US 3586809 A
Description (OCR text may contain errors)
United States Patent I 1 I 1 I 1 I 1 Inventor John D. Santl West Allis, Wis.
Appl. No. 818,913
Filed Apr. 24, I969 Patented June 22, 197i Assignee Briggs & Stratton Corporation Wauwatosa, Wis.
Continuation of application Ser. No. 714,490, Mar. 20, 1968, now abandoned.
REED SWITCH FOR RAPID CYCLE, HIGH POWER APPLICATIONS 25 Claims, 7 Drawing Figs.
Int. Cl. H0111 1/04 FleIdotSearch ZOO/166C,
166.l, 166 B; 335/154, 196; 337/] I I, 379
 References Cited UNITED STATES PATENTS 2,400,003 5/1946 Hensel et al. 200/166 (C) 3,3 I 7,869 5/1967 Funke 335/154 3,327,262 6/ I 967 Bongard et al.... 335/154 3,229,063 1/1966 Sakatos 335/154 Primary Examiner-H. 0. Jones Attorney-Ira Milton Jones PATENIED JUH22 l97l 3; 586; 809
sum 2 or 3 "NIROMET 42" "KOVAR" FIG-.7.
FULL GAP ZERO GAP (REEDS ENGAGED) 'NVELNTOR Juhn 17. 55min. BY
ATTORNEY REED SWITCH FOR RAPID CYCLE, HIGH POWER APPLICATIONS This application is a continuation of my copending application Ser. No. 714,490, filed Mar. 20, 1968, now abandoned.
This invention relates to magnetically actuated reed switches and more particularly to a reed switch that is especially well suited for use in a circuit that carries substantially high power and must be opened and closed at recurrent intervals of short but variable duration.
Magnetically actuated reed switches have been available for many years, and have long been recognized as potentially use- 1 ful in many applications. In practice, however, such switches have heretofore had limited utility because of their inability to operate satisfactorily in high volt-ampere circuits and to respond accurately to short, rapidly recurring cycles of buildup and decay of a magnetic actuating field. For example, one such switch, recently placed on the market and considered to represent an advanced state of the art, was rated for 240 operations per minute and to break 3 amps. at 125 volts, AC, and 2 amps. at 250 volts, AC.
To some extent this recent switch sacrificed high cycle speed to gain power handling capability, since previously marketed reed switches intended for lower power were capable of cycle speeds on the order of 60 operations per second.
Operating frequencies many times the 240 pm of this recent switch are needed for numerous applications in which reed switches are potentially useful (e.g., computers); and power ratings substantially higher than 500 volt-amperes at break are likewise essential in many cases (e.g., machine tool controls), and are desirable in all cases where they can be obtained without sacrificing other important characteristics.
For broadest utility, a reed switch should have a long service life as well as a high cycle speed and a high power rating. The above mentioned recently marketed reed switch had an advertised life expectancy of only 500,000 operations when breaking 3 amps. at 120 volts AC. Under conditions of rapid cycling this would represent an extremely short service life, e.g. less than 40 hours at 240 0pm.
The reed switch disclosed and claimed in US. Pat. No. 3,315,193, issued Apr. 18, 1967 to J. D. Santi, has been found to provide consistently accurate timing of make and break at frequencies of more than 200 operations per second, as well as at all lower cycle speeds. But while the invention of that patent achieved very high cycle speeds, it did not solve the long standing problem of providing a reed switch with both high cycle speed and the capacity for handling large amounts of power.
The present invention has as its general object to achieve a substantial advance over the state of the reed switch art represented by US. Pat. No. 3,315,193, and, specifically, to provide a reed switch that not only compares favorably in cycle speed with the switch of said patent, but in addition, has the ability to break circuits carrying on the order of 1,500 volt-amperes, and has a service life expectancy on the order of hundreds of millions of operations at its rated power.
In general, for a reed switch to afford reliable responses to changing magnetic fields, the resilient flexing bias of its reed or reeds, by which its contacts are separated for switch opening, should be a relatively light one; and in addition, for it to close and open in accurately timed relation to the buildup and decay of a rapidly cycling magnetic field, the normal gap between its contacts should be relatively small. A switch that meets these requirements tends to have relatively light contact pressure in its closed condition, and therefore, to avoid high contact resistance, prior reed switches usually had contacts of copper or precious metal operating in an inert gas. But such contacts tend to stick and erode when handling high power, and therefore the capability for controlling substantially high power has been regarded as incompatible with the requirement of a reed switch, and particularly one intended for rapid cycling.
Most of the causes of contact sticking'can be overcome with the use of tungsten contacts operating in a deep vacuum environment. Because of the high melting temperature of tungsten, tungsten contacts in deep vacuum can interrupt higher currents than contacts of any other known material without the occurrence of welding of the contacts or arcing between them; and because of the hardness of tungsten, frictional sticking of tungsten contacts is also unlikely.
However, when contact separating force is limited, as in a reed switch, tungsten contacts in a deep vacuum, when breaking a high voltage, tend to reclose under the influence of electrostatic attraction. To understand why such reclosure occurs, it must be borne in mind that almost every circuit is more or less inductive, and that tungsten contacts in a deep vacuum can separate without drawing an are through which energy in the circuit can be dissipated. Therefore, as the tungsten contacts initially separate, a high voltage surge is induced in the circuit and a large potential difference appears between the just-separated tungsten contacts. Whether or not the electrostatic charge between the contacts will reclose them depends upon the mechanical energy available for separating them. If the contacts can reclose, the high current surge due to the distributed capacitance in the circuit will usually be sufficient to melt the contacts and weld them together.
With the foregoing observations in mind, it is one of the general objects of this invention to provide an electromagnetically actuated switch that is responsive to relatively small actuating forces but is nevertheless capable of breaking circuits that impress very high potentials across its contacts, without switch failure due to contact sticking.
More specifically, it is an object of the present invention to provide a high cycle speed reed switch capable of controlling circuits that impress substantially high voltages across its contacts at circuit opening, and wherein means are provided for overcoming sticking of the switch contacts due to any of the causes of contact sticking, including reclosure due to electrostatic attraction between them.
A reed for a reed switch is a tuned element which vibrates at some natural frequency and which thus tends to have different responses at different rates of cyclical buildup and decay of a magnetic actuating field, depending upon whether or not the frequency of a building magnetic field is in phase with the vibration of the reed following an immediately prior switch opening.
Nevertheless it is another object of this invention to provide a reed switch that responds accurately to a cycling magnetic field of any frequency up to the maximum for which the switch is intended, said maximum being on the order of hundreds of cycles per second.
Another general object of this invention is to provide a very versatile reed switch that is adapted for very rapid cycling and is suitable for high voltage-high current applications, but which nevertheless has a relatively low resistance when closed so as to be well suited for low voltage applications where switch resistance must be kept to a minimum.
Another general and very important object of this invention is to provide a reed switch that has a very long service life expectancy even in high power DC applications.
The useful life of a switch depends upon the current that it is required to interrupt, particularly if it must always open when current is flowing in the same direction, as in a DC circuit.
It is well known that even when a switch opens without the occurrence of a spark or are between its contacts, some transfer of material between its contacts nevertheless takes place at each opening. This arcless transfer, which is known in the art as low voltage phenomenon," always takes place from the positive contact to the negative one. It occurs with tungsten contacts operating in air or in a deep vacuum, as well as with contacts of the other metals, operating in any environment. When a switch is used in DC service, arcless transfer of material always takes place in the same direction, and the transfers occurring in successive openings are cumulative so that the service life of the switch depends upon the volume of its contact material and the current to be interrupted. The same switch, operating in comparable AC service, will normally have a substantially longer useful'life because the switch does not always open at the peak current in the cycle and, further, because each contact is the positive one in about 50 percent of the openings of the switch so that material transferred during any one opening will be transferred back to its original site in a subsequent opening.
With the foregoing in mind, it is another object of this invention to provide a switch that has contacts which overcome to a very substantial extent the problem posed by the low voltage phenomenon just described, so that the switch of this invention has an extraordinarily long useful life even in high current DC applications.
More specifically, it is an object of this invention to provide a switch for relatively high power applications, having contacts which permit an incipient arcing to take place at each opening of the switch, and wherein the incipient arcing thus permitted offsets to a substantial extent the transfer of material from the positive to the negative contact that also occurs at each switch opening, to thus afford the switch of this invention a long useful life, even in DC service.
Still another specific object of this invention is to provide a reed switch having each of its reeds normally engaging a post or rebound stop under bias, so as to have a yielding preload in accordance with the teachings of the above mentioned US. Pat. No. 3,3l5,l93, and wherein the reeds can have such a preload of a predetermined value in the finished switch even though they are heated to a substantially high temperature during the course of manufacture of the switch.
It is also a specific object of this invention to provide a reed switch of the character described that can be operated under widely varying ambient temperature conditions without any change in the value of the biasing force which the reeds exert against their posts or rebound stops.
Another specific object of this invention is to provide a reed switch for rapid cycling having refractory metal contacts that operate in a deep vacuum environment and having reeds that can be outgassed before assembly of the switch is completed without losing their preload, so that the deep vacuum environment for the contacts will be maintained during a long service life of the switch.
With these observations and objects in mind, the manner in which the invention achieves its purpose will be appreciated from the following description and the accompanying drawings. This disclosure is intended merely to exemplify the invention. The invention is not limited to the particular structure disclosed, and changes can be made therein which lie within the scope of the appended claims without departing from the invention.
The drawings illustrate one complete example of the physical embodiment of the invention constructed according to the best mode so far devised for the practical application of the principles thereof, and in which:
FIG. 1 is a view in side elevation of a reed switch embodying the principles of this invention with a portion of its envelope shown broken away, the switch being illustrated in its normal open condition;
FIG. 2 is a view similar to FIG. I but showing the switch in its closed condition;
FIG. 3 is a disassembled perspective view of the components of the switch of this invention that are inside its envelope;
FIG. 4 is a view in front elevation of the reed assembly that includes the contactor;
FIG. 5 is a side view on an enlarged scale of a reed element of the switch of this invention;
FIG. 6 is a graph of the relationship between temperature and linear expansion of each of three metals used in the reed of the switch of this invention; and
FIG. 7 is a graph of the forces acting upon a reed in the switch of this invention.
Referring now to the accompanying drawings the numeral 5 designates generally a reed switch embodying the principles of this invention, comprising a pair of elongated, resiliently flexible, magnetically permeable reeds 6 and 6' enclosed in an elongated hermetically sealed envelope 7 of glass or the like. The envelope is evacuated to a deep vacuum. Extending through the sealed ends of the envelope are a pair of elongated terminal members 8 and 8, one for each of the reeds 6 and 6'. Each of the terminal members supports its reed at its inner end and has its outer end portion exposed to provide one of the terminals 9, 9' of the switch.
The reeds 6 and 6' extend axially inwardly from their respective terminal members 8 and 8, and their free tip portions are disposed in lengthwise overlapping relationship in the medial portion of the envelope, normally spaced apart laterally by a predetermined distance to define a gap 10. Under the influence of a magnetic field that threads the reeds and the gap 10, the reeds are flexed to bring their overlapping tip portions into contact with one another.
Also secured to the inner end of each terminal member 8,8 is an elongated, relatively stiff post or rebound stop which ex tends lengthwise along the reed at the side of the latter that is remote from the other reed, the post for the reed 6 being designated II and the post for the reed 6 being designated 11. Since each reed is formed from flat strip material, it can be said to have a rear surface that faces its post and a front surface that faces the other reed.
Each reed normally engages its post under flexing bias so as to have a rearward preload against the post in accordance with the teachings of the above mentioned US. Pat. No. 3,3 I 5,193, to which reference may be made for an explanation of the advantages of this arrangement and the manner in which it affords extremely high cycle speed in a reed switch.
In accordance with the present invention the reed 6 carries a contactor 12 which normally engages the rear surface of said reed 6 under forward flexing bias and which has a contact portion 13 that projects forwardly across the tip of the reed 6 to be normally spaced from the reed 6 by a distance somewhat smaller than that across the gap 10 between the tip portions of the reeds themselves. As explained more fully hereinafter, the contactor has several functions, one of which is to prevent failure of the switch due to sticking of its contacts under high current surges.
Turning now to a more detailed consideration of the switch structure, the envelope 7 is preferably made of a length of glass tubing that has its opposite end portions respectively drawn down around the terminal members 8 and 8' and fused to them to hold the terminal members 8 and 8' in rigidly fixed substantially coaxial relationship with one another and to provide hermetic seals around them.
As shown, the terminal members have their inner ends spaced apart axially by a substantial distance, and each has a post I I, ll welded to its inner end; but it will be apparent that each terminal member could comprise an integral axially outward extension of its post. The length of each post is such that its tip is near the tip of its reed. As explained in US. Pat. No. 3,315,193, the post can be of either magnetic or nonmagnetic material, but it should in any case be substantially stiffer than the reed.
In accordance with the present invention, each of the reeds 6 and 6' has a substantially flat anchor portion 15 adjacent its captive end, a substantially flat armature portion 16 that comprises its free end, and a medial neck portion 17 that has a smaller cross-sectional area than the anchor and armature portions and is bent at obtuse angles to them. The neck portion is of course integral with the anchor and armature portions and connects them. The anchor portion 15 of each reed flatwise overlies its adjacent post near the captive end thereof and can be secured to the post as by welding. The neck portion 17 extends obliquely forwardly, away from the post and toward the reed tip. The armature portion 16 is normally inclined toward the post to have its tip engaging the post under rearward flexing bias.
It is intended that most of the flexing of the reed of this invention shall take'place in its neck portion 17, and to this end the neck portion has reduced cross section area. Preferably the smaller cross section area of the neck portion is obtained by reducing only its thickness, as shown, to insure that the reed has a relatively flat spring rate, that is, to have the flexing stress forces in the reed increase relatively gradually with forward displacement of the reed tip toward the other reed. The significance of this will be evident from FIG. 7, which represents the forces acting on a reed in the switch of this invention as the reed is magnetically actuated towards its switch closing position. As a switch actuating magnetic field builds, it reaches a point at which the attraction between the tip portions of the reeds is sufficient to overcome their preloads and start them moving towards one another. From that point the convergence of the reeds is so much more rapid than the change in impressed flux that the impressed flux can be regarded as constant and the change in magnetic attraction force (which is represented by the curve 26 in FIG. 7) can be regarded as due entirely to the convergence of the reeds. The change in attraction force between the reed tips is approximately inversely proportional to the change in the flux leakage path area between them, and the curve 26 is therefore generally parabolic. The flexing stress in each reed, represented by curve 24, increases linearly with forward displacement of its tip. The difference between magnetic force and flexing stress at any instant is the force available to accelerate'the reed at that instant, and with a reed having a flat spring rate this difference grows rapidly as the reed moves toward the other, so that the reeds converge at an increasing acceleration.
Confining flexing of the reed mainly to its neck portion 17 further contributes to rapid response of the reed to a magnetic field in that only the armature portion 16 of the reed undergoes substantial movement. Since the mass of the armature portion is small as compared with that of the entire reed, the armature portion is readily accelerated by the magnetic and flexing forces that respectively act upon it in its switch closing and switch opening motions. To achieve a correspondingly low inertia in a reed of substantially uniform thickness along its length (like most prior switch reeds), the reed would have to be quite short and would therefore be so stiff that its spring force curve would cross the magnetic force curve and therefore the reed would not move to its closed position in response to a magnetic actuating fleld.
The neck portion 17 of each reed extends obliquely to its armature portion l6 in order to minimize contact bounce when the reed tips engage one another. The tendency of the armature portion to rebound is manifested in a vibration of the armature portion transversely to its flat faces. But because the neck portion extends obliquely to the armature portion, such vibration has a substantial component lengthwise of the neck portion. Furthermore, the armature portion is relatively short. For these reasons, the frequency of reed vibration due to any contact rebound tends to be high, and the amplitude of such vibration tends to be small. Such high frequency vibration dissipates energy rapidly, and because of its low amplitude there is little energy to dissipate. Therefore such make bounce as may occur in the reed switch of this invention poses no practical problem, even at the fast closure rates that are a concomitant of high cycle speed.
It will be understood, of course, that the provision of a post or rebound stop for each reed, against which the reed is preloaded, further contributes to the minimization of make bounce," permits the gap to be a small one, and makes possible the use of reeds having a low spring rate in a switch intended for successive operations at short intervals. Without the posts 11 and 11 each opening of the switch would initiate a high amplitude, low frequency vibration of the relatively supple reeds that would persist for a long time and would interfere with the timing of a directly subsequent reclosure of the switch. Instead, each reed rebounds from its post with a high frequency vibration of a very low amplitude, as explained in U.S. Pat. No. 3,3 I5,I93.
Since the reeds engage one another with substantial force upon closure of the switch, the front surfaces of their tip portions have tungsten coatings '18. Because of their hardness, these tungsten coatings resist mechanical wear and they prevent the reed tips from stickingtogether as a result of cold welding or mechanical friction; and because of the low current flowing between the contacts 18 as they engage and separate, due to the functioning of the contactor 12 as explained hereinafter, those contacts have no tendency to be welded together especially in view of the high melting temperature of tungsten. Because of the oxygen-free environment of the tungsten contacts 18 in the switch of this invention, they do not oxidize and therefore their contact resistance remains low during the life of the switch, instead of increasing like the resistance of contacts operating in air.
On the reed 6 there must also be a tungsten coating 23 that provides a contact engageable by the contact portion 13 of the auxiliary contactor 12, for reasons explained hereinafter. This contact 23 can of course comprise a continuation of the tungsten coated area that forms the contact 18 at the tip of that reed. (See FIGS. 3 and 4.)
As emphasized above, it is important that each of the reeds 6 and 6 be under a rearward preload of a predetermined value by reason of which it normally engages its post or rebound stop with a predetermined biasing force. However, the reed assemblies comprising the reed 6 or 6, the post I l or 11 and the terminal member 8 or 8' are necessarily subjected to substantial heating at some time during manufacture of the switch. A certain amount of heating of the reed assemblies is inevitable when the end portions of the envelope are melted and fused to the terminal members 8 and 8; and in fact heating of the reed assemblies is desirable in order to outgas them and insure that they will not emit gases that would destroy the deep vacuum in the envelope.
But if a monometallic reed were subjected to prolonged heating at or above its stress relieving temperature, it would of course lose its preload against its post.
A bimetallic reed, however, can be caused to have a desired preload after heating, and I have discovered two metals which possess properties that complement one another in a remarkable manner and cooperate to afford an extremely desirable bimetallic reed. One of these metals, sold under the trade names Therlo, Kovar and Rodar" (and hereinafter referred to as Kovar") consists of:
nickel 28.5 percent29.5 percent cobalt 16.5 percent-17.5 percent manganese 0.5 percent Max.
iron balance The other, sometimes sold under the trade name Niromet 42" (and hereinafter so referred to) consists of:
nickel 40.5 percent-4| .5 percent carbon 0.02 percent max.
silicon 0.25 percent max.
iron balance Both of these alloys have a high magnetic permeance and low retentivity so that they are well suited to be combined in a switch reed. They are also sufficiently resilient for the purpose. But the special suitability of this particular pair of alloys for the elements of a bimetallic reed has to do with their thermal expansion characteristics. FIG. 6 is a graph of the linear expansion with increasing temperature of these two alloys and of tungsten.
In the range from room temperature to about 600 F., the two alloys just described have very nearly equal coefficients of linear thermal expansion, so that a bimetallic reed that is made of them will not change its preload even under the most extreme changes of ambient temperature in which a reed switch could be expected to operate. Furthermore, within that temperature range the rate of expansion of these alloys is very nearly the same as that of tungsten, so that a bimetallic reed of these alloys can be very satisfactorily provided with the tungsten coating 18 on its tip portion without any danger that the tungsten coating will cause flexing of the armature portion of the reed under temperature swings by which the preload of the reed would be changed.
As may be seen from FIG. 6, the curves of expansion vs. temperature for these two alloys are substantially linear in the temperature range just mentioned. At about 600 F., however, the curve of expansion of Niromet 42" per degree rise in temperature turns upward and at values above about 700 F. it again becomes linear, but at a steeper inclination than below 600 F. The corresponding curve for Kovar continues upwardly in a substantially straight line to about 750 and then gradually steepens until from about 850 upwardly it again becomes linear and is parallel to the curve of Niromet 42 for temperatures about 700.
in accordance with this invention, a bimetallic reed is made by bonding together elements of Niromet 42 and Kovar and then rolling the bimetal billet to the desired thickness to form a strip from which the reeds are made. Each reed is secured to its post with the "Niromet 42" next to the post and with the reed initially under a substantially higher preload than is desired for it in the finished switch. This reed assembly is then heated to a temperature substantially in excess of 800 F., and preferably on the order of 1,400" F. or above.
The maximum temperature to which the reed assembly is heated is not critical, so long as the reed becomes hot enough to be incapable of supporting stresses. Nor is there any criticality in the time during which the reed is maintained at the maximum temperature. Preferably, however, such heating is performed in a deep vacuum and the reed assembly is brought to a high enough temperature and held there for a long enough time to outgas it.
As the temperature of the reed goes up through 600 to 800", the Niromet 42 at the rear of the reed expands more rapidly than the "Kovar," and consequently a portion of the initial preload of the reed against the post is relieved. With further increase in temperature to the maximum, the reed is completely stress relieved, remaining in contact with the post but without exerting any biasing force against it.
As the reed is cooled from its maximum temperature down to about 800", no stresses appear in it, since the two metals of which it is made are contracting at the same rate. At 800 and below, the reed is capable of supporting substantially high stresses. As it cools through the range of 800 to 600', the Niromet 42 contracts more rapidly than the Kovar and the reed is bimetallically flexed into engagement with its post under a biasing force that can be accurately predetermined on the basis of the relative thicknesses of the two alloys comprising the bimetal. Since the two metals expand and contract at substantially equal rates at temperatures of 600 and below, this biasing force or preload remains constant at all temperatures from about 600 down to room temperature.
It is emphasized that the reed must be initially assembled to its post with a substantially excessive preload, so that the tip of the reed will not move forwardly out of contact with its post under the bimetallic flexing that occurs as the reed is heated through the 600-800 range. if the reed tip were permitted to move away from the post during heating, due to an insufficient initial preload, the final preload on the reed would be indeterminate because it would depend in part upon the unknown spacing between the reed tip and the post at temperatures above about 800. lf high enough, the value of the initial preload is not critical, since the final preload on the reed is purely a function of its bimetallic flexing.
It will be understood that bimetal reeds according to this invention can be made of other pairs of metals than "Kovar" and Niromet 42." In every case, at least one of the metals comprising the reed must be magnetically permeable; the rates of linear thermal expansion of the two metals must be substantially similar up to a certain value, which value is within the range of temperatures at which both metals will support stress; and the rates of thermal expansion of the two metals must be different through a range of temperatures above that value, the metal having the higher rate of linear expansion in said range being at the side of the reed adjacent the post. For example, a fairly satisfactory reed according to the invention has been made of Niromet 42 having a tungsten coating along its entire length at its side remote from the post,
the tungsten coating serving as the bimetal element having the lower rate of linear thermal expansion in the high temperature range.
While the high cycle speed of the switch of this invention is in large measure due to the preload of its reeds 6 and 6' against the posts 11 and 11, its ability to control high volt-ampere circuits is mainly attributable to the contactor 12 that is carried by the reed 6, although as will appear in the course of the description, the contactor also performs certain other important functions.
The contactor should be light in weight so that it does not greatly increase the mass of the reed 6 by which it is carried, and it must be substantially resilient. These requirements are nicely met by molybdenum wire, which also possesses other very important advantages that are discussed hereinafter.
The molybdenum wire contactor 12 is bent generally to an elongated U that has its legs 19 curved inwardly back upon themselves near their free end portions, around loops 20, to provide coplanar anchoring elements 21 that comprise the end portions of the length of wire. These anchoring elements flatwise overlie the front face of the armature portion 16 of the reed 6 and they are welded or otherwise bonded thereto. By reason of joglike bends 22 in the legs 19 of the contactor, the anchoring elements 21 and loops 20 lie in a plane that is spaced forwardly from the plane of the remainder of the legs 19, and the portions of said legs that lie between the bends 22 and the bight portion 27 of the contactor define a spring arm which overlies the rear face of the armature portion 16 of the reed and which normally engages the armature portion under forward bias. At its bight portion 27 the legs 19 of the contactor are spaced apart a distance substantially less than the width of the reed, but they diverge substantially toward the neck portion 17 of the reed and are spaced laterally from the side edges thereof.
The bight portion of the contactor, which lies outwardly of the tip of the reed 6 and which comprises the contact 13 thereof, is bent at right angles to its spring arm portion to pro ject forwardly across the tip of said reed and beyond the front surface thereof so that with the switch in its normally open position the distance between the contact 13 and the reed 6' is less than that between the reeds themselves.
As the reeds converge under the influence of a magnetic field, the contact 13 on the contactor 12 engages the contact 23 on the reed 6' before the contacts 18 on the reeds themselves come into engagement. Because of the forward preload of the contactor against the reed 6, the force required to bring the reeds into contact with one another increases markedly when the contactor engages the reed 6'. In other words, once the contactor 12 engages the reed 6, the magnetic attraction force acting on each reed is yieldingly resisted by the flexing force of the reed itself combined with that of the contactor 12. in H0. 7 the flexing force in the reed itself is represented by the curve 24, while the additional flexing force contributed by the contactor is depicted by the curve 25. Note how the stepped curve for the combined flexing forces of the reed and contactor nicely matches the curve 26 for the magnetic force acting to bring the reeds into engagement.
As pointed out hereinabove, when the reeds move toward one another under the influence of a magnetic field, they are accelerated rapidly, right from the beginning of their converging motion. Hence the reeds have substantial forward momentum at the instant the contactor engages the reed 6'. F urthermore, the magnetic force acting upon them at that instant is substantially large in relation to the flexing forces that they oppose to it. Since the contactor has forward preload, these inertia and magnetic forces combine to very nearly eliminate contact bounce between the contactor and the reed 6'. Such slight vibration as may occur immediately following impact of the contactor against the reed 6 will be of very high frequency and small amplitude, due to the rapidly building force with which the contactor is urged against that reed, and such vibration will therefore be very quickly dissipated.
After the contactor engages the reed 6, the armature portions of the two reeds continue to converge rapidly, but the net spring force resisting their convergence does not increase quite as rapidly as the magnetic force attracting them into engagement, as will be apparent from FIG. 7. For this reason, as well as for the reasons pointed out hereinabove, the armature portions of the reeds will have little tendency toward "make bounce" when they engage; but of course if they should rebound, the switch will nevertheless remain effectively closed because the contactor will remain engaged against the reed 6 with a contact force substantially equal to the preload under which it normally engages the reed 6.
It will be noted that most of the flexing of the contactor takes place in its loop portions 20, which are located closely adjacent to the neck portion 17 of the reed. Hence the contact 13 of the contactor l2 and the armature portion 16 of the reed 6 tend to swing about a common center, with the result that the contact 13, once engaged with its cooperating contact 23, has substantially no wear producing movement across the surface thereof even though the armature portion of the reed 6 continues to swing.
When the switch of this invention is fully closed, the contacts on the two reeds, comprising the tungsten coatings 18, provide a second current path through the switch, cooperating with the contacts 13 and 23 on the contactor and the reed 6, respectively, to provide the switch of this invention with a relatively low net resistance whereby it is well suited for low voltage applications as well as for high power circuits.
While the contactor l2 performs very important functions in preventing make bounce" and in maintaining low net resistance through the closed switch, its role during switch opening is perhaps even more important.
As pointed out above, the reeds in the switch of this invention have a very flat and relatively low spring rate so that they respond promptly and consistently to a rapidly building magnetic field. But sticking is always a problem with switch contacts that must open a high power circuit, even with high melting contacts operating in a deep vacuum; hence reeds of low stiffness could not, by themselves, exert sufficient contact separating force to interrupt a circuit carrying high power. In the switch of the present invention, contact sticking from all causes is overcome by means of the contactor 12.
When a magnetic actuating field for the switch of this invention decays, separation of the reed tips somewhat increases the net resistance of the switch, so that there is some potential difference between the contacts 18. But the switch still remains effectively closed, owing to the engagement of the contactor with the reed 6'; and therefore the electrostatic attraction between the contacts 18 is negligible and does not interfere with their rapid separation.
As the reed 6 moves rearwardly towards its post 11, it acquires a high velocity, being accelerated by a combination of its own flexing force and that due to the contactor preload. By the time it reengages the arm of the contactor it has a large momentum. In consequence, the reed 6 literally knocks the contactor out of engagement with the reed 6', overcoming any tendency for the contacts 13 and 23 to stick to one another. In thus providing for abrupt, forceful contact separation at switch opening, the contactor enables the reed switch of this invention to interrupt as high as 3,000 volts at 0.5 amp., with surge voltages at break that are as high as about 15,000 volts.
The current that a switch is required to break normally determines its service life expectancy. The switch of this invention has been found to have a substantially longer useful life than prior reed switches even when breaking currents of as high as amperes. At 0.5 amperes DC a switch of this invention has been found to be still useable after more than 200 mil lion operations; and with lower currents its useful life would increase in a nearly linear inverse relationship to current.
This long service life is due to the particular metals of which the contacts 13 and 23 are made. As noted above, the contact 13 is an integral part of the contactor 12, which is made of a single piece of molybdenum wire, and it cooperates with a tungsten contact 23 on the reed 6'.
If the contacts 13 and 23 were both of tungsten, operating in a deep vacuum, they would open without sparking or arcing, but a certain amount of tungsten would transfer from the positive contact to the negative one at each switch opening. This transfer of material to the negative contact is the well known low voltage phenomenon that occurs in arcless, sparkless separation of any switch contacts through which current is flowing. The reason for it is fairly well established. When a pair of contacts begins to make an arcless separation, their resistance increases rapidly as the contact force diminishes. Just as contact pressure goes to zero, the contacts are engaging each other at an infinitesimal point through which all of the current in the circuit must flow, and the energy thus charged into the contact metal in this very localized area heats it above its melting point, however high this may be. The molten particle of metal thus formed is for a brief instant held in place by surface tension, bridging the separating contacts, but as the contacts continue to move apart its surface tension collapses and it settles back onto the contacts. However, most of the molten metal settles onto the negative contact, because the positive contact becomes hotter than the negative one. While the amount of metal thus transferred at each opening is relatively small, it varies in more or less direct proportion to the amount of current carried by the switch.
When the switch of this invention is connected in a DC circuit, or in any circuit in which current is always expected to flow in the same direction at the instant of switch opening, it should be so connected that its terminal 9 will be the positive one at the instant of switch opening and its terminal 9 will be the negative one. This is to say that the molybdenum contact 13 will be positive and the tungsten contact 23 will be negative.
Hence when the contacts begin to separate, molybdenum from the contact 13 will be transferred to the tungsten contact 23 in accordance with the usual phenomenon of low voltage material transfer. As the contacts continue to move apart, the bridge of molten molybdenum across the two contacts reaches its collapsing point, and a voltage difference appears between the contacts. Under the influence of this voltage difference an abundance of electrons is emitted by the molten molybdenum on the negative contact, producing a more intensified local heating there which causes ionized molybdenum to be emitted into the space between the contacts.
Apparently a full plasma arc does not develop between the contacts, but there appears to be an incipient arc in the nature of a cathode spot on that part of the contact 23 to which molybdenum has been transferred. From that cathode spot most of the molybdenum ions are emitted, and such ions tend to move toward the anode, i.e., the contact 13 from which they originated and from which they were transferred as molten molybdenum. Because of the deep vacuum environment of the contacts, they do not oxidize, and therefore the vaporized molybdenum which is redeposited on the molybdenum contact 13 readily adheres thereto.
Thus the arcless transfer of molten molybdenum from the contact 13 to the contact 23 that occurs in the initial phase of contact separation is in large measure offset or compensated for by the retransfer of molybdenum back to its original site, occuring by reason of the ionization that takes place in the immediately succeeding phase of contact separation.
The tungsten does not take part in the transfer and retransfer described above, because its melting and vaporizing temperatures are substantially higher than those of molybdenum and because in the very first phase of contact separation, when arcless transfer is occurring, the tungsten contact is the one to which such material transfer always takes place, so that molten molybdenum is being deposited onto it. At the power levels for which the switch of this invention is intended the temperature of the cathode spot is determined by the characteristics of molybdenum; or in other words there is not enough power to support ionization of the tungsten once the molybdenum is vaporized off of it.
The foregoing explanation for the long service life of the switch of this invention in DC applications may not be complete, and may be inaccurate in some respects, but it is consistent with the results that have been obtained with the switch of this invention and also with known phenomena of switch contact behavior and the generally accepted theories for them.
The explanation tends to be confirmed by what happens when the switch of this invention is wrongly connected in a DC circuit, with its terminal 9 negative and its terminal 9' positive. Under those'conditions the negative molybdenum contact 13 is sufficiently cooler than the positive one so that an accumulation of tungsten develops on it, and the switch has a substantially shorter useful life than when properly connected, failing when the tungsten contact becomes pitted or eroded to the point where the substrate metal is exposed and the contacts stick because of welding or mechanical friction. Retransfer of the tungsten cannot occur under these conditions because of the extremely high vaporizing temperature of tungsten, which prevents the occurrence of an effective cathode spot at power levels within the capabilities of the switch.
Further confirmation of the correctness of the above explanation appears from the fact that the switch of this invention actually has a somewhat shorter life in a low voltage, low inductance circuit, even when correctly connected, than in the higher power circuits for which it is intended. Thus when a switch of this invention was installed in a substantially noninductive circuit carrying 2 amperes at 1?? volts, molybdenum was found to transfer rather steadily from the positive contact 13 to the negative tungsten contact 23, with no evidence of retransfer. It should be pointed out that a circuit of this type (which would usually be in the nature of a heater circuit) is one in which a reed switch would not be likely to have particular value, and that even under these comparatively adverse conditions the useful life of the switch of the present invention was several times that of the best prior reed switches.
It will be seen from the foregoing explanation that the long useful life of the switch of this invention is due to the spread or difference between the melting and vaporizing temperatures of the two metals comprising the respective contacts 13 and 23 and the deep vacuum environment in which they operate. Contacts of other materials could be used in a switch embodying the principles of this invention provided that the material of the negative contact was one that did not produce an effective cathode spot within the power levels for which the switch was intended, and that the material of the positive contact did so; and provided, of course, that such contacts were in a deep vacuum. However, tungsten and molybdenum are the preferred metals because both of them melt at high enough temperatures to obviate the possibility of the contacts being welded shut, and both are hard enough to resist mechanical wear that might in time produce frictional sticking.
While the tungsten/molybdenum contact pair of the switch of this invention undoubtedly has a higher resistance than contacts of the softer metals, other factors pointed out hereinabove compensate for this to provide the switch of this invention with a relatively low net resistance.
Of course the switch of this invention will not last forever. A certain amount of the molybdenum that is vaporized at each switch opening is dissipated to the wall of the envelope 7, and not all of the molybdenum that is transferred to the tungsten contact 23 will be reengaged by the contactor 12. Obviously the molybdenum that is deposited in locations where it is never again contacted by the molybdenum contact cannot take part in cathode spot vaporization and therefore cannot be transferred back to the molybednum contact member. Hence after some hundreds of millions of operations in a high power DC circuit, the molybdenum contactor will fail because of such dissipation of the material of its contact portion. But with these considerations in mind, it is possible to prolong the life of the switch by providing the contactor with a contact surface of substantial area, to afford the greatest possible opportunity for transfer back to it of molybdenum deposited on the tungsten contact 23. To this end the wire ofthe bight portion of the contactor is preferably flattened as shown in FIG. 6.
Besides affording long service life as explained above, molybdenum wire makes a good contactor because it is duetile enough to be readily formed to the desired shape and has sufficient resiliency to provide the desired biasing force. A further advantage of molybdenum wireis that it continues to support bending stresses at temperatures as high as l,400 F. and above. This means that the contactor 12 can be bonded to the reed 6 before that reed is heated to outgas it and fix its preload against its post ll. Of course the contactor should be fastened to the reed 6 with a preload substantially higher than that which it is intended to have in the finished switch, since a certain amount of stress relieving of the contactor will occur as it is heated during outgassing of the reed assembly. However, if contactors are always assembled under an initial preload which is equal to the yield point of the material, and the reed assemblies are always heated to a particular maximum degassing temperature such as l,500 F., the preload of the contactor against the reed 6 will always have a value, in the completed switch, that is directly related to the temperature to which it was heated.
From the foregoing description taken with the accompanying drawings it will be apparent that this invention provides a magnetically operated reed switch that is capable of closing and opening in very fast and faithful response to the rapid buildup and decay of a magnetic actuating field; and that the switch of this invention is capable of controlling substantially high volt-ampere circuits in rapid cycling, with an extremely long service life even in DC applications.
What I claim as my invention is:
l. A switch having a pair of contact carriers arranged for relative movement to and from engagement with one another under the influence of a magnetic field and biased to normal relative positions in which they are spaced a predetermined distance apart, said switch being characterized by:
A. a first pair of contacts, both of high melting temperature material, one on each of said contact carriers, said contacts comprising the portions of the contact carriers that are engageable with one another;
B. acontactor 1. carried by one of said contact carriers for movement relative thereto toward and from the other contact carrier,
2. said contactor having a biasing preload against said one contact carrier in the direction toward said other contact carrier;
C. a third contact of high melting temperature material on the contactor, normally spaced less than said predeter mined distance from the other contact carrier;
D. a fourth contact of high melting temperature material on said other contact carrier, engageable by said third contact upon relative movement of the contact carriers towards engagement with one another; and
E. a hermetically sealed envelope, evacuated to a deep vacuum, enclosing said contact carriers, the contactor and said contacts.
2. The switch of claim 1, further characterized by:
A. the contacts of said first pair thereof being of tungsten;
B. said third contact being of molybdenum; and
C. said fourth contact being of tungsten.
3. A switch comprising a pair of contacts at least one of which is movable toward and from engagement with the other, for repeatedly opening and closing a circuit in which an electric current always flows in the same direction at the time of separation of said contacts so that at such times a first one of said contacts is always connected with the negative terminal of the source of current for the circuit and the other with its positive terminal, said switch being characterized by:
A. means defining an evacuated hermetically sealed enclosure enclosing the contacts by which they are maintained in a deep vacuum environment;
8. said first contact being of material having melting and vaporizing temperatures at least about as high as those of tungsten; and
C. said other contact being of a material having lower melting and vaporizing temperatures than the material of the first. 4. The switch of claim 3, further characterized by: said other contact being of molybdenum. 5. A reed assembly for a reed switch, comprising an elongated, resiliently flexible reed having a free end and an anchored end, and means defining a stop against which the free end portion of the reed is normally engaged under bias, said reed assembly being characterized by:
the reed being a bimetal comprising a pair of elements which are bonded together along the length of the reed, 1. one ofsaid elements being magnetically permeable, 2. the rates of linear thermal expansion of said elements being a. substantially similar at temperatures below a value which is within the range of temperatures at which both elements will sustain stress, and
b. substantially different through a range of temperatures above said value, and
3. the element having the higher rate of linear thermal expansion in said range being at the side of the reed adjacent to said stop.
6. The reed assembly of claim 5, further characterized by:
A. said element at the side of the reed adjacent to the stop being ofNiromet 42"; and
B. the other element being ofKovar."
7. The reed assembly of claim wherein said stop comprises an elongated post, substantially stiffer than the reed, arranged in side-by-side relation with the reed, further characterized by said reed having:
A. a substantially flat anchor portion adjacent to its anchored end that flatwise overlies the post;
B. a substantially flat armature portion adjacent to its said other end; and
C. an integral neck portion between said anchor and armature portions and connecting them, said neck portion having a cross section area less than that of the anchor and armature portions.
8. The reed assembly of claim 7, further characterized by:
A. the neck portion of the reed being of lesser thickness than the anchor and armature portions;
B. said neck portion being oblique to the anchor and armature portions of the reed and being inclined away from the post and toward the armature portion; and
C. the armature portion of the reed being oblique to the post with its end portion adjacent to the neck portion spaced from the post and its opposite end portion preloaded against the post.
9. The reed assembly of claim 8, further characterized by:
a contactor secured to the armature portion of the reed and having a spring arm overlying and preloaded against the surface of the armature portion that is adjacent to the post, said spring arm extending lengthwise beyond the tip of the armature portion and having a contact portion that projects away from the post across the tip of the armature portion to normally extend beyond the plane of the surface of the armature portion that is remote from the post.
10. The reed assembly of claim 5, further characterized by:
a contactor secured to the reed and having l. a spring arm portion overlying said free end portion of the reed and preloaded against the same in the direction away from the stop, said spring arm extending lengthwise beyond the free end of the reed, and
2. a contact portion projecting laterally from the spring arm in the direction away from the stop and across said other end portion of the reed a distance to normally extend beyond the plane of the surface of the reed that is remote from the stop.
11. The reed assembly of claim 10, wherein said contactor is of molybdenum wire, further characterized by:
a coating of tungsten on said other end portion of the reed at the side thereof remote from the stop.
12. A switch of the type comprising an envelope enclosing a pair of magnetically permeable contact carriers, one of which comprises an elongated, resiliently flexible reed that is normally spaced a predetermined distance from the other contact carrier to define a gap but is flexible into engagement with the other contact carrier under the influence of a magnetic field threading the contact carriers and the gap, said switch being characterized by:
A. the envelope being I. hermetically sealed and 2. evacuated to a deep vacuum;
B. a contactor carried by one of the contact carriers,
l. said contactor being preloaded against the contact carrier by which it is carried, in the direction toward the other contact carrier, and
2. said contactor having a portion which normally projects partway across the gap toward said other contact carrier;
C. a molybdenum contact on said portion of the contactor;
and a D. a tungsten contact on said other contact carrier in a position to be engaged by said molybdenum contact.
13. The switch of claim 12, further characterized by:
a tungsten coating on that portion of each contact carrier that engages the other, to provide an additional pair of cooperating contacts.
14. in a magnetically actuated switch:
A. an elongated, resiliently flexible, magnetically permeable reed having an anchored end and a free end; and
B. a contactor carried by the reed and comprising a single length of resiliently flexible wire, the medial portion of which is bent to a U-shape, with legs and a bight portion, and each end portion of which is connected with one of said legs around a loop and extends back along its leg towards the bight portion,
. said end portion being flatwise secured to the reed near the free end thereof,
2. portions of the legs overlying one face of the reed and normally engaging the same under flexing bias; and
3. the bight portion being bent out of the plane of said leg portions and projecting across the free end of the reed beyond the plane of the other face thereof to provide a contact portion of the contactor.
15. The magnetically actuated switch of claim 14, further characterized by:
A. said reed having 1. a substantially flat anchor portion adjacent its anchored end,
2. a substantially flat armature portion adjacent its free end, and
3. an integral neck portion of smaller cross-sectional area connecting its anchor and armature portions and obliquely inclined to them;
B. the contactor having its said end portions overlying and secured to one face of the armature portion; and
C. said loops being edgewise spaced to opposite sides of the neck portion.
16. The magnetically actuated switch of claim 15, further characterized by:
an elongated post, substantially stiffer than the reed, ex.- tending alongside the reed adjacent to the first mentioned face thereof, substantially to the free end of the reed,
l. the anchor portion of the reed being flatwise engaged with and secured to the post,
2. the neck portion of the reed extending obliquely away from the post and the anchor portion, and
3. the armature portion of the reed being preloaded toward the post.
17. The magnetically actuated switch of claim 16, further characterized by:
said reed being a bimetal comprising l.Niromet 42" adjacent to the post, and
18. The magnetically actuated switch of claim 14, further characterized by:
A. the contactor being of molybdenum wire; B. a tungsten contact engageable by the contactor upon flexing of the reed in the direction of its said other face; and
C. a hermetically sealed envelope, evacuated to a deep vacuum, enclosing the reed, the contactor and said tungsten contact.
19. The magnetically actuated switch of claim 16, further characterized by:
the neck portion of the reed having a width not substantially less than that of the anchor and armature portions but having a lesser thickness than the anchor and armature portions.
20. A magnetically actuated switch of the type comprising a pair of elongated, resiliently flexible, magnetically permeable reeds, each having an anchored end and a free end, and means securing the anchored ends of the reeds at locations remote from one another and disposing the reeds with their free end portion in overlapping but laterally spaced apart relationship to define a gap between them, said switch being characterized by:
A. a post for each reed, each of said posts being substantially stiffer than its reed and extending lengthwise along its reed at the side thereof remote from the other reed;
B. the free end portion of each reed being normally engaged under bias with its post;
C. a contactor secured to one of the reeds and having 1. a spring arm portion overlying and preloaded against the surface of said reed that is adjacent to its post, said spring arm extending lengthwise beyond the free end of said reed, and
2. a contact portion on said spring arm portion that projects across the free end of said one reed partway across the gap to be engageable with the other reed.
21. -An elongated reed for a magnetically actuated reed switch, said reed having anchored and free ends and being characterized by:
A. a reduced thickness neck portion intermediate the ends of the reed, defining a substantially flat anchor portion adjacent its anchored end and a substantially flat armature portion adjacent its free end;
B. said neck portion of the reed being at a substantial oblique angle to the anchor portion so as to be inclined lengthwise in one direction thereto; and
C. said armature portion being at an opposite oblique angle to the neck portion so as to be inclined lengthwise in the opposite direction to the neck and anchor portions.
22. The reed of claim 21, further characterized by:
an elongated post, substantially stiffer than the reed,
secured to the anchor portion and extending lengthwise along the reed substantially to its tip and substantially parallel to the anchor portion, said neck portion of the reed extending obliquely away from said post and said armature portion of the reed being preloaded toward the post.
23. The reed of claim 22, wherein said anchor, neck and armature portions of the reed are integral with one another, further characterized by:
said reed being a bimetal comprising 1. Niromet 42 adjacent to the post, and 2. "Kovar."
24. A magnetically actuated switch of the type comprising an elongated, resiliently flexible, magnetically permeable reed having an anchored end and a free end, a contact engageable by the free end of the reed, and means securing the anchored end of the reed at a location spaced from the contact and disposing the reed with its free end swingable toward and from engagement with the contact and normally spaced therefrom to define a gap, said switch being characterized by:
A. a post for the reed, substantially stiffer than the reed and extending lengthwise along the reed at the side thereof remote from the contact; B. the free end portion of the reed being normally engaged under bias with the post; and
C. a contactor secured to the reed and having I. a spring arm portion overlying and preloaded against the surface of the reed that is adjacent to the post, said spring arm portion extending lengthwise beyond the free end of the reed, and
2. a contact portion on said spring arm portion that projects across the free end of the reed partway across the gap to be engageable with the contact.
25. A switching comprising a pair of contacts at least one of which is movable toward and from engagement with the other, for repeatedly opening and closing a circuit in which an electric current always flows in the same direction at the time of separation, said switch being characterized by:
A. means defining an evacuated hermetically sealed enclosure enclosing the contacts by which they are maintained in a deep vacuum environment;
B. one of said contacts being of material having melting and vaporizing temperatures at least about as high as those of tungsten;
C. electrically negative terminal means connected with said one contact;
D. the other of said contacts being of a material having lower melting and vaporizing temperatures than the material of said one contact; and
E. electrically positive terminal means connected with said other contact.