Cathode surface effects and H.F.-behaviour of vacuum arcs

Cathode surface effects and H.F.-behaviour of vacuum arcs

Citation for published version (APA):

Fu, Y. H. (1990). Cathode surface effects and H.F.-behaviour of vacuum arcs. Technische Universiteit

Eindhoven. https://doi.org/10.6100/IR340471

DOI:

10.6100/IR340471

Document status and date:

Published: 01/01/1990

Document Version:

Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be

important differences between the submitted version and the official published version of record. People

interested in the research are advised to contact the author for the final version of the publication, or visit the

DOI to the publisher's website.

• The final author version and the galley proof are versions of the publication after peer review.

• The final published version features the final layout of the paper including the volume, issue and page

numbers.

Link to publication

General rights

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain

• You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:

www.tue.nl/taverne

Take down policy

If you believe that this document breaches copyright please contact us at:

openaccess@tue.nl

providing details and we will investigate your claim.

CATHODE SURFACE EFFECTS

AND

H.F.-BEHAVIOUR OF VACUUM ARCS

CATHODE SURFACE EFFECTS

AND

H.F.-BEHAVIOUR OF VACUUM ARCS

PROEFSCHRIFf

ter verkrijging van de graad van doctor aan

de Technische Universiteit Eindhoven, op gezag

van de Rector Magnificos, prof. ir. M. This, voor

een commissie aangewezen door het College

van Dekanen in het openbaar te verdedigen

op vrijdag 16 november 1990 te 14.00 uur

door

Fu, Yan Hong

geboren te Jiangxi, China

Dit proefschrift is goedgekeurd door de promotoren:

Prof.dr.ir. W.M.C. van den Heuvel

en

Prof.dr. W

.R. Rutgers

Co-promotor:

Dr.ir. R.P.P. Smeets

CIP-GEGEVENS KONINKLIIKE BffiLIOTHEEK, DEN HAAG

Fu, Yan Hong

Cathode surface effects and H.F.-behaviour of vacuum arcs/

Fu Yan Hong. - [S.l. : s.n.] . - Fig., tab.

Proefschrift Eindhoven. - Met lit.opg., reg.

ISBN 90-9003766-7

SISO 661.52 UDC 621.316.57.064.26(043.3) NUGI 832

Trefw.: hoogspanningsschakelaars

I

gasontladingen.

-Contents:

1.

IlfrR.ODUCI'Im ••...•••••...•••••....••...••.•..•...

1

1.1. General ...•... : ...•...•...•... ! 1.2. ~of relevant pbenomena ...•••... 2

1.3. Purpose of the present investigation ... 3

2.

DC YACUUM ARC

STABILITY AND

ARC VOLTAGE ...•...

5

2.1. Introduction ... 5

2.2. Experimental circuit ... 5

2.3. Contact surface condition ...•... 6

2.4. Arc lifetime measurement ... 13

2.5. Arc voltage measurement ... 15

2.6. Cathode erosion pattern ...•... 19

2. 7. Conclusions ... 21

3. JI(Jfi<If OF ~THODE SPOI' Wim TRANSVERSAL MAGNE'fiC FIELD ... 22

3.1. Introduction ... 22

3.2. Experimental set-up ...•... 23

3.3. Average velocity of cathode spot ...•.... 24

3.4. Influences of the cathode surface microstructure on arc voltage and on erosion patterns ..•••...•... 27

3.5. Correlations among arc voltage, cathode spot velocity and cathode erosion structure ...•... 29

3.6. Two dimensional measurement on cathode spot motion ...•... 39

3. 7. Conclusions ... 41

4. ~ISM OF ~THODE-SPOI' MOTION ... 43

4.1. Introduction ... 43

4.2. Growth of a cathode crater ... -... 44

4.3. Formation of a new emission site ... 50

4.4. Continuity of cathode spot renewal ... 53

5. EXPERIMENTS ON INTERRUPTION OF REIGNITION aJRRENT

AND YOLTAGE ESCALATION ... 64

5.1. Introduction .•...•••... 64

5.2.

An

investigation of high frequency current interruption by current injection method ... 65

5.3. The "cold" breakdown voltage of an increasing vacuum gap ... 75

5.4.

An

investigation of multiple reignitions and volatge escalation in a 7 kV AC circuit ...•••. 78

5.5. Theoretical interpretation on the dielectrical breakdown voltage of a small vacuum gap after arcing ...•...••• '86

5.6. Conclusions ...•• 88

6. SIMULATION AND CALaJLATION ON aJRRENT INTERRUPTION AND REIGNITION ... 89

6.1. Introduction ... 89

6.2. Analytical study on an inductively loaded single phase circuit ...••••... 89

6.3. Computer simulation of the voltage escalation process in a single phase circuit .•••... 100

6.4. Conclusions ...•••... 104 7. APPENDIX ••.•••...•••.••...••••••••.••••...•••••••.. 105 8. SUMMARY ••••••••••..•...•..•.•••••••••••••••••••••.•...•.•.•••••••• 110 SAJIENVATTING ...•.••.•••... 112 9. REFERENCES ...•... 114 10. ACKN~ ••••...•..•••••••••...••...•.•••••••••••... 119 11. a.JRRICULUM VITAE ...•.••...•••••••••...•...•... 120 - iii

-1.

INTRODUCTION

1.1. General

A circuit breaker is a switching device which has a conducting function in "closed" position and an insulating fnnction in "open" position. Its main duty is to interrupt electrical power transmission and distribution networks under fault condition.

The first vacuum circuit breakers (as one type of circuit breakers) were already built in the 1920' s. However, they only became c011111ercially available for power systems in the early 1960's. Since then, the vacuum circuit breaker has been developed rapidly and applied extensively in the medium voltage range (3 to 36 kV) of power systems. This is due to its distinctive advantages: simple structure, convenience of maintenance, absence of contamination to the environment and long service life

(Lafferty, 1980). There are, however, .still requirements from the users to be satisfied, such as raising the interrupting capability, lowering the overvoltage generated by switching, reducing the cost etc.. Therefore, development and research work on this subject still meets much interests. A vacuum circuit breaker includes a vacuum switch as the key part and an operating mechanism as the auxiliary part. A vacuum switch mainly consists of a fixed and a moveable contact. When it is used to break a certain current, the moveable contact is moved apart 'from the fixed one. A so called "vacuum arc" will be formed between the two contacts at the moment of separation. This vacuum arc is sustained by the metal vapour evaporated from the cathode which is locally heated by the "cathode spots". In an AC supply system, a fault can occur at any time. The separation of the interrupter contacts is, therefore, random with respect to any given point within the AC half-cycle. The extinction of the arc, however, mostly occurs near a natural zero of the sinusoidal current. Immediately after the arc extinction, a recovery voltage builds up across the vacuum gap. If this voltage can be withstood by the gap, the interruption of current is successful. A qualified vacuum interrupter should not only be able to break a short circuit current, but also have a high dielectric strength. The processes occurring around AC arc current zero are the most important events during the current interruption operation.

1.2. Swm!ary of relevant phenomena

The behaviour "Of a vacuum arc is strongly determined by the cathode IIIB.terial which is the source of the arcing medium. The vacuum arc has two different modes: diffuse arc and constricted arc respectively. The diffuse

I

mode arc occurs at currents lower than 7 to 8 kA (for a

co~r

cathode)

when only the cathode is active.

The~

arc appears at currents higher than 10 kA (for copper) when both cathode and anode are involved in vapour production (Reece, 1963). This study will be restricted to the diffuse mode arc, in which the cathode spot phenomena are dominating. It is known that a single cathode spot can only carry a limited arc current

;·

which is highly dependent on the cathode IIIB.terial. The average number or cathode spots is proportional to the arc current (Djakov and Holmes, 1971). The cathode spot is a small (diameter of a few tens of JJIII) luminous region which is the emission source of electrons, positive ions, neutral vapour and microparticles into t.he arc region. Moreover, the cathode spot moves with a high velocity over '·the cathode surface {Daalder, 1978). It is a zone or high particle density, and ~s consequently difficult to investigate experimentally.

For a vacuum are of a single cathode spot at low currents (

<

100 A for copper). the are stability plays an important role. At a certain current the vacuum arc spontaneously extinguishes after a certain duration from the ignition, which is the "arc 1 He time". In. power frequency current interruption, the failing stability of the arc causes the arc current coming to zero abruptly and prematurely before the natural zero. This phenomenon is known as "current chopping". The AC current chopping phenomenon has been found to be governed by the instability of the DC arc at low currents (Smeets, 1987).

As a consequence of the current chopping, large overvol tages can ·be generated over the vacuum interrupter. This is not only harmful to the rest of the network components but also may cause multiple reignitions when the gap length is small. Such reignitions, on the one hand, protect the network. But on the other hand they may be the cause for generating new and higher overvol tages. This phenomenon is called "voltage escalation"

(Greenwood, 1980).

2-1.3. Purpose of the present investigation

A better understanding of the essential processes occurring during a vacuum arc interruption, will be helpful for the further development of the vacuum circuit breaker. It is therefore the purpose of the present investigation to contribute to the knowledge on cathode spot behaviour, current interruption, dielectrical recovery and overvoltage generation.

This work consists of two parts. The first part deals with the single cathode spot behaviour of a OC vacuum arc (chapter 2, 3 and 4) which is relevant to the current zero phenomena. In this part, particular attention has been given to the influence of the cathode surface microstructure. It starts with an experimental study on the DC vacuum arc stability and the arc voltage in relation to the cathode surface microstructure. Then the mutual relations between the transient instabilities in the arc voltage and cathode spot motion in relation to the cathode erosion pattern are studied. A qualitative physical interpretation is given to explain the experimental results.

The second part of this work deals with the high frequency current interruption behaviour at small gap length {chapter 5 and 6) which is relevant to application aspects. From a large number of experiments on high frequency current interruption ability and dielectrical strength of the vacuum gap after current interruption, data have been obtained for the draw up of the interruption and the reignition criteria. These results are introduced into a computer program which calculates the overvoltages generated during the operation of a vacuum circuit breaker for a practical single phase circuit.

PART mE:

-2. DC VAaruM ARC STABILI'IY AND ARC VOLTAGE

2.1. Introduction

The influence of the cathode surface condition on vacuum arc behaviour can be distinguished by two aspects for a given cathode material: the surface contamination and the surface microstructure.

The influence of surface contamination on various properties of vacuum arcs

has been surveyed by JUttner et al. {1982): on contaminated cathode

surfaces there is predominantly one type of cathode spot with a fast motion

and a small crater size (the so--called type I spots); on the clean cathode surface there is predominantly another kind of cathode spot with slow motion and large crater size (the so--called type II spots) .

JUttner has found that the influence of surface microstructure is reduced after sufficient surface cleaning (JUttner 1984). Daalder et al. (1973) has

shown that the arc voltage has different values when using polished. half-polished and roughened copper cathodes.

In this work. the electrical roughness and geometrical roughness of a contact surface will be distinguished. Since i t is well known that the smoother the cathode surface. the higher the breakdown voltage. the vacuum circuit breaker manufacturer usually makes the contact surface as smooth {electrically) as possible. But it is doubtful whether the vacuum interrupter with more smooth (electrically) contact surface has better characteristics for the current chopping level with respect to the overvoltage generation. Therefore, it is worthwhile to study systematically the influence of the degree of contact surface roughness on the arc stability {directly related with current chopping), arc erosion, etc ..

2.2. Experimental circuit

The experimental circuit is shown in Fig. 2. 1. The OC supply consists of six batteries (total voltage of 72 V). The master switch (MS) limits the maximum time of current flow. Its closing duration is normally 100 ms. The variable carbon moulded resistor (R) is used for adjusting the arc current. The circuit inherent inductance (L) and parastic capacitance (C) are about 20

pH

and 400 pF respectively. The arc current is measured by means of the

I

shunt (Rs, 2. 78 mO). The test vacuum interrupter is indicated as VB. This vacuum interrupter consists of one fixed and one movable contact. The arc is initiated by separating the movable contact apart from the fixed one. The average contact opening speed is about 2 m/s. The final contact distance is 5 11111.

The arc voltage and arc current are measured (at V and I post tions in Fig.2.1) by a LeCroy 9400 oscilloscope with a sampling frequency of 100

IlHz.

Fig. 2.1. The experimental circuit.

2.3. Contact surface condition

The contacts are made of OFHC copper. Their configuration is shown in Fig. 2.2. The diameter of each contact is 30 11111. The edge curvature radius is 1.35 1111. Both contacts have a flat surface. Before assembUng. the contacts

have been cleaned in an absolute ethanol (C2HsOH) bath with high frequency vibration for five minutes. The vacuum interrupter has been degassed and baked out for at least 16 h with the maximum temperature of 400 °C. The inside pressure is then lower than 10-6

Pa.

l

~

r=30mm=-:)

Fig. 2.2. The contact configuration.

2.3.1. Contact surface microstructure

The contact surface roughness has been varied by using different emery

-papers (P60. P220. P400 and

llSoo.

these numbers mean that there are 60, 220, 40 and 800 silicon particles in 1 wa2 of the emery paper) and a rotating polishing disc (grain size 1 pm).

The

aver.age local electric field enhancement factor {j from the Fowler-Nordheim theory (lafferty, 1980) is taken as a measure for the degree of contact surface roughness electrically.

The

electric field can be expressed:

E

=

{jV/d (2-1)

where V is the voltage over a pair of parallel electrodes and d is the distance between two electrodes.

The Fowler-Nordheim equation describes the current density from a field emitting tip as a function of the electric field at the tip. For practical application the F-N equation can be written as:

aE2 brp3/2

J

=

"""9>

exp(- E ) (2-2)

where a

=

1.54x10-6

AJV-

2

• b

=

6.83xl09 VT3/'2m-1 and , the work function (in eV) of the electrode metal (4.5 eV for copper}.

Substituting Equ. (2-1} .into Equ. (2-2} and dividing the total field emission-current by an average emitting area {S} instead of J, Equ. {2-2) becomes:

(2-3}

The

slope of the so-called Fowler-Nordheim line or "F-N plot", i.e. I

ln(V/d)2 as a function of (V/d}-1, can be obtained experimentally. Then {j can be derived as:

- - brp¥2

{j - slope (2-4)

In the experiment, the gap distance was measured to be 0.25 mm. The emission current was limited to 500

JJA.

The maximum applied voltage over the gap was 5 - 10 kV. The measured {j value is only of order of magnitude,

cause inaccuracy in the gap distance measurement. Nevertheless it bas been found that the magnitude of Jj is larger for the contact surface treated with rougher emery paper. According to the Jj value, the contacts have been classified and named after.

The geometrical roughness of a contact surface bas been measured by the surface scanning method wi tb a metal probe. The measuring resolution is

1 J.UD. The geometrical roughness profiles are shown in Fig.2.3.

The horizontal axis indicates the dimension along the contact surface. The vertical axis indicates the height of the local position with respect to the center line {dotted line in Fig.2.3).

A COIIIIIOnly used geometrical roughness parameter is the centre-line average roughness height (R ), which is defined according to DIN standards as:

a

{2-5)

where 1 is the total scanning length in the x direction and b{x) is the

Ill

height at x post tion.

1fe

geometrical roughness Ra and the electrieal roughness Jj for six contact surfaces are listed in table 2.1.

The

way of obtaining the desired surface is also described in this table.

Table 2.1 The geometrical roughness R and the eletrical roughness Jj for a

six contact surfaces

surface very rough roughened rough normal SIIIOOth eroded

Ra {pm) 4.2 1.2 0.8 0.4 0.02 5.0

>800 800 600 400

(*)

(**)

treated

by P60 P220 P400 PSOO

(*): polished by a rotating polishing disc, plus an additional conditioning up to a ~~~a.xiiiiUIII voltage of 10 kV and an emission current lower

9500

pA. at the gap length of 0.25 111111.

(**): starting with "rough surface and subs~quent!Y processed by DC arcs with the total . sferred charge of~ leaving 80% of the cathode surface covered by arc traces.

-20 10 0 -10 -20 {a) -30 6 0 500 1000 1500 2000 2500 3000 3500 41XXl 4 2 0 -2 -4

_{_,}

_(b) 1500 0 500 1000 2000 2500 3000 3500 4000 4 2 0 -2 -4 0 200 400 600 800 1000 1200 1400 1600 1800 2000 2 E !:l, 1-0

~

-1 (d) 200 400 600 800 1000 1200 1400 1600 1800 2000 ~0+---~--~--~----~--~---r--~---,----.-~ 0 200 400 600 BOO 1000 1200 1400 1600 1800 2000 X !)Jm)

Fig.2.3 Geometrical rouglmess profiles for the contact surfaces of "very rough" (a), "roughened" (b). "rough" (c), "normal" (d), "smooth" (e) and "eroded" (f).

'1

·~

It can be seen from table 2.1 that the more geometrically rough the surface is; the more electrically rough the surface is except for the "eroded" surface. The "eroded" contact surface is geometrically extremely rough but electrically rather smooth.

2.3.2. Qonta.ct surface contamination

The release of absorbed contaminating' gases of the contact surface has been checked indirectly by mass-spectrum analysis. A Balzers

QMS

311 Quadrupole mass spectrometer was mounted on the vacuum interrupter as shown in Fig. 2.4. analyzer ion soLrCe 1 grid detector I

Fig. 2.4. Schematic diagram of vacuum interrupter and mass analyzer.

A. Residual ms analysis

By energizing the filament with a certain current. the ion source is applied. The neutral species of the residual gases in the vacuum interrupter will then be ionized. These ions pass through the analyzer. then to the detector. This ion current can be enhanced by the electron multiplier and finally be measured with an oscilloscope. Fig. 2.5 shows a signal of the residual gas in Ampere unit. The pressure inside the interrupter was lower than 10-4

Pa.

One can see from this figure that there are five humps in the ion current. + Their mass numbers and possible ions are respectively: 1) Mle = 2 for H2 ;

+ + + +

2} Mle

=

12 for C ; 3) Mle

=

14-18 for CHn and H20 ; 4) Mle

=

28 for CO or

+ +

N2 ; and 5) Mle = 44 for 002 .

4

10-1oAJ

1 !i.

l

2 3

_...

---"

0 10 20 3) 40 ~ t!J~e

Fig. 2.5. Residual gas species from the interrupter.

B. The products from the contact surface released by the arc

As described in 2.3.1. the different degree of contact surface roughness is achieved by treatment with different emery paper or a rotating polishing disc. The grinding material of these emery papers is silicon carbide (SiC).

The polishing material of the rotating polishing disc is a diamond compound. AI though all contacts have been well cleaned and baked out after the mechanical treatment, some particles from the polishing material which were pressed in the contact surface layer could still be there. These particles are one kind of the surface contamination. When an arc is drawn between the contacts, these impurity particles will be heated and evaporated by the cathode spots together with the contact metal. Then they can be examined by the mass spectrum analysis.

Since it can not be avoided that roughening a contact surface will cause the surface contamination, a question arises: do the contact surfaces with a different degree of roughness have the same degree of contamination? To answer this question, dynamic mass spectrum analysis have been done before, during and after a 10 ms OC arc (30 A).

First, the mixture of neutral and ion species has been detected. This means that all species, received by the mass analyzer, are ionized by the ion source first (see Fig.2.4) and then detected. Fig. 2.6.a. shows the repeated mass spectrum with mass number between 0 to 70. The first cycle

was run before arc ignition: the second one was during the arcing, and the rest three cycles were run after arc extinction. The scanning speed of the mass spectrometer is one mass number per millisecond. It can be seen in the second cycle that ~ is detected. This is only a small part of the total

amount of ions because the arc lifetime was nruch shorter than the total scanning time. The noise between the second and third cycle of Fig. 2.6.a. is caused by the vibration due to contact movement.

(a) _Cu++ 50ms (b)

o40A

.J:':~·.,

.A

0 10 20 30 40 50 60 70Hie

Fig. 2.6.a. A repeated mass spectrum of neutral species. Mass number (Mie) for each cycle is from 0 to 70. A "normal" contact surface of copper is used.

b. Enlarged signal of first and fourth cycles from a.

Fig. 2.6. b. shows the enl-arged mass spectra of the first {before arc} and the fourth (after arc) cycle. The increase of the impurities by arcing can be seen. The largest yields were at mass number 28. They could be

ro

or N2 , or Si from the cathode surface.

2

.•

'

5 10 15 20 2S 30 35 40 ltSMie

Fig.2.7. The impurity yields increased by arcing, AI = id- ia. "o" - "smooth" surface; "o" - "normal" surface; "x" - "rough" surface; "e" - "roughened" surface.

Fig. 2. 7 shows the impurity yields increased by arcing. The arc current was 30 A. The data are' the average values of the first five measurements. The choice of a low· number. of measurenaents is made to· avoid contact surface conditioning by .the ~urement itself, The impurity yield before and after

-arc (ia and id) is defined as the peak height of ion current as in Fig. 2.6.b.

One can see in Fig. 2.7. that the impurity yields increase are nearly the same for four contact surfaces with diffe~ent degree of roughness. Therefore the answer to the question proposed earlier could be: the contamination is not correlated with

the

roughness.

~

10-BA

0 10 20 J) 40 g) 60 70 I"Ve

Fig.2.8. Mass spectrum of the copper ions directly from the arc of 30 A. "roughened" contact surface was used.

Secondly, the ion species have also ~n detected (without use of the ion source}. From the measurements, only copper ions were detected. and they can be one of:

cli.

~.

a1+

and

o¥.

Fig. 2.8 shows an example of ion yield during arcing. The amplitude of the copper ions signal is in the order of

3x10-a A. Comparing· Fig. 2.6 with Fig. 2.8, it can be found that the maximum impurity yield is about 2X of the copper ion yield by arcing.

2.4. Arc lifetime measurement

It is known, that the DC arc lifetime is directly related to arc stability. There are many factors affecting the arc stability (Lafferty, 1980; Smeets, 1987). This section only focuses on the influence of contact surfaee microstructure on DC arc lifetime.

Fig. 2.9 shows typical arc voltage and arc current oscillograms for the contact surfaces with a different degree of roughness and the same DC arc current. The arc lifetime T is defined as the duration from the arc

ignition to its spontaneous extinction. and T is indicated in Fig. 2.9.c.

Four contact surfaces have been used: "smooth", "normal". "rough" and

"roughened".

-Fig. 2.9. The arc voltage (upper trace) and arc current (lower trace) for some contact surface microstructures. 1=10 A. a. "smooth": b. "normal". c. "rough". and d. "roughened".

10

lmsl

0.1

0,01

4 5

8 10 15 20 25 3l 40 l!Al

Fig; 2.10. Average arc lifetime {i} versus arc current (I) for several degrees of roughness. "•" - "roughened": "x" - "normal"; ''o" - "smooth"; "A" - "eroded".

versus arc current for "roughened". "normal", "smooth" and "eroded" contacts are shown in Fig. 2.10. The experiments were carried out by increasing the current, in order to reduce surface conditioning by charge transfer from the measurement itself.

I"OllJhene1

o.s

12.8A 2 5 15 20 li 10 15 20 25 'T 180 smooth 26C

lJy4

(JJs)160 22C 140 160 A 10 15 20 25 5 10 15 20 25 N N

Fig. 2.11. The typical distribution of four kinds of contacts. T is the arc life. N is the number of measurement.

From Fig. 2.10 it can be seen cl~rly that for a certain arc current in the range of 4

<

I

<

20 A. the average arc lifetime is in following sequence according to the contacts used:

T

(roughened}

»

T

(normal)

> T

(smooth}. The average arc lifetime for "smooth" contacts is n~rly the same as that for "eroded" contacts. In the current range above 20 A, the average arc lifetimes for four kinds of contact surface do not differ significantly. The individual arc lifetime measurements have some statistical spread as can be seen from Fig. 2.11. There is an apparent decrease of limetime during measurement on the "roughened" contact and in a less obvious way on the "normal" contact. But there is no such tendency on the "smooth" and "eroded" contacts. the lifetimes are "constant" in time in both cases. From this point of view, the arc 1 tself has a tendency to reduce the arc lifetime on rough surfaces.

2.5. Arc voltage measurement

The arc voltage. as well known, is strongly dependent upon the cathode material. Arc voltage dependence on other factors, such as the electrode

diameter. electrode gap distance, and axial magnetic field etc. has also

been found (Mitchell, 1970; Kimblin, 1969). Here only the influence of

contact surface microstructure will be studied. The arc voltage for low-current arcs consists of a certain DC level and a SUPerimposed high-frequency (HF) component (Smeets, 1988). Both DC and HF components of

the arc voltage are indicated in Fig.2.9(c).

2. 5 .1. DC CO!!!QOnent

Figure 2.12 shows the volt-ampere characteristics of the arc voltage DC component. with the degree of contact surface roughness as a parameter. Five contacts have been used: "very rough", "roughened", "normal", "SIIOOth" and "eroded".

The slopes for the four different contacts are approximately the same, except for the "eroded" contact. The arc voltage at a given current is higher for smoother surface. The difference of more than 4 V between "smooth" and "very rough" contacts is remarkable. The arc voltage of the "eroded" contact is near

io

that of the "SIIOOth" one.

Detailed measurements of the behaviour of the V-I characteristic for the "roughened~' contact influenced by an extra transferred charge Q, are shown in Fig. 2.13. Each series of V-I measurements, composing one line, is

taken

by increasing the arc current after a certain charge transfer (Q = I:»r, I

=

31 A). It can be seen that the arc voltage for the "roughened" contact increased with the total transferred charge.

v

20 !Vl 18 14 A_,---·r--·· 10

__,_

..

-_.t>,--·• 20

..

...--30 UAl

Fig. 2.12 The arc voltage DC component versus the arc current for five contact surface microstructures. A - "very rough"; B -''roughened''. C - "normal", D - "smooth"; E - "eroded".

Fig. 2.13 V-I characteristics for an "eroded" contact and for "roughened" contact varies with the extra transferred charge

Q,

following each measuring series from low to.high current.

2.5.2. HF component

From Fig. 2.9, the difference in HF behaviour and lifetime for different contact surfaces is obvious. It has been noticed (Smeets, 1987), that the

peak value and frequency of occurrence of the voltage spikes is strongly

related to arc lifetime and current chopping of vacuum interrupters. The results of measurements on the frequency of occurrence of arc voltage spikes higher than a given voltage level for some arc currents and some contact surface microstructures, .are shown in Fig. 2.14. The contact surfaces used were: "roughened", "smooth" and "eroded".

It is clear that the arc voltage spikes are higher and more frequently appearing for a lower arc current at a given contact surface microstructure. They are also higher and more frequently appearing for a smoother contact surface at a given arc current. The eroded contact surface again shows a similar behaviour as the smooth surface in this respect at currents larger than 30 A.

When the arc current is higher than 30 A. the occurrence of the arc voltage spikes is influenced only by the arc current. This is because that, the contact surface conditioning effect by the arcing itself is dominating. Figure 2.15 shows the frequency spectra .of the arc voltage for a "smooth"

contact surface with the, arc current of 30, 50 and 70 A, obtained by Fourier transformation. It can be seen that, the higher the arc current is;

the lower the high-frequency component (with respect to DC component) is

and the less the arc voltage spikes there are. The influence of the contact surface roughness on the arc voltage HF component (or on the occurrence of arc voltage spikes) becomes less important at an arc current above 30 A.

103

runbersd

spikes jill" ms

10

2) 30 40 50 (i) 80 100 .120 voltoge(VI

Fig. 2.14 The frequency of occurrence of arc voltage spikes higher than a given voltage level for some arc currents and contact surfaces.

"o ... -"eroded"' ... x"'-"'smootb'*; '*A .. -"roughened ....

-Fig.2.15 The frequency spectra of the arc voltage for a "smooth" contact surface with the arc currents of 30 A (a). 50 A (b) and 70 A (c) (vertical scale shows the relative amplitude of various frequency components).

2.6. Cathode erosion pattern

Fig. 2.16 shows the cathode erosion patterns for "smooth", "normal" and "roughened" contacts left by arcing. The different eroded areas on the cathodes of different surface microstructure have been observed. The arc

eroded area has been estimated to be 15% for the "smooth" cathode, 30% for the "normal" cathode and 60% for the "roughened" one. The total transferred charges were 50 C for the "smooth" cathode, 34 C for the "normal" cathode and 32 C for the "roughened" cathode respectively. It can be seen that the total eroded area on the "smooth" cathode is smaller comparing with the total eroded area on the "roughened" cathode. A more detailed observation has been done by a scanning electron microscope (SEM). Figure 2.17 shows two SEM photos of both "smooth" and "roughened" cathode surface after arcing.

Fig. 2.16 The cathode surface after arc lifetime measurement. a. "smooth", b. "normal" and c. "roughened" cathode surface.

Fig. 2.17. SEM photos of a "smooth" (a) and a "roughened" (b) cathode surface after arcing.

-It can be seen that the craters on the "smooth" cathode are large and

overlapping, while the craters on the "roughened" cathode are small and

separated.

2.7. Qonclusions

It is necessary to distinguish between an "electrical" and "geometrical" roughness of a contact surface by means of the electrical field enhancement factor

p

and the centre-line average roughness height Ra respectively. For an unarced contact surface, a geometrically rough surface is also an electrically rough one; while for an arc eroded contact surface, a geometrically rough surface can be an electrically smooth one.

By mass spectrum analysis, it is found that the treatment on a contact surface with the emery papers with different degree of roughness does not significantly influence the contamination of the contact surface.

The rougher the cathode surface, the longer the arc lifetime and the lower the arc voltage (both DC and HF components), thus giving a more stable arc. The arc itself has a tendency to condition the cathode surface, which in the way is unfavourable for a sustained arc lifetime, and also increases the arc voltage.

The crater size for a rougher (geometrically) cathode surface is smaller than for a smooth one, but.the number of craters is much larger, and they are more separated. Thus the area where arc traces are found is larger for a rough cathode surface, compared with a smooth cathode.

The conclusion for the electrical behaviour during interruption is that each geometrical rough contact surface will be electrically smoothed by arc conditioning. Its behaviour will be transferred towards that of an eroded contact. this means towards a higher current chopping level but also higher breakdown voltage. The eroded state is already reached after a charge

3. IIOfiON OF CA11100E SPOI'

UNDER

TRANSVERSAL MAGNETIC FIFLD

3.1. Introduction

A prominent feature of the cathode spots of a vacuum arc is their rapid and erratic movement over the cathode surface. The movement causes a variable energy exchange between the spot and the cathode surface in the form of

local eaission and erosion, leaving traces of craters and droplets.

In previous investigations. it has been found that the motion of the cathode spot strongly depends on the cathode surface condition and so does the cathode erosion structure (Sethuraman and Barrault. 1980: 1982}. It has also been found that the arc voltage is influenced by the cathode surface microstructure (chapter 2 of this work).

In practical application, Szulczyk and coworkers (1986) found that the current chopping is tightly related with the arc erosion microstructure left by the molten layer on the cathode.

The aim of this chapter is to study the correlations among the cathode spot motion, the arc voltage and the cathode erosion microstructure and to relate thea with the cathode surface microstructure.

In this work, two kinds of contact materials have been studied: OFHC copper and copper chromium (CuCr) with weight 25X of chromium content. Cu is a basic and simple contact material often selected for research purposes because most of its physical properties are already known. The metal alloy

CuCr is extensively applied as a contact material in modern vacuum circuit breakers because of its following advantages: low chopping current level, low welding tendency. excellent recovery and voltage withstand characteristics, as well as large current interrupting capability (Ok:awa and coworkers, 1987).

In order to conveniently observe the cathode spot motion and the arc traces on the cathode surface, a magnetic field is applied parallel to the contact surface. The cathode spots will then be forced to move in the retrograde direction. i.e. in a direction opposite to that expected on the basis of Ampere's law for the electron current direction.

-22-3.2. Experimental set-up

The experimental circuit is shown in Fig. 3.1. The arc, is ignited by applying a post tive voltage pulse (6kV, 400 ns} at the trigger pin (TP}, and is then supplemented by a precharged capacitor (C, 5 f.!F} across the gap, and is finally supported by a set of batteries of 144 V. The duration of the supplementary current can be adjusted by the thyristor {T}. The arc current is limited by a resistor (R). The applied transversal magnetic field is generated by a Helmholz coil placed outside the vacuum chamber (VC}. The magnetic induction along the cathode surface is approximately uniform and its value can be up to 0.15 T.

Fig. 3.1. Experimental circuit. VC-vacuum chamber; TP-trigger pin; MS-master switch: D1 , D2-diodes: T-thyristor: Rs-shunt: r1 • r2 , R-resistors: C-capacitor.

Fig. 3.2. Contact configuration. GO-gear drive: TP-trigger pin: ~etic induction; v-spot velocity.

The contact configuration is shown in Fig. 3.2. The anode is a quarter of a disc and the cathode is a full disc of diameter of 4 em which can be rotated by a gear drive connected to the outside of the vacuum chamber. The cathode spot moves from the center to the edge of the cathode surface with a transversal magnetic field, the arc trace is restricted to a narrow

/

region. This contact configuration has the advantage, that many arcs can be initiated, each one on a virgin area before opening the vacuum chamber to observe the individual arc traces. The processes of the contact cleaning and the vacuum chamber degassing are the same as described in chapter 2.

The displacement of the cathode spots was recorded on a ( 400 ASA) 35 mm film by means of a rotating mirror type ( HITACHI sp-1) high speed streak camera. The arc voltage and current were recorded by a LeCroy 9400 digital oscilloscope with sampling frequency of 100 MHz. The detailed observation of the crater traces after arcing was performed by means of a scanning electron microscope.

3.3. Average velocity of cathode spots

The average velocity of the cathode spot motion was determined from the length of an arc trace divided by the arc lifetime. Fig. 3.3. shows such arc traces on both "smooth" virgin OJ cathode (a) and OJCr cathode (b).

The ~c s on "smooth" virgin cathode of a. Copper,' b.

~ ~

Copper-chromium. Fig. 3.3.

For the virgin OJ cathode. the results of the average velocity of the cathode-spot for three kinds of surface microstructures as a function of magnetic field are shown in Fig. 3.4. The cathode surface used were:

"smooth". "rough II" and "rough .L". The "rough II" and "rough .L" surface is

obtained by grinding the cathode disc in the radial direction and in the tangential direction respectively (grinding material is the emery paper P400).

From Fig. 3.4. the influence of the surface microstructure on the average

-velocity of the cathode spot shows to be significant. The average -velocity of the cathode spot for a "smooth" cathode is about ten tiJDes sDBller than that for a "rough II" cathode for a saJDe given are current. magnetic

induction and gap distance. Moreover, the direction of the grinding also creates SOJDe influence on the average spot velocity.

When

the direction of the grinding is in radial direction of the cathode disc, the average spot velocity is about 2-3 tiJDes larger than that when the grinding is in tangential direction of the cathode disc.

The JDeaSUreJDent of the average velocity of the cathode spot (v) on an arc eroded copper cathode has been done by Fang (1980). His result showed that v increases linearly from 3 m/s to 5 m/s with the increase of B from 0.06 to 0.1 T. at a 60

A

arc current and 1.5 DID gap distance. This result is

very close to our result for a virgin "smooth" cathode as shown in Fig.3.4 although the experimental conditions are different.

v

200

1

I I )100

1

m s SO

-}---I--I--i--1---1-20 10 5 2 SO 60 70 80 90 100 B(mTl

Fig. 3.4. Average velocity of cathode spot (v} versus magnetic inductance (B) for Cu cathode with three kinds of surface microstructures. The arc current is 50 A and the gap distance is 3 DID. "1"-"smooth". ""-"rough II" and" "-"rough .L".

The influence of the cathode surface condition (both surface contamination and microstructure) on the average velocity of the cathode spot motion is shown in Fig. 3.5 for a CuCr metal with "smooth" surface, at an arc current of 50 A, gap distance of 3 DID and magnetic induction of 0.04 T. It can be

seen that the average velocity of the spot motion (v) decreases with the number of JDeaSUreJDent (N) from 10 - 100 m/s to 4 - 7 m/s within N<lOO.

After one hundred of measurements (total transferred charge is less than 5 C). an extra total charge of 100 C was transferred. Then the average velocity of the spot motion is more or less a constant around 3.5 to 6 mls. Therefore, the more charge is transferred, the lower the average velocity of spot motion is.·

200

v

100 0

°

0 0 (mJsJ 50 0 0 0 0 0 _{extra 100(} 20 oo odbodo 0 transterrerl oO ao o

l

10 0 oo 0 ooooo ' 5 'l> 0~ 0 0 0 Rf<xll'!l9 0 0 ~q9qj 0

°

20 40 60 80 100 120 140N

Fig. 3.5. Average velocity of the cathode spot versus the number of measurement for CUCr cathode, d = 3 Dill, B

=

0.04 T, I =50 A. The average velocity of the spot motion versus the magnetic induction on an arc eroded CuCr cathode is s~ in Fig. 3.6. The gap distance was 3 11111 and three current values were used: 25, 50 and 75 A. It can be seen from this figure, that the log(v) increases with B linearly for a certain arc current. The slopes of log(v) versus B for three current levels are approximately the same. For a certain magnetic induction, the higher the arc current is, the higher the average velocity of the spot motion is.

v

100 lm/slso 20 10 5 2 0.08 0.12 B!Tl

Fig: 3.6. Average velocity of the cathode spot for sever!!l arc current!(! on eroded

Cue-f.

cathode,. d = 3 mm ..

-26-v

lm/sl 20

10

2 3 4 dlmml

Fig. 3. 7. Average velocity of the cathode spot versus gap length for arc eroded CuCr cathode, I = 50 A, B = 0.08 T.

Fig.

3.7

shows the average velocity of the spot motion (v) as a function of gap distances (d). A positive linear relationship between

v

and d bas been found for a certain arc current and magnetic induction.

The following empirical relations can be obtained from Fig.3.6 and Fig.3.7:

38.5B

v = a e 0

<

B

<

0.1 T (3-1)

where a= 0.5, 1.1 and 1.7 {mls) when I = 25, 50 and 75 A respectively. And for I = 50 A. B = 0.08 T:

- 3

v

=

5.88x10 d. 0

<

d

<

4xl0-3 m and

v

in mls (3-2) 3.4. Influences of the cathode surface microstructure on arc voltage and

on erosion patterns

From the arc voltage measurements and cathode erosion observations for copper metal. distinct phenomena have been found for the "smooth" and "rough !" cathode surfaces, as shown in Fig. 3.9. The magnetic induction (B), the arc current (I) and the gap distance (d) for both cases are the

(u) (c)

Fig. 3.8. The arc voltage oscillograms and typical erosion SEM pictures for copper metal with a "smooth" cathode surface: (a), (b), and with a "rough .l" cathode surface: (c), (d). Arc current of 50 A, magnetic induction of 0.1 T and gap distance of 3 nun. For (a) and (c): vertical scale of 25 V/div. and horizontal

scale of 12.5 ~s/div.

First type, on the "smooth" cathode: the arc voltage consists of a low

oc

level and a few spikes in high-frequency component; the crater size is

large and the craters are overlapping. Second type, on the "rough .l"

cathode: the arc voltage has many spikes with high amplitude in

high-frequency component; the craters are more separated and are formed on the convex ridge of the grinding trace in most cases.

Therefore, the appearance of spikes in arc voltage is likely to coincide with the appearance of separated craters. This is just the opposite of what has been found in the case without applied magnetic field. For detailed observation of the crater trace at the moment when an instability peak appears in the arc voltage, a feasible way for copper metal is to use a

28-cathode with both "smooth" and "rough l." surface microstructures. This was obtained by polishing and grinding different parts of the same cathode surface.

3.5. Correlations among arc voltage. cathode spot yelocity and cathode erosion microstructure

3.5.1. On virgin

Cu

cathgde

A cathode has been used with a special surface structure. In this, a number of groove regions on the cathode surface are provided. Each concave groove region has a width of 0.2 BIB and a depth of 0.1 BIB, the surface

microstructure is similar to the "rough l." surface described before. The rest of the cathode surface is "smooth". During one arc lifetime, the cathode spot will be forced by the transversal magnetic field to pass through these groove regions and "smooth" regions. With the aid of streak photographs, which showed the position of the cathode spot as a function of time, the arc voltage oscillograms were related to the SEM pictures of erosion traces on the cathode. An example of the results is shown in Fig. 3.9. Figure 3.9.{b) is enlargment of Fig. 3.9.(a). The arc current was 50 A, the applied magnetic induction was 0.05 T, and the gap distance was

3 BIB.

There are three typical erosion structures and arc voltages which can be distinguished as indicated in Fig. 3.9.{b).

{1} In region I {on the "smooth" cathode surface}: from the macroscopic view, the spot motion seems to be dominately governed by the magnetic field as a more or less straight crater trace shown in Fig. 3.9.(a}. But from the detailed observation, the erosion trace is wide and consists of many large

and heavily overlapped craters. This means there still ~xists some random motion in the direction perpendicular to the retrograde motion. The average velocity of the cathode spot in this region is about 2.5 m/s. The correlated arc voltage has a high-frequency component not higher than 10 V

and a OC component of 30 V caused by the magnetic field in the "column" region. It seems that a new emission site can more easily be formed at the old crater's edge in the direction of motion than in the zero-field case, where a high HF component in the voltage is present (see Fig. 2.8.a).

1mn

Fig. 3.9. a. The cathode erosion trace and correlated arc voltage oscillogram. Arc current of 50 A, magnetic induction of 0.05 T and gap distance of 3 rmt. Vertical scale: 40 V/div. and horizontal scale: 250 ~s/div.

b. A part enlarged from (a) and a synchronous streak

photograph. Vertical scale: 30 V/div. and horizontal scale:

28 ~s/div. {for both voltage and streak photo).

30-(it) In region II (on a scratch with a direction the saJDe as the spot motion): the cathode spot motion is facilitated considerably. The erosion trace is of about the same width as the scratch. The correlated arc voltage oscillogram during this period also suggests an efficient process of cathode spot rene11q1.l. This is because that the tendency of the cathode spot motion enforced by the magnetic field is the saJDe as that enforced by the cathode surface irregular! ty. New emission sites can easily be formed along the scratch. This aakes the vel oct ty of the cathode spot extremely high.

(iii) In region

III

(in the "ring-shape" region of cathode surface): the emission trace is discontinuous and goes in the perpendicular direction. Only a few craters can be found. This means that the cathode spot moves erratically in this region with a velocity of about 15 mls in the retrograde motion direction. The correlated arc voltage is strongly varying with peaks even higher than

80 V.

For the same reason as in region

II.

the cathode spot tends to move in the direction of the grinding. On the other hand, the retrograde ampere force of the applied magnetic field tends to push the cathode spot out of this region into the "smooth" surface. Because of a lack of field-enhancing microprotrusions. a

new

emission site will be formed with difficulty on the "smooth" surface. Thus the spot motion tendency enforced by the magnetic field has to compete with the tendency enforced by the surface irregular! ty. During this competition the discharge is unstable, leading to a peak in the arc voltage. This is correlated with the dark part in the streak photograph of Fig. 3.9.b. When the B field is strong enough, the cathode spot will move. to the "smooth" region .and the arc voltage might be stabilized and go back to a low value. I£ not, the arc voltage 111ay build up to exceed the supply voltage and then the arc extinguishes as occurs in the fourth of the "ring-shape" region in Fig. 3.9.a.

3.5.2. On virgin

CuCt

cathode

1. Time-resolved cathode soot velocity

The first step is to record the displacement of the cathode spot on film. Since the luminous source (of the cathode spot} is located at the cathode, the camera image is a view parallel to the cathode surface through a 0.5 • wide slit. The experimental parameters were: arc current 50 A. contact gap

length 3 IIID and magnetic induction 0.06 T, and camera recording speed of about 3 J.lS/IIID. Fig. 3.10 shows an example.

The second step is to transform the optical image on the film ·into nUIIIerical data, suitable for computer analysis. The position (xi) along the direction of motion as a function of time (ti)

was

measured from film (800 points in total) with the help of a ten times enlarging projector. The resolution is 1 ~s in time and 0.1 IIID in position.

The third step is to calculate the velocity from the post tion data. The time resolved cathode spot velocity v{t)

was

computed by using the lagrange's rtve-point formula (Kreyszig 1972):

{3-3)

where At = ti+₂ - ti+l : t₁_₁ - ti_₂. Figure 3.10.b shows the result obtained from Fig. 3.10.a. The time dependence of the velocitY of the cathode spot can be clearly seen. The waveform of v(t) consists of many pulses. This may indicate that, during most of the time the cathode spot moved relatively slow or stayed stationary, and the spot moved fast only during short time periods. The appearance .or this "fast motion transients" is irregularly.

2. The correlation between cathode soot yelocity and arc voltage

An arc voltage signal recorded simultaneously with the streak photograph is shown in Fig. 3.10.c. A method to check a correlation between this cathode spot velocity v(t) and arc voltage u{t) is to .calculate their cross correlation function as follows:

~B«{t) =

t

J!

V(T-t)U(T) dT {3-4) In this calculation, bOth signals V(t} and U(t} are the AC components of v( t) and u(t).

Therefore, the more pronounced a peak in ~ is, the better the correlation

Ba

of spot velocity and arcing voltage is. For T=800 ~s. four pairs of signals have been checked and the computed results are shown in Fig. 3.11.

-32-80 -u(V!

6S-Fig. 3.10. Simultaneous streak photograph {a), spot velocity {b), arc voltage {c), spot brightness {d), and SEM pictures {e) for a

It can be seen clearly that the maximum of the """' function is at t:O. Thus, the spikes in the spot veloei ty v( t

t

seem to occur synchronously with the spikes in the. arc voltage u( t) . However, the amplitudes of the HF spikes of u{t) and v(t) show no obvious correlation.

Fig. 3.11. The cross-correlation function of spot velocity and are voltage.

3. The correlation between cathode soot velocity and soot brightness

film

l[

(Q)

Fig. 3.12. Scheme of the cathode spot brightness measurement.

The spot brightness measured here is a relative value during the arcing lifetime. It was obtained by measuring the relative darkness of the spot signal on the negative film. The evaluation method is shown in the scheme of Fig. 3.12.a. A 10 times enlarged image of the streak film is projected onto a video camera. scanning the image with 625 lines. The electrical signal of each odd line (312 in total) is evaluated according to Fig. 3.12.b. From the oscillogram of a single-line signal, the darkness of the

-film (thus the brightness of the cathode spot} can be obtained by measuring the relative amplitude of the corresponding peak. In this way. the spot brightness Br as a function of time was extracted in Fig. 3.10.d. The time

interval between the measured data was 0.5 J..lS.

Comparing Br(t) with the spot velocity v(t) as in Fig. 3.10 b. a coincidence of peaks in v(t} with dips in Br(t) can be found. This can also be checked by the calculation of the cross-correlation function:

ll .. (t-!h.)

=

t

J!

v(T-t)[l-Br(T)] dT (3-5)

The result is shown in Fig. 3.13 with the maximum of ~~ .. (1--!h.) at t

=

0. Since the spot velocity spikes are synchronous to the arc voltage spikes. The coincidence of dips in spot brightness with peaks in arc voltage is. however, an indirect effect. Due to the relatively low feeding voltage, a spike in arc voltage (AU) results a dip in arc current: AI= AUIR. From previous work it is clear that a declining arc current causes an

instantaneous decline of spot brightness (Smeets 1988).

1-0.5

Fig. 3.13. The cross-correlation function of spot velocity and spot darkness.

4. The correlation between cathode spot yelocity and the cathode erosion structure

From the observation of the crater trace. both large overlapping as well as small separated craters have been found. With the aid of the streak photograph which shows the spot position against time in one dimension,

each section of the crater trace (with 100 ~ uncertainty) can be related to the cathode spot velocity, arcing voltage and spot brightness at the moment this trace section was actively emitting.

Three typical SEM pictures of crater trace sections have been shown in Fig. 3.10.e and their correspondent time periods have been also indicated. In the case of Fig. 3.10.e.(1), a large number of small and separated craters can be seen. The cathode spot velocity and arc voltage was peaked, but the arc current and the brightness of the cathode spot decreased. In Fig. 3.10.e.(2) the craters are large and mostly overlapping. During the formation of the latter pattern, the spot velocity and arc voltage were low

and stable, the spot brightness was high and instable, and the arc current

was high and stable. For Fig. 3.10.e.(3), both large and small craters are present. The spot velocity, arcing voltage and spot brightness showed fluctuations.

From the detailed photograph (for example Fig. 3.14) of the cathode surface microstructure, it can be clearly seen, that the craters are preferably formed at the boundary of Cu and Cr. This is also observed by Burrage and Guertin (1973). Craters on the Cu regions of the contact are large and the spot motion seems to be slow in that place. While the spot moves on the Cr regions, the craters are small and the spot motion is fast and often along

irregular scratches. The area covered by the craters on Cr islands is

larger than on the Cu region. Concluding, the mobility of the cathode spot seems to be very high along the Cu-cr interface, moderately high on pure Cr and low on pure Cu parts.

Fig. 3.14. SEM picture of the craters on a CuCr cathode surface ( ti 1 t

400).

To explain the motion behaviour of the cathode spot on CuCr, an observation

36-of the virgin surface microstructure has been done and the SEM picture is shown in Fig. 3.15.

It can be seen that the "island" like Cr region is higher than the "sea" like Cu region on the CuCr surface. The cathode surface is not very smooth especially scratches are left still on Cr "islands" although the contact as a whole has been finely polished. The presence of scratches on the protruding Cr "plateaus" could explain the ease of motion of the cathode spot there. The apparent preference to form craters on the eu-cr interface may be due to the surface irregularity and/or contaminations.

Fig. 3.15. SEM picture of a virgin CuCr surface (tilt 70°}.

3.5.3. On eroded CuCr cathode

Fig. 3.16 shows a time resolved measurement of the cathode spot velocity and arc voltage on an arc eroded CuCr cathode, at arc current of 50 A. gap

length of 3 mm and magnetic induction of 0.06 T. It can be seen that the motion of the cathode spot is mainly based on a uniform motion in the order of 10 m/s and superimposed by an erratic motion within short time intervals. Comparing Fig. 3.16 with Fig. 3.10, the cathode spot motion is more uniform on an arc eroded CuCr cathode than on a virgin one.

The transient erratic motion of the cathode spot on a virgin CuCr (in the order of 100 m/s} is faster than that on an eroded CuCr (in the order of 40 m/s}. The arc voltage behaves similarly. The spikes in the arc voltage for a virgin CuCr contact have higher amplitude than for an arc eroded CuCr with a certain magnetic induction, gap length and arc current. The arc voltage and the cathode spot velocity are also correlated in this case.

Fig. 3.16. A simultaneous measurement of the cathode spot displacement (a), the cathode spot velocity (b) and the arc voltage (C) for a 50 A arc at B--Q.06 T and gap length of 3 mm .

38-tf

3.6. Two dimensional measurement gn cathode soot motign

3.6.1 Experimental arrangement

Figure 3.17 (a) shows the construction of the vacuum chamber. The anode is a stainless steel mesh with a diameter of 46 11111. The cathode is a OF'~~: copper disc with a diameter of 30 11111. The distance between the two

electrodes is fixed at 1 Dill. The are is ignited by sepa.ratilig the molybdenum trigger pin (grounded} with the anode and is then driven by an applied magnetic field B moving to the copper cathode. Two dimensional measurement on the cathode spot motion has been done by a framing camera through the mesh anode and the glass window. The interframe time used was 91 and 182 ~s respectively. The coordinate of the spot motion measurement is shown in Fig.3.17 (b).

(a J ( b l

Fig.3.17. The construction of the vacuum chamber (a) and the coordinate of the spot motion measurement with respect to the cathode surface (b).

The cathode surfaee had been cleaned by arcing before the measurements were taken.

3.6.2 Experimental results

A number of frames, depending on the arc lifetime, can be taken during one

*The measurements in this section were initiated by

Dr. J.

E. Daalder and

shot. Both x and y components of' the displacement of' the cathode spot between each two frames have been measured. These are in fact the step-widths in x and y directions (Ax and Ay) of the cathode spot motion in the interframe time (At

=

91 or 182 ~s).

It has been found that, for a given magnetic f'ield B and an arc current, both step-widths Ax and Ay obey the Normal distribution function:

1 [ 1 x-x:z F(x)

= -

exp[- - ( - - ) ] dx

a$ .-

2

a

(3-6)

where x i s the average value and

a

is the standard deviation. The average velocity of the spot motion is extracted by the average step-width divided ~ the interf'rame time:

v

_y

=

Ay/At {3-7}

the results of vy and vx as the functions of' an applied magnetic field are shown in Fig.3.18 (a) and {b).

(m"%1

lal 16 _~ ( b} 0 lm/SJ

\

12 0.4 8 0 0 -0.4 0 0 II 0 ₀ 0.1)4 0.08 0.12 0.14 0.2 BITI 0.1)4 0.08 0.12 0.16 0.2 BITI

Fig.3.18. Average cathode spot velocity as a function of · the applied magnetic field on an arc eroded copper cathode surface for a 50 A arc at gap length of 1 am. {a) y-component and (b} x-component. The standard deviation 2o' of y-component is shown in the error bar but 2o' of x-component is too large to be drawn in the scale.

It can be seen from Fig.3.19 that the value of vy' which represents the