US2222485A - Electric discharge lamp and starting device therefor - Google Patents

Description

Patented Nov. 19, 1940 PATENT OFFICE ELECTRIC DISCHARGE LAMP AND STARTING DEVICE THEREFOR Henry J. McCarthy, Danvers, Mass, assignor to Hygrade Sylvania Corporation, Salem, Mass. a corporation of Massachusetts Application May 6, 1939, Serial No. 272,246

1 Claim.

This invention relates to electric gaseous discharge lamps, and in particular to such lamps having thermionic cathodes, and to apparatus for starting and operating such lamps.

An object of the invention is to provide an electric gaseous discharge lamp with thermionic cathodes and with apparatus for allowing the cathodes to preheat for a definite interval before the discharge is started between them. A further object is to provide a switch with the proper time delay to allow preheating oi the cathodes.

A feature of the invention is a switch actuated by the stretching of a heated wire.

Other objects and advantages of the invention will be apparent from the following description, taken in connection with the accompanying drawing, in which:

Figure l is a schematic View of the circuit and device of the invention.

Figure 2 is a view of the contacts of the device in a direction at right angles to Figure 1.

Figure 3 is a view of a modified form of the switching device of Figure l.

in Figure l, a tungsten wire i is stretched be tween the nickel supports 3 and i. The support wire is short and relatively fixed, while the wire t is longer and hence more flexible. The glass tube 8 is placed around the wire l. A wire Ci, which may also be of nickel and is preferably of a smaller diameter than the wires 3 and t rests against the glass tube 8, preferably with its end bent as shown in Figure 2 to rest on the glass tube at right angles to it, in order that the place where it rests can be easily predetermined. Another wire I rests against the bent portion of wire 6. The wires 3, d, 6 and l are mounted in a glass stem 5, being sealed through said stein by the lead-in wires 9, H], H and 12 which are of a material capable of being sealed through glass. The stem is sealed into the glass bulb 2, and the whole is sealed by the exhaust tube l3. The bulb may be evacuated, or it may be filled with an inert gas at a pressure sufficient to reduce arcing at the contacts 8 and I.

An electric discharge lamp M has the tubular glass envelope IS with the electrodes "5, comprising a coiled-coil of tungsten wire coated with one or more of the alkaline earth oxides, sealed through each end of the tube by the lead-in wires l1, Hi. The discharge tube l4 preferably contains an inert gas and mercury vapor, and may have a coating of fluorescent material on the interior surface of its envelope. A wire I9 may be connected to the lead-in wire I! and extend along,

but out of contact with, the filament l8, toward the other lead-in wire it. The lead-in wires it are connected to a suitable power supply line by the conductors 253, one of which has the reactarice coil it in its circuit to ballast the lamp during g operation and to provide an inductive surge at starting. The same conduct-or has the tungsten wire i of the switching device connected in series with it, although this wire may be connected in series with the conductor to the w other side of the line, if desired. The lead-in wires ii are connected together through the normally closed contacts l of the switching devices.

in operation, when the ends of the conductors N are connected to a source of voltage, the contacts 3, being closed, a current will flow through the reactance coil 2i and the electrodes l6, heating the latter to an electron-emitting temperature and allowing a small current to flow so from wire it to the end of filament it nearest the lead-in wire it as described, for example, in my copending application Serial No. 269323, filed April 22, 1939. To permit such a discharge, the voltage across the filament must be above the excitation voltage of the gas in the tube, but it should be only slightly above it, to insure that the discharge will not be sufiicient to damage the filament l6. JV'hile the filamentary electrode iii of the discharge tube is heating up, the tungsten filament l of the switching device also is heating up, and is adjusted so that when the electrode i5 reaches the proper-electron emitting temperature and is surrounded by the ions produced the small discharge to the wire IS, the tungsten "wire i has just expanded sufficiently to allow the wire t to spring away slightly from the contacts l allowing them to open, thus producing a voltage surge by discharging the inductance across the discharge tube i i, and starting the lamp.

The contacts 6 and l are positioned so that they would be opened slightly if the wire l did not push the glass tube 8 against them. The wire l is then made tight enough to just pull the support l, with its surrounding glass tube 5, over against the contacts enough to hold them closed when no current flows through the wire I. When current does flow through the wire i, the latter expands in length, allowing the wire t to pull 50 away from the glass tube 4, thus opening the contacts 6, l. The support 4 should be sufliciently springy and under the proper tension, to spring back as the wire l stretches, and the contacts t, l, or at least the contact 1, should be sumciently springy to open as the wire 4 springs away.

I have made a device such as shown in Figures 1 and 2, for operation on a current of one-quarter ampere, in which the rods 3 and 4 were of about 3*; inch diameter nickel and the contacts 3 and I were of about ,4 inch diameter nickel. The glass tube 8 had an inside diameter Just large enough to slip over the wire 4, and an outside diameter of about inch. The wire I was of the non-sagging, non-offsetting tungsten used in incandescent lamps, had a diameter of 0.01 inch, and a length of 1% inches and was welded at its ends to the supports 3 and 4. The contacts 3, I opened about one second after current began to flow through the wire I; if the lamp bulb was evacuated. I sealed the bulb to a four prong radio tube base of a standard tube, but the device, of course, will work without any base.

Figure 2 shows a side view of the contacts 8, I, at right angles to the view shown in Figure 1. Wire 1 is seen to be straight while wire 6 crosses it and the glass tube 5 at an angle. This bending is done to give a sharp point of contact, and a definite place when the wire meets the glass, but other arrangements may be used if desired. Silver or tungsten contact points may be placed on the wires 6, I for example, and the wires themselves may be pieces of fiat spring metal. The wires 6, I may, of course, be themselves of tungsten to give better contact, if desired.

Figure 3 shows a modified form of the switching device 23. In the embodiment of the invention, the device is mounted in a glass bulb 2 as before. Support wire 3 is short and relatively fixed as before, and support wire 4 is longer. Now, however, a flat thin spring iron contact piece 23 is mounted on thesupport 4 and is held against the contact rod 1 by the pull of the tungsten wire I, which is connected to the nickel tab 24 welded to a second spring metal piece 32. A thin strip of mica 33 insulates the spring piece 32 from the spring piece 23 so that there will be no electrical contact between the contact circuit I-23-34 and the actuating circuit 3I24-324. The tab 24 passes through an insulating hole in the piece 23. The spring piece 23 is adjusted so that it would tend to spring away from the contact piece I if the tungsten wire I did not pull the piece 32 down to close the contacts. The tungsten wire I may be welded to the tab 24 and the support 3. This enables the tungsten wire to be kept straight. The mica insulating piece 25 holds the end of contact rod I steady with respect to support 4. Metal eyelets may be placed in the mica and welded to the supports to hold the mica in position. A hole is provided in the mica to allow free movement of the piece 24. The counterweight 26 may be attached to the spring piece 23 to change the natural frequency of the spring, if desired. A four prong base 21, such as used on radio tubes may be afilxed to the bulb. The lead-in wires or supports 4, 34, 3, I may be aflixed, respectively, to the contact pins 23, 29, 30, 3I.

My device may be used in open air without the enclosing envelope 2, and other wires I than tungsten may be used. However, enclosing the bulb in an evacuated envelope, or in onefilled with an inert gas, and using a filament I of tungsten, and arranging the parts as shown, makes a longlived device, obviates the need for special contact materials, and eliminates any need for continually adjusting the device to compensate for ambient conditions, wire fatigue and the like. I have operated my device for exceedingly long periods, over one thousand hours, with no appre ciable change in its operating characteristics, and have run a device 22, with the wire I connected in series with the contacts I5, I to provide automatic on and oi! operation as rapid as 100 times a second, for 1000 hours. Such operation means that the device has operated as a switch 36,000,000 times successfully and without adjustment, whereas most lighting switches are ordinarily tested for no'more than 10,000 operations. As a matter of fact, most lamps would not be turned on and off more than three times a day, in which case my device could last 36,000 years. Its permanence is thus apparent, and this permanence is a characteristic of my invention.

Another remarkable feature of my invention is the switching mechanism, its low power consumption, the device 23 of Figure 1, for example, requiring only 0.05 watt for operation.

Instead of being separate from the discharge tube I4, a device such as switch 22 may be mounted inside the lamp, preferably at one end with an insulated connection from the filament at the other end coming down the center of the tube, where it will not cast a shadow, to the proper contact. If the switch is not made a sealed unit when placed in the lamp, the contacts should preferably be surrounded by a glass tube to prevent the discharge going directly across them at breaking, instead of across the main lamp electrodes.

I have already explained how my switching device can be arranged, by connecting the contacts and the expanding wire I directly in series, to give a repeated cycle of making and breaking contact. I have found that when my device is so arranged it will act as a current limiting device. As the voltage across the device rises above a certain value, the make and break will occur more rapidly, keeping the average value of the current constant. Accordingly, the device can be connected in series with the lamp to limit the current therethrough, in place of a resistance or reactance, and will limit the current without much loss in the device. My device may be incorporated In the lamp itself, the whole sealed device preferably being inserted at one end of the lamp and in series with the main current carrying lead to the electrode at that end, thus providing a self-ballasting lamp which may be directly connected to a power line, without the use of a ballast impedance. If, in addition, a starting switch is placed at the other end of the lamp, a complete lamp results which may be placed in any convenient socket, directly across a power line, and which has only two outside terminal connections. In Figure 1, the heated element I is connected in series with the line, but between the lamp and the choke. In this way, the voltage between tungsten wire I and the contacts 6, I is kept low enough to prevent any arc or electrical discharge between the wire I and the contact circuit. If the wire I were connected in lead 20 between the reactance 2i and the line, the entire voltage of the reactance 2I would bepresent between the circuits of wire I and contacts 6, I, and might produce a discharge in device 22, with harmful results.

What I claim is:

A switch for preheating the filaments of an electric discharge lamp, said switch comprising: A sealed hollow container; an atmosphere substan-' tially free from non-inert gases within said container; a plurality of lead-in wires sealed through said container; a contact wire afiixed to one of said lead-in wires and extending in a. direction approximately parallel to the longitudinal axis of the container; another contact wire aflixed to the second ofsaid lead-in wires and extending transversely to the other contact wire and being normally slightly out of contact with said other contact wires; a glass tube placed loosely around a third and longer lead-in wire, and a small diameter tungsten wire stretched between said third lead-in wire and a fourth and short lead-in wire to push the glass tubeagainst the pair of contacts tightly enough to hold them normally closed yet loosely enough to allow them to open when the 5 small diameter tungsten wire expands on heating.

HENRY J. MCCARTHY.

Images