184] is the multiple rod structure shown in Figure 3.67(a). A high-current diffuse arc forms between the metal rods during interruption of high currents. A triggered version of this design illustrated in Figure 3.67(b) by Wang et al. [185] showed the potential to interrupt ЮОкА. Lamara et al. [186] propose other designs using a combination of AMF and TMF structures, which separate the normal current conduction and the high-current interruption functions.

It is obvious from the technical literature that each of the vacuum interrupter manufacturers is actively developing their new generation of AMF contact designs. Certainly, the AMF modification ideas I have discussed above present intriguing possibilities that could lead to improved vacuum interrupter performance. The continued search for the optimum Cu-Cr composition or even multiple compositions across the contact’s face [187] may also enhance the vacuum interrupter’s performance. As vacuum interrupters are designed for sub-transmission and transmission voltages; i.e., 72kV to 242kV, the contact gaps will necessarily become much greater. This will place a greater

FIGURE 3.67 The rod vacuum interrupter structure [183,185].

burden on the vacuum interrupter designer to maintain the AMF at a value greater than Bcriir Xui et al. [188] Zhang et al. [151] and Yao et al. [149] present examples of a double coil design to achieve this at large contact gaps. The challenge will be to limit the contact structure’s impedance to an acceptable level. As I will discuss in Chapter 6 in this volume, the vacuum interrupter placed in a vacuum circuit breaker does not open as soon as a fault or short circuit current occurs. There is a delay that results from the mechanism itself and from the need for coordination with the other protection devices in the electrical system. This delay allows time for the AMF to be somewhat out of phase with the fault current. Thus, in a practical situation the maximum AMF will never occur at the maximum current flowing through the vacuum arc and there will be a nontrivial AMF present at current zero. At present, there is no consensus on the best contact design or even the best contact material structure. Each proponent for a particular position has presented limited experimental data to support their claim for superiority. Meanw'hile the AMF contact designs presently in production continue to be successfully and economically applied in the widest range of circuit switching applications.