US2011605A - Central-blast rectifier and water temperature controlling means therefor - Google Patents

Description

Aug. 20, 193:5. A L ATHERTN 2,011,605

CENTRAL BLAST RECTIFIER AND WATER TEMPERATURE CONTROLLING MEANS THEREFOR Filed Jan. 28, 1952 Z SheetS-Sheet 1 iJgij /z 26' 24 /8 I A324 25 z 4a 2 1 L 54 l I INVENTOR -WITNESS ATTORNEY Aug. 20, 1935. A. L. ATHERTON 2,011,605

CENTRAL BLAST RECTIFIER AND WATER TEMPERATURE CONTROLLING MEANS THEREFOR Filed Jan. 28, 1952 z Sheets-Sheet 2 ATTORNEY Patented Aug. 20, 1935 TE 'ENT OFFICE CENTRAL-BLAST RECTEFIER AND WATER TEMPERATURE CONTROLLING MEANS THEREFOR Alfred L. Atherton, Verona, Pa., assignor to Westinghouse Electric & Manufacturing Company, a corporation of Pennsylvania 7 Application January 28, 1932, Serial No. 589,415 1'7 Claims. (01. 250-275) My invention relates to means for improving simultaneously operating arcs in the same the operation andreducing the space requirechamber. ments of mercury-arc metal-tank rectifiers and The present invention it hasparticular relation to means for influencis based upon thetheory ,5 ing or controlling the flow of vapor in a multiinsulating gas pathsin a rectifier, which causes anoderectifier tank and for providing a blast them to break down, occasionally, at voltages of of. sufliciently deionized vapor in the region of the order of or less,'of their ordinary breakthe anodes to prevent arc-backs, while helping down potential, is attributed to the presence of i to provide sumcient vapor for the main arc path. ionized mercury vapor f om the a c-pa s 10, The thought behind my invention is to provid other anodes which are active at the moment. means for preventing a phenomenon which con- This ionizedmercury p a o g to my D stitutes a peculiar difierenc e between. the insulat- 911T belief, s a r ed about in a Variable and ing gas paths from the cathode to the anodes e or less n l d way y th last f (and from one anode t th when t parmercury vapor from the cathode, in rectifiers as '15 insular anode is notcarrying current t h heretofore constructed. Atintervals there will some other anode is active, on the one hand, and be the combination of a Sufiicienfi density of an ordinary insulating gap in a s, o 'th th ionization concentrated in the space around an h In y r ctifi r the Object is, of com-Se, anode on which the reversepotential is applied 1 to have the gas paths between the inactive anodes With Sufiicient instantaneous potential to cause and the other parts of the structure f m the breakdown wh chdevelops into an arc-back. the function of insulators toprevent current fiow lS comblnatlon of circumstances occurs only to and from said inactiveanodes during the por- Occaslonany because 1t reqllllfes h acculllulatlon tion or" the cycle Whenthey are not supposedto i gg 3 3: 2 5 2 22 fi n3: 2

r" rre t. If itwere not for the resenoe Y 25 gi gg gn orjanodes carrying cufrent as cause the movement of the'positive ions is conthe terminal of an arc or'arcs, in the same tank, l d y blast vapor m h cathode stand a potential of somew ere around 10,000 volts, with the spacings, vapor pressures, and mai Into i i g of a large terials of the anodesand the gas, commonly met 2 32 5 2 gi g; 5 2 2 2 etc 0 fifii gi i 'f Wi metal'ta'n'kiecmfiers' 3 m lectlfiet plication is introduced by the fact that along the inwh1ch the maximum potential dlfference 1s path of flow of the Vapor there are portions in of order 9 9 volts no fi should which energy losses take place because of the flow ever occur, but 1t1s Well known-that such backfires, or arc-backs, do occasionally occur. The 9 F current" Whlch r t local h theory has been suggested that the remanent 22:2 :25 gifg j if gpg gg gg and ionization; which remains in the region of an The conception that thgsourceof the'se Ceca? q f alter P 'tgrmmatlop thls penqd sional breakdowns is the concentration of ioniza- 40 activity explains these occas onal back-fires, and tion from an outside source b1 ought about b means have been added with the view of quickly r y deicnizm thmse remanem ions the flow, into the high voltage-gradient space, of

v g 1 i an abnormally large concentration of positive observation has however that; the ions, seems to me to be reasonable. It also seems back-fire m PQ f f Phemmemn, to me to be reasonable that the extremely varied d my 5 Of t b fiw of arc-backs flow of gas, resulting from the high velocity and during the period of inactivity have not shown extremely irregular path will explain the long a conslderable F e y i nes 'F time intervals and the variation in time interinltial remanentomzation period of transltlon vals, rem conductivity to insulation characteristics Considerations Such as the foregoing lend i a gljen path y, F 9 m creased support to the corrective measures dismd, that the remanefit 10111294111011, at least 111 closed in my application Serial No. 560,722, filed the mod n typ s f fi a Ordinarily n- September 2, 1931. If we can intr'oduce such a struoied, although a Contributory Cause f a flow of tin-ionized vapor that the ionized vapor fire, Cannfit a'pledominant e f b cannot" enter the high-gradient portion; ,this

fire in the presence of ionization from" other source of aroing' backis' removed. Inthe rectior belief that the peculiar characteristic of the to supply the flow of un-ionized vapor.

' My present invention is partly in the nature of-a continuation, in part, of. my above mentioned application, and, in part, it is directed'to modifications and improvements thereon.

to a separate mercury pool, and I havedis covered', more definitely, the control of the temperature gradients within the rectifier, if the un-ioniizedeblast is to beused'" effectively,whether itoriginates from the oathode poolor from an auxiliary mercury pool. Of course, it is knownthat there are. many 7 other cases of arc-backs, each of "which'hasto For example, the presence 7 be taken care of. of too muchforeign gas will cause troubleand it fis necessary tog'et-this-kind of thing under control, regardless of what other mechanismsof "vapor-flow" may besetup; The {propen design f and controlof means for providing an'un-ionized mercury; blast, in accordance with my present invention, afiords one of the most effective ways of "which 1' know,

for avoiding the troublesdue to foreigngaso This maybe explained as follows gas'which is givenoff, 'therebydepositing' said 7 foreign gas at.-a portion of -vacuum pumping connection isfmade. This re-: 1 sults-in ireeingthe space within the anode shield from entrance ofserious quantities 'of foreign .-.ga i W are carried along'in I vapor flow an'dhavefaccumulated in pockets, in. previous constructions; thereby preventing, the mercury vapors from .the tank, thus lowering the efliciency ofthe recti-" lier, bu Whatis worse, wagary. of the vaporpufiwthrough-such agas pocket and blows the Thevapor from the cathode may be assumed to be relatively free from foreign gas. .The principal sources of foreign gas, assumingthat Iany'rubber gasketsi which arejutilized' ior vacuum-sealing processes are *s'ufiiciently cooled, I are the anodes andthe hot parts'surrounding them. I believe that-"these gases, which are given:- off -by;itli'e hotter parts of the apparatus,

the general direction-of the reaching the cold walls of accumulating until-some flow sends an. eddy or gas into a 'region where back-fire may occur. ;In

l.my.:..vapor-blastcontrol system, a gas-free blast.

of vapor from the cathode :is-definitely directed across the anodes, and particularly across the mouths of the anodeshields,-where such shields central l2 cooled by water which entersthrough an inlet are utilized,-so'as ..to carry alongany foreign There is "another factor in the behavior of a h v rectifier which has to be taken into account in determining 'the amount. and density of vapor current flowsf'in an anode shield to theanode -head,-the vapor in theshield is heated and tends v at'thebottom of the "shield; When the current flow ceases, the vapor re-enters""the anodesh'ield due'to the reduced temperature lof the gasin the shield. 'I'his renecessary in the auxiliary stream. When the to expand and. flow out sults' in a'breathing-efiect which tends to draw into the shield whatever. vapor may surround theopening, just ati thetime when the'neg'ative potential is being applied." It is .thus necessary that the vapor-surrounding the opening of the fanon ie'v shield shall, .be' to a sufficient extent unionized, Doubtless the; preach the proble'rn of determining the amount only way forlus to ap- I havebeen able, as described in the present application, to obtain a blast of sufiiciently un-ionized vapor. from thecathbde pool itself, without resorting central blast from th the nature and importanceof ameans forincreasing the" central blast,

which lies in the direct arc'path between the cathode and 'thefsev'eral anodes. also showswhat is now my preferre'dmethod of controlling the temperatures of the various coolthe section plane being indicated on the line IV- IV; I V W Fig; "5 is a partial vertical, sectional view tureresponsivedevice associated therewith.

"continually being the tank to which the aroundthe upper portion of -bafiie l2 and above the conical bafiie i3. Each I.J.. T. Mathews Serial No. .26, 1932 and assigned is the use'of a quartz v anodelead or shank 263 and counter-sunk into the anode head i6." Surrounding the quartz tube l9, and resting'upon the top of the anode head 96,

i the lower end of space, as well. as

I 7 2,011,605 fier constructions shownin my copendingappli cation just mentioned, use was made of a separate mercury pool, heated by an auxiliary source,

of auxiliary vapor flow which will be required will be by trial, at least for the present.

With the foregoing and otherobjects in view, my invention consists in the various structures, combinations, [methods and arrangements of parts which are hereinafter described and claimed and shown in the accompanying drawings, where- Figure l is a sectionalview of a rectifier embodying my invention in a form which utilizes cathode pool itself, with a series connection 'of the various coolers for the purpose of obtaining proper maintene'nce of the blast.

Fig.2is a'similar View showing the use of an auxiliary mercury pool or auxiliary portion of the mainpool which is segregated as the source of the un-ionized blast.

' Fig. 3 is aview similar to Fig. 1, illustrating the relative strength of as compared to the radial blast This figure ers, with automatic means for this purpose.

,Fig. 4 is a sectional view of the center cooler, F 3 by through the center cooler, showing the tempera- Fig'. dis a diagrammatic View illustrating the water flow system for the'coolers, 1 In'Fig. l'my invention is shown applied to a metal-tank rectifier comprising a main tank portion! having an insulated cathode receptacle 8 in the bottom thereof. Ajliquid mercury cathode pool Bais disposed in the-bottom of said cathodereceptacle 8. The side and bottom walls of the" tank are provided? with a tank. cooler or water jacket 9 'and the cathode receptacle is :provided with a water'jacketor cathode cooler i9. Depending from the central portion of the 'coverplate H of the tank is a central cooler or tubular baffle l2 hich terminates, atits lower 'end, in an annular conical baiiie l3. Both the and the conical ba'file 53 are pipe l' liand leaves through-anfoutlet pipe vl5.

A "plurality 'of anodes fiii'are'disposed in a ring the centraltubular anode may be providedwith I an anode shield H,

:the bottom ofwhichis protectedby a grid 18.

also, I-utilize an anode construction Preferably, as shown and claimed per se'in an application of 588,915, filed January to. the Westinghouse Electric & Manufacturing Company. One of the most significant things about-this anode construction tube l9 surrounding the is a quartz'disc or ring 2! which is spaced from the anode porcelain 22. This the quartz tube, i9, seem to be necessary, in addition tothe arrangement of the fmercury blasts in my .present ap plication, in order to secure the fullest measure of freedom from back-fires which .has .tofore.

'not f been attainable herethe approximate temperature gradients which are necessary for the or thermostat which is The ring of anodes' lfi is disposed close in to- Wards thecentercooler I2, being much closer to 'said coole'rthan to the side walls of ,the'tank l. Surrounding'the ring of anodes is an intermediate cooler 24', and surrounding that there may bean outer cooler 25. The outer cooler 25 and the tank cooler 9 provide what I call the final con-- Toy "and leaves it at a Y with the velocity.'

, densing walls; or

surface of the mercury,

, ing a downwardly and walls of which the predominant characteristic is thatthe mercury condenses and runs as a liquid back to the cathode, as distinguished from a wall on: which the mercury condenses and re-evaporates." For brevity, I shall refer to these. final condensing walls as the side walls of the tank 1, regarding the outercooler 2'5,

if it'is used, as simply a continuation of the side walls of the tank; To make my meaning more clear, I shall refer next to the temperature gradients which I maintain within the tank, and to the general phenomena of mercury vapor flow.

1 Mercury vapor leaves the cathode at a high velocity, probably muchhigher than the velocity corresponding merely to the temperature of the although this temperaof the mercury vapor ture is the. hottest part blast; As soon as the mercury vapor leaves the cathode surface, it expands with molecular movements in all. directions, but with a preponderance of the molecules or particles moving in the direction' of the temperature gradient, ortowardthecooler portions of the rectifier. When mercuryvapor touches the central cooler, which, according to myinvention, is the hottest cooling surface, being intermediate in temperature be tween the temperature of the surface of the meri cury pool and the temperature of the final condensing walls, the vapor" approaches the surface of i the cooler at the high velocity abovedescribed much lower velocity corresponding to the molecular velocity corresponding to the temperature of the cooler. I describe this process as a condensationandcre-evaporation of the mercury on the walls of the cooler. During this process, the vapor also loses the ionization which it had when ode.

it left the region of the cath- In its progress up through the. [central cooler l2, the mercury vapor thus becomes deionized and it loses much of its velocity and, of course, much of'the vapor pressure which goes When the central blast of mercury vapor reaches the top of the central cooler, itiagain expands, as afree gas which is released, under pressure, in a space,in a manner similar to that which was described for the blast astitifirst left the mercury cathode. Some of the mercury vapor comes down and condenses on the top walls of the ,conicalbafiie l3, and some goes out totheintermediate cooler 24, thereby providoutwardly directed blast of'substantially un-ionized mercury vapor which is fairly free of foreign gases, sweeping past the bottom mouths of the anode shields l1, and carrying along with it any foreign gases which are liberated from the anodes l6.

In'addition to-the central blast which rises from, the cathode, there is also .a radial blast which passes outwardly under the conical baiiie i3 and thence out to the side walls of the tank and to the outer cooler 25, the two blasts joining and condensing, finally upon the final condensing walls of the tank cooler 9 and the outer cooler 25. r Y

In order to maintain .the vapor flow just described, it is essential that the central cooler shall have a temperature cooler than the initial temperature of the vapor blast, so that the mercury vapor will condense on it, and yet warmer-than the final condensing walls, so that the expanding vapor, as it loses its velocity and pressure, will re-evaporate and move onas un-ionized vapor. The intermediate cooler a temperature more nearly like that of the inner cooler l2.

In the particular system shown in Fig. 1, the approximate temperature gradients of the coolers are maintained by introducing the cooling water first into the cathode cooler in, as indicated by the inlet pipe 30. The water leaves the cathode cooler 10 through 'a rubber hose connection 35 and entersthe inlet pipe 32 of the tank cooler 9. It leaves the tank cooler through the outlet pipe 33 and enters the outer cooler 25. Leaving the outer cooler through the outlet pipe 35, the cooling water then enters the intermediate cooler 2 rom which it passes, through pipes 36,,to the inlet pipe #4 of the central cooler l2 and conical bafflc l3. The water leaves the central cooler i2 through the outlet pipe l5 and enters a coverplate cooler 38, from which the water is discharged through a suitable outlet or drain pipe 39. In the foregoing system of water connections, it will be observed that the hottest water is at the'inner cooler l2 and that the coldest water is in the tank cooler 9 and the outer cooler-25.

The above-described scheme of controlling the temperatures at the various parts of the rectifier so as to produce the conditions necessary for the vapor to flow from the cathode to the cooling surfaces, through paths apart from the danger zones, (the regions of the anodes where back-fires are likely to occur, is very important in reducing the frequency of these back-fires. The particular structure shown in Fig. 1 illustrates the principle, although I do not believe that it carries it out as far as possible, For example, if we make the center baffle hotter by, say, 10 degrees, than the balance of the condensing surface, then the vapor pressure just at the surface of this baffle is approximately twice as high as the vapor pressure at the cooler surfaces. This certainly produces a tendency for vapor to flow from the center bafile to the cooler surfaces and this vapor, flowing in thisdirection, will carry foreign gases and ionized mercury vapor with it toward the main condensing surfaces and thus into pieces where less harm is done. t is, of course, necessary that the hot walls be cool enough to ensure condensation, so that saturated mercury vapor will be present.

I also find'that the groupingor nesting of'the anodes l6, close in towards the center cooler if,

has an important bearing on the freedom from back-fire. In other words, it is important to provide a free radial space between the ring of anodes 7 l6 andthe side walls of the tank 1. This is pari the bottom of the intermediate cooler and out to the side and bottom walls of the tank 7. I

' The main arcing path between the cathode and the several anodes is around thecentral conical baflle l3, although it must be understood that the arc, taking place, as it does, in an extremelylowpressure gas or vaponspreads out pretty largely over the wholetank space, some of it probably even 26 should, I believe, have 2' passing .at least occasionally, through the cen a tral'blast path hereinabove described."-

A further point to be considered is'that the v2 c-' uurn pumping connectionqshouldbe arranged so as to open'into the tank at the point where. the gases accumulate.

In the present: rectifier, as above described, thispoint isat the. top of the side walls of; the tank portion. l;

i inear the side walls :of the tank, inFig. 1.

I have found some indicationoi a. tendency oi the design'shown in Fig.1 to be sensitive torapidly applied loads.

While thenurnber of arc-backs is relatively small, all of thearc-backs which-have mercury pool spread occurred soiar have occurred at a time when the loadfwas suddenly increased, say, from 25% of full load to 260% of full load, the frequency of I back-fires, under such circumstances, being about one in' a thousand times.

iew minutes; after such an increase :inloadthe arc dropis considerably higher and is considerably erratic.

. I; also find that, for a I attribute these phenomenan ainlyto the slowness with which the cathode spotsonthe out over the surface of the pool, so as to carry the time which it takes for the average tem peraturegof the top surface the mercury pool to increase-tea valuenecessary to supp y the increased mercuryvapor which is necessary to be present inthe arc 'paths. In other words,

the central blast, in the particular design shown,'

does not seem' to-be large enoug to supply. a 'suf-f ficient number of uneionized mercury molecules or vaporin the, regionunder theanode shields ll,

' sothatthere arenot sufiicientgas'particles there. j to carry the increased load current without excessive arc drops. i

* There are .twoways 'in which might correct this'defect, according to" my prese'nt theories.

- baffle 43 is provided between the central cooler 12' i and a central portion 0 the mercury pool, which is heat-insulated from indicated at M. A heatert l is provided ior this One isto utilize an auxiliary source of mercury vapor, for theun-ionized last, :as I haveindif cated in Fig. 21 "Inthis case; atubular quartz thev cathode cooler 19, as

1 0 wh r byfmer u fmar-be central .mercury boiled off therefrom vat anyrate which maybe form of embodimentof vmy indesired. In this 'vention, one or more starti 'g'anodesf it are provided, in the outer active annular portion of the I mercury poolil'l-"instead of using asingle cen trally located starting anode {it as inthe Fig. I

construction. Fig; 2""a1so showsafeature which-may boot utility in any of the forms-f embodiment'o'fmyj invention, namely, a depending tubulari baffle 59 creased central mercury of thej'ring'oi anodes depending from the mouths of the anode shields ll.

An additional method oi' obtaining" the? in endoi the central'cjoolerfltd isfjopenedouthke,

a tunnel, as indic'atediat 51, ,so that itsp'reads oversubstantially thewh'ole area oiithe cathode pool, so as totake instant advantagejof the increased mercury ebullitionjrom the'regiojns of V the cathode spots, no matter in load. Inthis way, it is believed that more mercury'vapor will immediatelypassup through" the'centralbafile or'cooler 12,without waiting 7 h o o I haveindic'ated a j vacuum pumping connection l l'in the cover plate 1 1'0 the increased load, and to 7 41; cover plate l tin the region, 3 l6 ior'the purpose of assist;- ing" in directing the'blast downwardpast the I blast which I believe to be needed in'the particular designshown in f e Fig. l, is indicated inFig. fi wher'ein the bottom I v where they arelo ated, upon the occurrence of a sudden increase for the temperature .of the cathode, spotsfto spread out over thesurface ofthe mercury,upon

the occurrence of.asdddenincreaseiin load." I

. Figure 3.-also shows what I nowlbelievev tobe the best watercirculating systemin the coolers.

As in Fig. 1, the coolers are connected in series,

' althoughin the broadest aspects of 'my' invention I am not to be limited toaiseries connection.

InFig. 3, however, the cathode. cooler i0 is placed in thethird position insteadofiihthe first positionv in. the series oi jco'olers which constitute .the

path or thecooling watenandwl provide means for artificially heating allot the. coolers except the tank'cooler 9 and the outer cooler 25,'during periods of no load," so that the parts arekept at a temperature suitable for immediate assumption ofload. j 2. As shown in -Fig. 3,"as Wellasinthe diagramvalve or" so-called water-flow regulatorllland thence'tothe'tankcooler 9 and outer cooler ZEl. From thefdischarge pipe 350i the outer cooler, a

hose connection 72 leads bothptoithe' cathode cooler ID and to a heater tank'le. Tracing first the normal-path of the water through the cathode cooler, a second hose, connection 116 carries the water to the inletpipe? of the. intere center cooler and-the cover-plate to the outlet pipe 39 which is connectedboth to the drain Ti "and totheinlet side of a pumpf-lawhich, when operatinggdischarges water into the heater tank .HLirom which it is. recirculated .thrOughthe cathode, intermediate; center? and "cover-plate I coolers" H3, 2t, "ltwand 38;. The pump 19 is adapted to be driven by -a .motor:8 l which may be energized, whenever current is supplied to. an

electric heater 53in the heater tank i l, in re-' sponse to the operationof a contactor switch 8d. It will be understood that, normally, whenithe the electric water heater- 83 and the pump'l'fi 'will be unenergized, so'that no water will flow therethrough.

To give some 'idea of the temperatures involvedy l shall indicate certain temperatures which'ha've been found-'to' be satisfactory, al-. though it willbe understood thatI am not to be 'matic viewpf Figw6, water isiled first through'a mediate cooler 2d, whence itpasses through the altogether limited tothese particular tempera-- I tures, Inlet water is supplied,at almost any desire able temperature, which preferably should be lessthan 35 0., as indicated in Fig. 6. After passing through the tank cooler, a water, temperature of as large as 40 C'. may be utilized. The'waterwill be slightly furtherheatedcinpassing through the sol outer'cooler; reaching a. temperature of possibly Subsequently, the water temperatures will be- 42, at the end; of thecathode cooleri 45;at

the end of 'theintermediate .cooler; 52 at the end of the center cooler and possibly. 55 at the discharge terminal 39.*rThese teniperatures' are merely given for 'illust ative purposes,

During periods of no1oad,; or extrem'elylight loads, .it is very desirable to'preventthe temperatures; of the cathode and of the intermediate and center coolers-from falling totoo low a value.

I, thereforaprovide means 'for shutting ed the water-flow regulator" l8 andfstarting' 'an inter mediate circulationfby means of the pump heater tank, as previously described. In such a case, the inlet water which is fed first to the cathode cooler from the heater tank may have a temperature ,ofabout '51? C., after which it is conducted intothe intermediate cooler at a tern perature of, say, 48 0., then to thecenter cooler and I rectifier tank is carryingany appreciable load, t

Bands and contracts in .the pump motor 8|.

. ga'gement of the contact finger 92 plate at a temperature of 47, returning to the pump at a temperature of possibly 46 C., from which itis discharged againinto the heater tank 74. I

It will be observed that, at alltimes, the intermediate and center coolers are kept at a temperature intermediate between that of the cathode surface and that of the tank cooler and outer cooler, so that the vapor fiow, as previ ously described, may be satisfactorily maintained.

It will also be observed that the principal source sensitive to load changes.

y In devising an automatic heat control system, therefore, I utilize a thermostat or heat-responwhich is responsive to the water temperature in the center cooler I2a. As shown in Figs. 4 and 5, this heat-responsive element may take the form of a liquid container which is inserted in the center cooler |2a and which is connected, externally of the tank, by means of a pipeconnection 81, to a metallic bellows 88 which exresponse to temperature changes in the center cooler. This metallic bellows is mounted as an integral part of the waterfiow regulator, so that it turns off the water at a temperature of, say, 50 in the center cooler, and turns it open wide at. a temperature of 58 C. in the center cooler.

' Upon either the same or a duplicate thermostatic device 86, 88, an insulating support 89 carries a contact finger 90 which moves between two stationary spring-contact fingers 9| and 92. Energy is supplied from a relay battery 93 to the movable'contact finger 90, so that when the back contact 9! is engaged, at a temperature of, say, 45 C.,' the contactor switch 84 will be energized in order to turn on the water heater 83 and As soon as the contactor switch 84 picks up, a hold circuit 95 is ener' gized through the normally closed contacts 96 of an auxiliary relay 97 which is tripped by the enat a temperature of, say, 48 C., so that the water heater and pump are cut off at. this temperature. Whenthe temperature of the central cooler l Zincreases two more degrees, or to 50 C., the thermostatic device begins to turn on the'water-flow regulator, as previously described.

very excessive temperatures should be reached, a final limit-switch contact 98 is reached, by the contact member 99 of the thermostatic success of'a sectionalized-type design which constitutes the subject matter of my copending application Serial No. 589,414, filed January 28, 1932.

sectionalized design isa. much smaller size than has heretofore been possible. I feel that it is the herein-described control of the directionand magnitude of the fiow of mercury vapor within the rectifier tank, which has made it possible for me to provide a rating of 750 kilowatts at 600 volts in a volume only one-half as great as has previously been required for a 500 kilowatt rating at 600 volts.

While I prefer to utilize an automatically controlled temperature regulation in the various I contemplate that numerous changes will be made within the scope of my invention, such as the possible elimination of a part of the shielding around the anodes, which would result in a further reduction in internal losses. I have also obtained a material gain in efliciency by utilizing a 2" depth for the cathode receptacle instead of an 8" depth.

tion is increased by my vapor-flow means for preventing back-fires, thus making it possible to dispense with many of the shields and other obstructions which have been heretofore'needed for the purpose of preventing back-fires and which consume several volts of arc drop. 'For example, previous Westinghouse designs of 500 kilowatt rectifiers, as well as present competitive designs of the same capacity, have a full-load arc-drop of approximately 21 volts. The design shown in Fig. 1,..which is now in service, has a full-load arc-drop of 20 volts, at a 50% increase in rating over the previous and competitive designsof twice the size. An experimental modification, making full use of the principle discussed hereinabove, has been made and tested and found, so far as the restricted tests indicate, to be satisfactoryyhavihg a full-load arc-drop of approximately 16 volts. I believe that a reduction in the anode shielding and an increase in the central blast over that shown in'Fig. 1 will reduce the full-load losses to possibly 13 volts.

In connection with the disposition of foreign sembly. My own experience, as well as that of others. has clearly shown that the quality of a rectifier may be determined by the degree of freedom from foreign gases. In building my rectifiers, I go to rather extreme lengths to decrease the amount of foreign gas. One of the most important items in this regard is the practide of high-temperature pre-treatment of the graphite anode-head under vacuum. All other materials which might emit foreign gases are pro-treated before assembly wherever this can reasonably be done.

A practice which I believe is novel in metaltank rectifiers is to determine the completion of the treating-out process by a measurement of pressure-rise within the rectifier at a maximum contemplated load current for a given period of time, with the pumps not in operation. The previous practice of treating until a given degree of vacuum could be maintained by the pumps did not give complete information as to the condition of the rectifier from the standpoint of freeof said central tubular' bafile, also spaced from the'cathode, a plurality of anodes spaced in apresncd' of foreign gas,

10 stitute of? Electrical! Engineers on January' 28,

- the cathode, a iplurality'of circle close "touthe upperportion: of said central tubula'rsbafile and for the-most part. over said dom from 1 ooclude'd '1 'g'ases i- I have found that:

even though a high'degree 10f vacuum may be maintained by: the operation rectifier from the hot surfaces 'to'the' pumping connection, is definitely and seriously detri mental.

:.The subject-matter of the present application is more or less completely shownand discussed in my "paper presented beforethe American In- Iclaim; as my invention:

.1.'A multi-anode, metal-tank, mercury arc rectifier characterized by1 .'a central tubular bafile cverthe cathode in spaced relation thereto, an}- annular bafile'mountedi around the bottom end QfT said central tubular.

the cathode, a'plurality "of anodesspaced in a qcircle'close to' the upper portion of "said central tubularbaflle and fo'r'the most part over said annular, baille, and a second tubular baffle suri roundingsaid-circlefof anodes. J

2. A multi-anode, metal-tank,umercury-arc rectifier character-ized'by a central tubular bafile overthecathode in, spaced relation thereto, an annular-baffie mountedaroundithe bottom end of saidicentral tubular bailie, also spaced from anodes spacedin a annular baffle, and a liquid-cooled jacketior said central tubularbafile;

over the cathode in spaced relation thereto, an annular bafiiefimounted'around thebottom end of said central tubular baffle,a1so spaced from the cathode, a plurality of anodes spaced in a cooled jacketior the tubular liquid-cooled jacket for the side; walls of said tank. i a

, rectifier characterizedby'a central tubular bailie circle close to the upper portion ofsaid central tubular baffle and for the most part over said annular baflie, a tubular ,bafile; an annular liquid top of said tank, and a said central tubular '4. A -multi-anode, 'metal-tank', mercury-arc over the cathode in spaced relation thereto, an annular bafilemounted around thebottom end circle close to the upper portion of said central tubular baffle andfor the" most part over said annularbaflle, a second tubular 'bafile surrounding said circle of anodes, a tubular liquid cooled jacket for said central tubular bafile, a tubular liquid-cooled jacket for said second tubular baflie; anannular liquid-cooled jacket for 3 the top of said tank, and a tubular'liquid-cooled jacket for theside walls-of said tank. r r

5. A multi-anode, metal-tank, mercury-arc rectifier characterized by a central tubular baffle 7 the cathode; a plurality over the cathode in spaced relation thereto, an:

annular bafile, mounted around: the, bottom end of said central tubular baffle, also spaced from 1 of anodes spaced in a circle close to the upper portion of said central tubular baffle land for the'fmost part over said annular baffle,and an individual shieldand grid for each of the anodes.

; 6. A multi-anode; metal-tank, mercury-arc rectifier comprising a main tank, an insulated, centrally disposed, cathode-receptacle in" the bottom thereof, liquid mercury in said cathodeof" the.pumps,. the. bular baffle in spa'ced .fiowing through the bafile, also spaced from r cury-vapor flow path metal-tank, mercury-arc rectifier characterized by'a centraltubular baflie liquid-cooled jacket for receptacle,

a centralvertical ftubular baffle, an!

annular baffle mounted" around said central. tu-;

said annular baffle and at of said central tubular bafiie, and a pluralityj'of anodesspaced around said central tubular bafile" above said annular baflie.

'1; A multi-anodaj metal-tank, *mercuryearc rectifier 'comprisinga main tank, an insulated,

' centrally. disposed, cathode-receptacle inthe bot-v tom thereof, liquid mercury in saidcathode-rew;

' as, reducing the vapor-pressure of, a mercury vapor flow from, the liquid mercury centrally upward and out past the regions of the anodes, and an annular baflie surrounding the central meranddisposed between the regions of the anodes and the bottom of the tank forproviding a relatively ionized f'main arc-path,

under said annularbafiie.betweensaid regions of said anodes and said centrallyv disposed: liquid mercury. T r v L '1,

'8. A lrnulti-anode, metal tank, mercury-arc rectifier characterized by a'central tubular baffle over the cathode inspaced relation thereto, an annularbafile mounted arourldthe bottom end of said central tubular baffle, alsospaced from the cathode, aplurality'of anodes-spaced inv a: circle close to the upperpor'tionof I said central tubular balTle' and forthemost part :over saidannular baiile, a liquid-cooledjacket for said central tubular ballle and-for said annular battle and an aux,- iliary-tubular'bafile depending iro'm'said central liquid-cooled baffle. 1' I ,9. "A ,multi-anode, metal-tank; "mercury-arc rectifier characterized by an insulated cathode re ceptacle, a-rnercury'catho-de therein, acentral tubularbaffle' over thecathode in spaced relation thereto, an annular; baflle mounted around the bottom end of said' central tubular baffle, also spaced from the cathode, a plurality of anodes spaced in av circle close to the"upper portion of saidlcentral tubular shafts and forv the most part over said annular battle-[a second tubular'bafile surrounding said ci'roleo'fi anodes, liquid-cooled jack'ets'for the cathodefreceptacle, for an of said the side walls of the tank, and an from said therebetween, from said cathode pool, said-blastproducing means comprising a'coolerfor the outer tank walls,la cooler for the annular space surrounding said anodes, and acooler forthe central space'over said-cathodep'ool and within the liquid-cooled'baffie.

1l. A multi-anode, metal-tank, mercury-arc rectifier characterized by a" central tubular baflle relation to the liquid mer-;; jcury, a circulating liquid cooling-mediumkforleast the top portion annular space occupied by said anodes, said firsta V ,mentioned cooler being over thecathode-in spacedrelation thereto, 'an

annular baffle mounted around thebottom end of said central tubular baffle, also spaced from the cathode, a plurality of anodes spaced in a circle close to the upper portion of said central tubular bafiie and for the most part over said annular baifie, a liquid-cooled jacket for said central tubular baffle, means for substantially segregating the surface of the central portion of the mercury cathode from the annular outer portion thereof, and a heater for said segregated central portion.

12. A multi-anode, metal-tank, mercury-arc rectifier comprising a main tank, an insulated,

centrally disposed, cathode-receptacle in the bottom thereof, liquid mercury in said cathode-receptacle, a central vertical tubular baflle, an annular baffle mounted around said centraltubular bafile in spaced relation to the liquid mercury, a circulating liquid cooling-medium for said annular bafile and at least the top portion of said central tubular baffle, a plurality of anodes spaced around said central tubular baffle above said annular baiile, means for substantially segregating the surface of the central portion of the mercury cathode from the annular outer portion thereof, and a heater for said segregating central portion.

13. A multi--anode, metal-tank, mercury-arc rectifier comprising a main tank, an insulated, centrally disposed, cathode-receptacle in the bottom thereof, liquid mercury in said cathode-receptacle, a plurality of anodes spaced in a circle in said main tank, means, including a circulatingliquid cooling-means disposed in the central space of the main tank, for largely deionizing, as well as reducing the vapor-pressure of, a mercuryvapor flow from the central portion of liquid mercury centrally upward and out past the regions of the anodes, an annular baifie surrounding the central mercury-vapor flow path and disposed between the regions of the anodes and the bottom of the tank for providing a relatively ionized main arc-path under said annular baflle between said regions of said anodes and said central portion of the liquid mercury cathode, means for substantially segregating the surface of the central portion of the mercury cathode from the annular outer portion thereof, and a heater for said segregated central portion.

14. A multi-anode, metal-tank, mercury-arc rectifier having a liquid mercury cathode pool, a plurality of anodes, there being a main arcpath between the cathode and the several anodes, and means for producing a blast of relatively un-ionized, low-pressure mercury vapor blowing over the regions of said anodes and the arc-back spaces therebetween, from said cathode, said blast-producing means comprising a cooler for the outer tank walls, a cooler for the annular space surrounding said anodes, and a cooler for the central space over said cathode pool and within the annular space occupied by said anodes, means for substantially segregating'the surface of the central portion of the mercury cathode from the annular outer portion thereof, and a heater for said segregated central portion.

15. A multi-anode, metal-tank, mercury-arc rectifier having a liquid mercury cathode pool, a plurality of anodes, there being a main arc-path between the cathode and the several anodes, and means for producing a blast of relatively un-ionized, low-pressure mercury vapor blowing over the regions of said anodes and the arc-back spaces therebetween, said blast-producing means comprising a centrally disposed source of mercury vapor segregated from the active annular cathode surface of the mercury pool, and a succession of coolers in the path of the blast from said mercury-vapor source, past the region of the anodes, and on beyond to the outer walls of the tank, terminating in a cooler at said outer Walls, a cooler which is disposed in the path between said mercury-vapor source and the region of the anodes being at a temperature intermediate between the temperature of the outer-wall cooler and the higher temperature of the mercury-vapor source.

16. A multi-anode, metal-tank, mercury-arc rectifier characterized by a central, cylindrically disposed cooler, a ring of anodes surrounding the same, an intermediate, cylindrically disposed cooler surrounding the ring of anodes, a cathode cooler, a tank cooler for the side walls of the tank, a main circulating system for supplying a liquid heat-interchange medium in series through the tank cooler, cathode cooler, intermediate cooler and central cooler, in said order, starting with the tank cooler, a liquid-flow regulator for said main circulating system, and an auxiliary circulating system comprising a heater and means for circulating the liquid heat-interchange medium in series through said heater, the cathode cooler,

the intermediate cooler and the central cooler, in said order, starting with the heater.

17. A multi-anode, metal-tank, mercury-arc rectifier characterized by a central, cylindrically disposed cooler, a ring of anodes surrounding the same, an intermediate cylindrically disposed cooler surrounding the ring of anodes, a cathode cooler, a tank cooler for the side walls of the tank, a main circulating system for supplying a liquid heat-interchange medium in series through the tank cooler, cathode cooler and central cooler, in said order, starting with the tank cooler, a liquid-flow regulator for said main circulating system, an auxiliary circulating system comprising a heater and means for circulating the liquid heatinterchange medium in series through said heater, the cathode cooler and the central cooler, in said order, starting with the heater, and automatic thermally responsive regulator-means responsive to successively decreasing temperatures for shutting off the liquid-flow regulator and subsequently turning on said heater system, and vice versa for'increasing temperatures.

ALFRED L. ATHERTON.

and auxiliary circulating

Images