US2352893A - Cooling of vacuum devices - Google Patents

Description

July 4, 1944'.

' c. w. HANSELL 2,352,893

' COOLING OF VACUUM DEVICES Filed Dec. 18, 1941 2 Sheets-Sheet 1 F/lg. I N Z 1 RAD/0 c TRANSMITTER THROTTL/NG RADIATOR 340/0 TRANSMITTER INVENTOR CLARENCE W HANSELL ATTORNEY y 1944- c. w. HANSELL ,8 4 COOLING OF VACUUM DEVICES Filed Dec. 18, 1941 2 Sheets-Sheet 2 POWER SUPPLY V CONTROL CONNECT/0N5 I 70 OTHER POWER SUPPLY CIRCUITS INVENTOR CLARENCE M- INA/SELL ATTORNEY Patented July 4, 1944 UNITED STATES PATENT OFFICE g.

' ooomno F VACUUM DEVICES Clarence Hansell, Port Jefferson, N. Y., assignor to Radio Corporation of America, a corporation of Delaware Application-December 18, 1941, Serial No. 423,448

' 3 Claims. (01. 250-275) This invention relates to the cooling of vacuum tubes and to a system with means for controlling both temperature and the temperaturedistribu- An object of this invention is'to reduce'electrical flash-over of high' vacuum'devices when they are operated atrelatively' high electrical potentials.

Another object of this invention is to simplify and provide more efiicient cooling of vacuumdevices including those of the mercury vapor type and also to provide greater economy in operation.

Still another object of this invention is to 'provide means in which the cooling fluid temperature of vacuum tubes as used; for radio transmitters will be maintainedat a temperature which is higher than the normal room temperature so that the tube envelope will remain at a lower temperature than the tube electrodes.

Still further objects of this invention is to provide cooling means which embraces a water cooling system and an air cooling system, separate, or a combination of both.

A further object of this invention is to maintain the anodes of a vacuum tube at an elevated temperature at all times with respect to the normal room temperature, so that the amount of stretching and compression of the envelope material is greatly reduced, particularly when such vacuum tubes have metal-to-glass seals. The elevated temperature reduces the number and range of temperature cycles caused by starting and stopping operation. The reduction in temperature cycling at the seals will not only reduce cracking and failure at the seal, but will reduce gas release where the metal material in and near the seals is repeatedly stretched within its elastic limit, particularly when the metal employed within the seals is a material having a relatively low elastic limit, such as for examplecopper.

A feature of this invention is the means to maintain a high anode operating temperature by providing means for increasing the pressure of the cooling fluid sufiiciently to raise its boiling point above that of the anode temperature.

Another feature of this invention is to maintain an elevated temperature on the cathodes of the vacuum device by having a low heating current continuously flowing through the cathodes when they are in a normal inoperative state.

For many years one of the most frequent causes of failure in operation of high voltage vacuum tube amplifiers has been fiashover or short circuiting between the operating elements. Flashover dilficulties increase very rapidly when potentials exceeding 10,000 to 15,000 volts are employed and become so great at higher potentials that substantially no operation of 5 vacuum tube amplifiers is carried on at potentials exceeding 20,000 volts. In fact, radio transmitter vacuum tube amplifiers have been substantially limited to operation at potentials below 16,000 volts because experience all over the world has demonstrated that use of higher potentials results in rapid loss of reliability and early tube failures due to fiashover. This limitation of operating potential is practiced in spite of the fact that higher potentials, if they could be used, would permit higher efilciencies and higher power outputs from the amplifiers.

When using tubes of the UV-207 type, for example, I have found that internal fiashover is a frequent cause of trouble which results in service interruptions and in the loss'of a considerable number of tubes. relatively high potentials of 10,000 to 15,000 volts for considerable periods and then, without warning, will flash over. If the fiashover does not destroy the filament, or cathode, then after a fiashover, the tube often will operate normally again and, for a time, can withstand higher potentials than before. However, after an interval of time ranging from days down to hours or minutes, the tube will usually flash over again. The tubes seem to flash over after time intervals and these time intervals between fiashovers decrease in length successively until they become so frequent that the tube has to be removed, if the fiashover had not already destroyed the filament.

This tube fiashover is known all over the world, by those skilled in the radio art, as the Rocky Point effect because it was first found and recognized as a fundamental problem during the course of research and development being carried out at Rocky Point, Long Island, New York.

It is my belief that the Rock Point effect in high vacuum tubes is very largely due to the presence of migratory gas which is normally condensed or absorbed on the inner tube surfaces and which for that reason is not detected by the usual gas tests given to the tubes. This gas slowly transfers itself from one surface to another, alwaystending to accumulate upon the coolest surface. The rate of transfer from one surface to another may be very slow, perhaps like the transfer of ice fromone place to another which can occur in closed vessels having a difference in temperature even though all temperatures in the vessels are below the freezing point. In other words, I picture the migration of gas in Tubes frequently operate at v the tubes lighted during idle periods frequently helps to eliminate flashover when anode-and gridpotentials are applied again. I believe this may be very largely due to heat of thelighted filament preventing or: reducing migration of gas back to the active tube surfaces. According to my invention, it is therefore desirable to keep the cathodes of the tubes heated at all times, though the heating may b reduced during idle periods.

In water cooled tubes as customarily operated, the anode is often much cooler than other parts,

particularly just after power is removed. As aconsequence, there is a continual force, due to temperature difference, tending to make the gas accumulate on the anode surface. The hotter we allow the other parts of the tube, such as the glass portion of the envelope, to become, and the colder we operate the anode, the greater is the rate of transfer of gas to the anode sur face.

This surface, in operation, is subjected to intense electron bombardment which tends to vaporize the gas and to ionize it. The ions strike the grid and cathode, particularly th grid, often with sufficient force to knock out both electrons and metal molecules. Often the ion and secondary emission grid current has been observed to lower or even reverse the normal grid current fiow, forcing the grid potential momentarily positive while the anode potential is high. A burst of gas and metal vapor from the anode and other parts resultsra-nd the tube. flashes over. If the flashover has not destroyed the filament, it is very often found that, immediately following the flash, the tube will operate normally or even withstand much more than normal cur-,

rent potential. I believe this is because the flashover has cleaned. gas off. the anode. and other parts, causing it to' be transferred to inactive surfaces. However, when the anode is kept. cold, the gas migrates back due to temperature difference and conditions are re-established for another fiashover. Flashovers may be repeated at intervals ranging from seconds up to days or weeks, depending upon the quantity and nature of the gas, the temperature distribution and the mean operating temperature.

It seems probable that the gas which causes fiashover is not always or exclusively made up of molecules of material we commonly list as gases. I believe metal molecules may take part in a migration process or at any rate tend to settle most on cool surfaces. Evidence of this is often observed in glass envelope vacuum devices in which an internal coating or blackening takes place usually on the coolest surfaces but not on hot surfaces. It seems probable that a cold anode tends to accumulate metal vapor from the filament and that the vapor forms sub-microscopic crystals and sharp protrusions which are easily revaporized by electron bombardment. Presence of thorium in the tungsten filament, a

common impurity, may result in thorium accumulations on a cold anode surface.

Regardless of the detail nature of the causes of flashover, I believe it is conclusively proven that one means of reducing the prevalence of the phenomenon is to control the temperature distribution of the tubes to keep gas away from the active surfaces which are subjected to electron and ion bombardment. That is, the surfaces of active parts should be maintained at higher temperatures than inactive parts.

One means according to this invention for maintaining higher anode temperatures in the case of water cooled tubes is to raise the temperature of'the cooling fluid. If the cooling fluid is water; which is customarily used, increasing the water temperature tends to reduce the permissible anode-power dissipation because of smaller temperature rise permissible before boiling starts. I may overcome this difficulty by subjecting the cooling system to an increased pressure suificient to raise the boiling point of the water above the anode temperature.

Another means for maintaining higher anode temperature is by an air cooled system including a circuit arrangement for partially heating. the. surfaces of the active parts during idle periods.

This invention will best be understoodby referring to the accompanyingdrawings, in which:

Fig. 1 is a fluid arrangement of'asystem of. this. invention;

Fig. 2 is another arrangement. of this invention similar to that of. Fig. 1 exceptv that an elevated reservoir is provided; and

Fig. 3 is a. powerv controlled circuit for the system of this invention.

Referring now in. detail. to Fig. 1 of the drawings, wherein one of. the simplest possible. ways of operatingthe cooling fluid and therefore the anodes of the water-cooled vacuum tubes at higher temperature, is illustrated, in this figure there is indicated a radio transmitter I which includes Water-cooled vacuum tubes, only one of which has been shown and. indicated as 2.

The tube includes anode 2A cathode 2B.and.

grid 20. Cooling water for the tubes is circulated in a, closed circulating. system comprisinga water pump 3, piping. 4,. an operating valve. 5;,

a throttling valve 6,.radiator T, and athermostat conventional mounting and.

8. Unlike the cooling systems of the prior. art, the anode. is mounted so as to be at a higher elevation than that of the tube envelope, and the water pump for the system is chosen to provide higher than the usual pressure so that the boiling point of the water is raised. The increased pump pressure, after the fluid passes through the vacuum; device 2, is absorbed by placing a' resistance to the flow at some point on the outlet; side of the connection to the vacuum tubes, either by means of a relatively small connection. or preferably.

by the adjustable throttling valve 6. The extra pump pressure appearsas an. increase in pressure of the water in the vacuum tube: cooling jackets water temperature may be increased about 34.4

centigrade without increase in the probability of boiling. at the vacuum tube anode surface.

In order that the temperature of the cooling water will always be maintained at an elevated temperature, the thermostat 8 is employed for controlling the flow of water through the radiator 1. This thermostat may be similar to those commonly used in automobiles to control the operating temperature of water circulated through the engines. Then, if when all the water flows through the radiator an amount of power equal to or greater than the power carried away from the radio transmitter by the cooling water can be dissipated by the radiator, I may maintain substantially constant elevated temperature of the circulated water.

If the'transmitter is shut down and remains idle, the thermostat automatically closes off circulation through the radiator and only a small amount of electrical power in addition to water flow friction losses is likely to be needed to keep the circulating water up to operating tempera ture. I propose to supply this power by dissi pating electrical energy in the tubes even when the transmitter is idle. Usually it will be sufficient to keep a reduced amount of power in the vacuum tube filaments but power may be dissipated in the grid and anode also with still better results in controlling gas distribution, if desired.

The radio transmitter should be provided with air blowers (not shown) which maintain relatively low temperatures of the glass and other relatively inactive portions of the tubes so that these parts will be maintained at all times at lower temperatures than the active parts.

The arrangement shown by Fig. 2 is somewhat similar to that of Fig. 1, except that it does not require increased power to drive the circulating pump 3, for the reason that a standpipe 10 or an elevated reservoir which is located substantially above the circulating system provides the means for increasing the pressure, which when full of water subjects the whole circulating system to increased pressure.

Other methods of increasing the pressure would be to connect the cooling system to a source of water supply, such as, for example, a public water works system, in which a higher pressure is maintained than the normal pressure needed for water circulation. Such a system would automatically replenish any leakage in the system and prevent the possibility of failure of r the cooling system due to loss of pump priming.

In operating conventional water cooled tubes, it is customary to mount the tubes with the glass ended portions up. With this arrangement, when elevated cooling liquid temperatures, or air cooling fins are used, there is a tendency for heated air to rise up around the glass portions and to increase the glass temperature. Under these conditions, continuously operated forced air cooling systems should be employed to force down the glass temperatures to less than the anode temperature at all times. I have found that it will help to reverse the customary tube mounting method by placing the glass end down. In this case, it may be possible to omit forced draft cooling, particularly while the tube is not in use, because natural convection may provide sufficient cooling for the glass during these periods.

In Fig. 3, there is shown a portion of the power control circuits'of a radio transmitter employing vacuum tubes of the 'type designed for water may. be fused at l8 or connected in series with.

a circuit-breaker (not shown). Switch ll isno'rmally kept closed at all times when the transmit-. ter is in service or available for service. The,

tubes l9 and .20' each includes an anode 20A, cathode 20B. and grid 200. Power inputto the cathodes of the, tubes l9 and 20 is maintained while the transmitter is not in service; :This is. accomplished by any adjustable impedance;

preferably that of a variable or adjustable rea'c-. tor. When the tubes I9 and 20 are to be placedin active operating service, the stop-start control switch 22 is closed, thus causing ajcontactor relay 23 to close. Contactor relay 23 is provided with a plurality of contacts 24 which close circuits for applying other potentials, such as gridbias and anode potentials. Closing the stopstart control switch 22 also starts the fans or blowers 25, 26 which force a rapid flow of air through the anode radiators so that the high anode dissipation during normal running can be accomplished with adequate air cooling, as described in more detail in the above mentioned Finch patent, particularly that portion relating to Fig. '7 of the patent.

In the arrangement shown by Fig. 3, cool ingoing air passes over the glass envelope surfaces of tubes l9 and 20 and maintains these surfaces cooler than the anodes. It is desirable to employ auxiliary blowers or fans, conventionally represented by 21 and 28, to force a current of air intimately in contact with the cathode stems, seals and tubulations around the grid leads where considerable heat may appear due to conduction from active portions of the tube and also to highfrequency resistance and dielectric losses. In the practice of this invention, all portions of the glass should be maintained at a temperature well below that at which electrolysis in the glass becomes a factor in tube operation and life and low enough to cause the migratory materials inside the tube to accumulate on the glass surfaces, particularly surfaces removed from metal parts, seals and high potential gradients.

An incidental advantage of maintaining the anodes at an elevated temperatur at all times is that stretching and compression of envelope material, particularly at the metal-to-glass seals, is reduced by reducing the number and temperature range of temperature cycles caused by starting and stopping operation. It is my belief that reduction in temperature cyclin at the seals will itself not only reduce cracking and failures at the scale but reduce gas release Where the metal material in and near the seals is repeatedly stretched beyond its elastic limit, particularly when the metal is a material having a relatively low elastic limit, such as copper.

The principles I have described for controlling the distribution of condensed volatile materials in high vacuum tubes apply even more forcefully in the case of mercury rectifier tubes and like devices in which case the presence of 1iquid mercury condensed on the surfaces in unwanted places, due to incorrect temperature distribution of the surfaces, is often plainly visible. When suitably designed, these low pressure mercury de.

Power for the transmitter.

vicem. employ; thm expedients: described: forruse with;-hiah;vacuumtimes.v

whilezonlw-aa few: modifications; of this inventiom are.- ShOWIL; to; he distirmtly5= understood.-

that. the inwentiomshoutld notibealimiteilnmoises lytheretm.

What is: claimedis:

L Apparatus. for; reducing; the-tendency to: flash; over by avacuum tu'be, havingganoda and. cathode electrodes, com-prisinga; jacket for; the: anode. eithetube, a. pumr and connections; for circulating: water; through; thexjacket under pres-: sure; means for maintaining: the temperature, above the normal boiling temperature at the: water; and means fordissipating electrical err-- ergy i'n someot the electrodesmf said tube during idle operating-periods-of said tube.

22 Apparatus for reducing: the tendency to flash over by a vacuum tube having anode andcathode electrodes, comprising a jacket for the anode-ofthe tube, a pump and connections for circulating water through the-jacket underpressure" for maintaining the temperaturesubstantially above thenormal boiling-i temperatureof the water; meansfor dissipating electrical ener- 9Y in" some: of: the: electrodes offsaidtube: during:

idl operating periods of sat-(L tube, and! anthem mostatically controlledscooling device associated Withs said connections.- ion increasing. the cooling; Ofi the-waiter: duringcperiodsc ofi operation of said 150% ofhthe; time, a pump and pipaconuections; an

operatin zvatlve-in the, inlet water pipe. for said jackemathrottlingjvaluein the-outlet water pipe, aradia-tor. flow conneeted across a sectionof the outlet Dining on the; sidaoi the throttling valve removed irom the-tube and a. thermostatically controlled valve for controlling water. flow throughsaid radiator for circulating water through the -jacket.under pressure. for maintaining thetemperaturesubstantially above the. norl ma-l=boilingtemperature. ofthe Water, and means,

for. dissipating electrical; energy inv some of the electrodes of saidltube during idle operatingperiods of: saidvtube.

CLARENCE W. HANSELLL

Images