ELECTRIC MELTING OF COPPER AND BRASS 
 
 C. A. HANSEN 
 

 
_s 
 
 cC) 
 
 A paper to be presented at the Annual : 
 Meeting of the American Institute of 1 | 
 Metals, September 24th to 27th, 1912, 
 
 Buffalo, N. Y. 
 
 (Advance Copy. Not to be published before the date of the meeting. Discussion Invited.) 
 
 o 
 i: ELECTRIC MELTING OF COPPER AND BRASS. 
 
 C. A. Hansen. 
 
 Dr. W. R. Whitney has turned over to me a request from 
 the American Institute of Metals for a paper on the above 
 subject, one that I hesitate to write upon since I have had but 
 very little experience with copper and its alloys in any type 
 of furnace. Mr. W. G. Rothwell, Superintendent of our 
 Schenectady Works Brass Foundry, who yearly turns out 
 from 3,000,000 to 4,000,000 lbs. of castings, and Mr. L. G. 
 Cooper, our Fuel Engineer, have, however, supplied me with 
 considerable information concerning the more commonly used 
 melting methods, the publication of which will undoubtedly 
 interest many. 
 
 The general purpose of this paper, then, will be to summarize 
 theoretical requirements, results obtained under ordinary and 
 extraordinary conditions in furnaces of the Rockwell, Charlier 
 and Schwartz types; results obtained in a few experiments 
 with an electric furnace; and a comparison and discussion of 
 the results. In this discussion I shall try to confine myself to 
 the few things I feel more or less competent to discuss. 
 
 THERMAL DATA. 
 
 Specific heats—Calories per gram per degree Centigrade. 
 
 Grvheriuatee. cick 0°C 0.0939 Richards’ 
 BE) Pet ete et at este, 2 os 300°C 0.09846 Naccari 
 DOPREr Sti au ty ees: 900°C 0.1259 Richards’ 
 Popper sirom. e...-  0FCto -3008C09) 104 LeVerrier” 
 ORY, Wag NE tee ee 20°C to 1300°C 0.110 Assumed 
 Zine from.......--. 300°C to 500°C 0.122 — LeVerrier’ 
 Pew dk LOI cs ks rans GA COeh LOS Goaes1 30 Assumed 
 
 7 36678 
 
Melting Points. 
 
 Goppeia ae) 2 eee ce eens 1084°C 
 LICR RANT | ORO oe ee cea ee 419°C 
 Brasses— 
 90. CopperelO: Zines. = ave ee ete er 1040°C 
 SU LC OpDeracU sine. ee ea eee 1000°C 
 JUS OD eres 0 SAincke mete ee eee. 940°C 
 60. Copper vO Fin Gra. fens eee eo 
 DU “COPper soU SAI Gute we neon reat ee 860°C 
 
 Heats of Fusion—Calories per gram. 
 
 06 6 010 @ £2 © © © 0 © © 0 6 ae « © (© © 6 © © «16 @ 6 © © © 
 
 Zine 
 
 Shepherd* 
 Shepherd*® 
 Shepherd 
 Shepherd® 
 Shepherd’ 
 
 Richards’ 
 Persons 
 
 Heat of Solution—Zine in copper—Calories per gram alloy. 
 32% copper—68% zine D2. Baker* 
 
 eoseeere ee eeeeweeeve @ 
 
 Baker found that the above ratio—CuZnz—evolved more 
 heat than any other combination and found a possible sub- 
 maximum corresponding to CuZn. I shall assume that all of 
 the zine in the brass melted combines with copper in the ratio 
 CuZnz. 
 
 Vapor Pressures. 
 
 COP Dereater a 1700°C— 0.002 atmospheres Greenwood’ 
 CoOpnperaiin eas 1900°C— 0.009 atmospheres Greenwood’ 
 Coppercatacw. 2h 2100°C— 0.022 atmospheres Greenwood’ 
 ANIC RO Lee ae 900°C— 0.732 atmospheres Barus 
 PAW ACE IN a el ere, Erte 1000°C— 2.5 atmospheres Greenwood 
 PAAUCW At Rete am 1100°C— 6 atmospheres Greenwood 
 JAN GRR ee asets 1200°C—12 atmospheres Greenwood 
 Brass— 
 76 copper—24 zine 1000°C+15°—0.29 atmospheres Hansen*® 
 76 copper—24 zine 1084°C+15°—0.66 atmospheres Hansen* 
 76 copper—24 zine 1150°C+15°—1.18 atmospheres Hansen’ 
 55 copper—45 zine 900°C+15°—0.24 atmospheres Hansen*® 
 55 copper—45 zine 950°C+15°—0.44 atmospheres Hansen*® 
 55 copper—45 zine 1000°C+15°—0.72 atmospheres Hansen’ 
 55 copper—45 zine 1100°C+15°—1.55 atmospheres Hansen’ 
 
 TLandoldt und Bornstein—Physikalische Tabellen. 
 
 2Chemiker Kalender, 1905. 
 
 ®’Jour. Phys. Chem., 8-421-1904—Approximated from Shepherd’s curve. 
 
 4Proc. Royal Soc., 68-9-1901. 
 
 5Trans, Faraday Soc., 1911. 
 
 ‘Boiling points in Arsem Vacuum Furnace in Hydrogen atmosphere (not pub- 
 lished.) From these boiling points, the calculated heat of solution of a gram 
 equivalent of zinc in copper is 
 
 (a) 5400 ealories for 24% Zn alloy. 
 
 (b) 2500 calories for 45% Zn alloy. 
 
 Baker’s value for 68% alloy was 5040 ealories. 
 
Heating Materials Used in This Paper. 
 
 Texas crude oil—7.25 lbs. per gallon. 
 
 Heat value—19,000 B. T. U. per lb. 
 Heat value—34,700,000 calories per gallon. 
 
 1 kilowatt hour—850,000 calories. 
 
 Thermal Calculations. 
 
 Heat required to bring 100 lbs. copper to 1300°C pouring 
 temperature. 
 
 PPLE OM rece ie ts acs PLS | 6,290,000 calories 
 Ve PGI gre ee ak ee os 3 1,950,000 calories 
 POG a tem crt td wise a Acs, 8,240,000 calories 
 
 Heat required to bring 100 lbs., 80 copper, 20 zine brass to 
 pouring temperature of 1100°C. 
 
 Lentil ges GOP Dele ssa... > 4. 4,310,000 calories 
 WeltinenGoppenan si. 1,560,000 calories 
 laksa niion WANG eae ote nee 1,270,000 calories 
 eee iC eee eie 4. 255,000 calories 
 
 7,395,000 calories 
 Less heat of solution of Zn 476,000 calories 
 
 OLE Me enier ita. oT tdotte, 8 evs 6,919,000 calories 
 
 Theoretical Requirements. 
 
 To bring 100 lbs. copper to 1300°C........ 0.237 gallons oil 
 To bring 100 Ibs. copper to 1300°C......... 9.70 Kw. hours 
 To bring 100 lbs. 80:20 brass to 1100°C.... 0.199 gallons oil 
 To bring 100 Ibs. 80:25 brass to 1100°C.... 8.13 Kw. hours 
 
 Oil Fired Furnaces. 
 
 Strictly comparative tests were made on Schwartz, Rock- 
 well and Charlier furnaces in which particular care was taken 
 to get the best results out of each furnace as regards metal 
 recovery and fuel consumption. 
 
Typical Mixture Melted. 
 
 200 Ibs. Copper scrap. 
 200 Ibs. Copper turnings 
 200 lbs. Brass turnings 
 300 lbs. Brass gates 
 
 8 lbs. Tin 
 
 12 lbs. Lead 
 
 80 lbs. Zine 
 
 1000 lbs. Total. 
 
 Typical Analysis of Metal Poured. 
 
 COppeine 4 tare ee 80.7 % 
 AN Cri aia th etae pe ore 16.4% 
 LUCE TGR oe Soar ra eee ae 1.2% 
 A a bores Sa aio atta Tear oe Lt% 
 UO Tiger cae ee crea, trace 
 
 Results Obtained Under Strict Test Conditions of Operation. 
 
 Furnace A Bes C 
 Number -Oteneats. POUreC esata teres D 8 
 Average weight charge, lbs............ 1000 700 
 Average time per heat minutes........ 113 64 
 Totala Metal chareereel bs = cows metder rere 5000 5650 
 Total Metal recovered? lbs:2. se esac r= 4886 5d40 
 Total svietalsloss "Liss eee ar. eee eee 114 110 
 Perycent.c Meta la1Ose ren at. ae ee eee 2.28 1.81 
 'TOLa bE bss Pele 01) Siig ec sme ence yee ates 765 760 
 Gallons-oN persGalpse metal meer een 2.02 dee: 
 
 Similar results not obtained under strictly test conditions 
 but under conditions of ordinary practice were as follows: 
 
 Furnace A B C 
 Numbersot heats spoured 5. eee ee 1 6 6 
 Average weight charge.:..:;:.......3> 1000 750 750 
 Average time per heat minutes........ 88 71 85 
 otaleMetal teharcved saps. weee see oe 17009 4487 4500 
 Total Metal recovered, Ibs......:3..... 16020 4371 4364 
 Potalaetal alosss. Hes uwi te. fom ae., aimee 989 116 136 
 Rerécents, Wietale loss: so. ieee eee 0.82 2.0 3.02 
 otal Eosmfuel soul sused.ss nwt: Ree 4216 682 696 
 
 Gallons soil perjOrlbs* metab... meee 3.46 nes 2.00 
 
“uotIoI jo esinor) ul soeuin J AW AJe1OYY $ JRA ch me) hf 
 
 
 
Concerning B Furnace—the cost of lining is about $14.00, 
 of which $10.00 is labor. Special Munroe fire brick shapes are 
 used and the life of a lining averages slightly more than 1000 
 heats, making this item about 2/10 cents per 100 lbs. metal 
 melted. One man looks after each furnace and, beyond a few 
 minutes spent each morning punching tuyeres and claying 
 up pitts in the lining, his attention is given wholly to the metal 
 he is melting. All of the copper alloys are poured into heated 
 graphite crucibles and the 70-lb. pots used last from 60 to 79 
 pours at 200 lbs. metal each pour. 
 
 Blast is furnished to seven furnaces of an aggregate capa- 
 city of 6,000 lbs. metal by means of a positive pressure blower 
 having a displacement of 19.6 cu. ft. per revolution, and 
 driven by a 35 H.P. motor at an average speed of 140 R. P. M. 
 The displacement is therefore 2740 cu. ft. of air per minute 
 at atmospheric pressure and the blast pressure is limited to 
 16 oz. per sq. in. by suitable valves and by-pass. 
 
 Theoretically, a 750 lb. heat melted with 2.00 gallons of oil 
 per 100 lbs. metal requires 1240 Ibs. air if the carbon in the 
 oil is burned to carbon monoxide and about half again as 
 much if burned to carbon dioxide. The former condition is 
 probably the one which interests us, since too strenuous en- 
 deavors to economize oil would only lead to the substitution 
 of zine for oil as fuel, and zine is still the more expensive. It 
 requires energy to compress this air to the average value used 
 of 14 oz. pressure and this energy is just as much a part of the 
 fuel bill as the oil. Theoretically it amounts to 0.825 Kw. 
 hrs. per 750 Ibs. heat, or more correctly 0.055 Kw. hrs. per 
 gallon of oil burned. Practically, since the blower is run con- 
 stantly and the air is by-passed when the furnaces cease taking 
 it, some 15 Kw. is expended during 10 hrs. each day for the 
 melting of some 20,000 lbs. of metal, 1.e., for the burning of 
 some 450 gallons of oil and, instead of the theoretical 0.055, 
 we have 0.33 Kw. hrs. per gallon of oil. 
 
 Summarizing the results and comparing them with the 
 theoretical requirements for the 80:20 brass, which we assume 
 is the equivalent of the alloy made, we get that much abused 
 and generally useless term—thermal efficiency. 
 
r 
 
 Oil Fired Furnaces Melting 80% Copper-Brass. 
 
 Test Conditions Ordinary Practice 
 
 Furnace A B Ea as B C 
 Sateoiumer C lbs. metal....2.02 1.78. .... 346 .. 1.97 2.00 
 Kw. hrs. blast per C lbs. 
 
 LN, gag ose SEG (ear ay. cr), 24 Oe COG 
 Oil equiv. of blast Kw. 
 
 Lk EE ROR ern a 0.016 .014 .... 0.028 0.016 0.016 
 Total oil or equivalent. .2.036 1.794 .... 3.488 1.986 2.016 
 Theoretical requirements. 
 
 Gallons oil per C lbs.....0.199 0.199 .... 0.199 0.199 0.199 
 Thermal efficiency %..... GeTiteiee Le Ueto a eh ee LL Ue Oaca 
 
 Furnace A was lined with a clay bonded carborun- 
 dum which at least partially explains the decided dif- 
 ference from the others in thermal efficiency. In the above 
 the blast power required is converted to oil equivalent on the 
 basis of 40.7 Kw. hrs. per gallon of oil, i. e., on the basis of 
 perfect interchangeability of theoretical heat values and not 
 on the basis of the actual amount of electrical energy that can 
 be obtained from a gallon of oil with any known engine sys- 
 tem. With the modern oil engine 12.75 Kw. hrs, per gallon 
 of similar oil represents excellent practice. 
 
 The electric furnace experiments were conducted on copper, 
 and for the sake of comparison the following oil furnace results 
 are approximated by using the comparative theoretical re- 
 quirements. for brass and copper and by assuming the same 
 furnace efficiency. The latter assumption favors the furnace 
 since the higher temperatures used in melting copper will most 
 certainly entail decreased thermal efficiency, 
 
 Oil Fired Furnaces Melting Pure Copper. 
 Test Conditions Ordinary Practice 
 
 Furnace A B Cie es) B C 
 POG CTONG YC Ure ee ee ase 6 3 O77. 11.0852... 070° 10-00-3987 
 
10 
 
 
 
 Week's Rotary Arc Furnace. 
 
Theoretical requirements. 
 Gals. oil per 100 lbs. ecop- 
 
 TES | ok ORS er D3 fen] ee 2 eo la ot 
 Actual gals. oil per 100 
 PIECE DOD TIC? © o5r.s <a'nore Be eee a mee ot Ac Om Dil feel) 
 
 Note—The gallon of oil is used instead of the calorie or the B.T.U. simply be- 
 cause it is a médre concrete unit to those outside the laboratory. 'The 100 Ib. 
 metal unit is chosen since it is quite general for cost sheets, etc., to be made 
 up on this basis. 
 
 Electric Melting of Copper. 
 
 The furnace used was one designed and built by Mr. C. A. 
 Weeks of Philadelphia, primarily for the distillation of zine, 
 and it was set up and tested for him in our laboratory in the 
 winter of 1909-1910. Figures 1 and 2 give a good idea of its 
 general appearance. It consisted of a horizontal, cylindrical 
 drum, 8 ft. in diameter by 6 it., lined for the copper melting 
 tests with a 9 in. course of Munroe brick and a 2 in. layer of 
 Dixon’s clay graphite mixture. The drum was mounted on 
 rollers which could be motor driven through an appropriate 
 ‘gear reduction. Stationary heads, water cooled, concentric 
 with the furnace axis served to pass 6” diameter graphite 
 electrodes intu the furnace interior. The heat was supplied 
 by radiation from an are and the rotation of the furnace body 
 was intended to equalize temperature of the furnace lining 
 by periodically bringing all parts. of it in contact with the 
 charge. Incidentally, thorough mixing of the charge would 
 be secured by this rotation. The copper melting experiments 
 were of purely secondary consideration to Mr. Weeks and 
 both the furnace and results obtained would be very materially 
 improved if the design were made with special reference to 
 convenience in melting copper and its alloys. 
 
 The results obtained were as follows: 
 
 a. Furnace at room temperature at start. 
 2,000 lbs. copper scrap charged. 
 1,926 lbs. copper ingots recovered. 
 74 lbs. or 3.7% not accounted for. 
 Time required, 3 hrs. 50 min. 
 
12 
 
 Voltage 110, 40 eyecle. 
 Current, 2,100 amperes. 
 Power used, 630 Kw. hrs. 
 
 b. Furnace at room temperature at start. 
 No copper charged. 
 Furnace interior heated to 1350°C. 
 Time required, 3 hrs. 
 Power used, 512 Kw. hrs. 
 
 The furnace requires 36 hrs., at an inside wall temperature of 
 1300°C. (zine experiments), for the outside surface to reach 
 constant temperature on account of the great heat storage 
 capacity of the lining material so it is not true that the dif- 
 ference in power consumptions for the two experiments above 
 detailed will give directly the amount necessary for melting 
 the copper. It will be seen that this difference is less than is 
 theoretically necessary to melt copper. It follows, however, 
 that if the furnace were operated continuously the amount of 
 energy required per heat would approach closer and closer to 
 that difference during the first 36 hours. 
 
 There are two ways of getting at the amount of power re- 
 quired to melt copper after the furnace has been warmed up 
 both of which will be given. 
 
 After the furnace had been in operation 36 hrs. with an in- 
 side temperature of 1300°C, the total loss of energy from the 
 furnace was 70 Kw.—this including direct electrical losses 
 in apparatus, losses in cooling water and losses from the 
 furnace shell. 200 Kw. can be supplied to the furnace on a 
 two ton charge of copper without seriously overheating the 
 lining, hence an efficiency of (200—70)-—+200 or 65% can be as- 
 signed by this method. At this efficiency it would require 298 
 Kw. hrs. per ton of copper or 14.9 Kw. hrs. per 100 Ibs. as 
 against the 9.70 Kw. hrs. theoretically necessary. 
 
 The other method of getting at the same result is by analogy 
 to results obtained in a steel melting furnace where the oppor- 
 tunities for heat loss were very similar in all respects to the 
 
13 
 
 opportunities for heat loss in the copper melting furnace. In 
 the latter furnace the following results—each figure the result 
 of averaging many figures—have been obtained. The steel 
 
 furnace had a nominal capacity of two tons as did the copper 
 melting furnace. 
 
 STEEL FURNACE. 
 
 Cold Furnace and Cold Scrap Charge. 
 
 1 ton charge melted, 1,480 Kw. hrs. per ton.......... 100 % 
 2 tons charge melted, 980 Kw. hrs. per ton.......... 66.2% 
 
 Hot Furnace and Cold Scrap Charge. 
 1 ton charge melted, 782 Kw. hrs. per ton 02.9% 
 
 2 tons charge melted, 570 Kw. hrs. per ton............ 38.5% 
 
 COPPER FURNACE. 
 
 Cold Furnace and Cold Scrap Charge. 
 
 1 ton charge melted, 630 Kw. hrs. per ton........... 100 % 
 Above value determined experimentally. 
 
 By analogy to steel furnace results: 
 2 tons charge melted, 417 Kw. hrs. per ton........... 66.2% 
 
 Hot Furnace and Cold Charge. 
 
 1 ton charge melted, 333 Kw. hrs. per ton........... 52.9% 
 2 ton charge melted, 243 Kw. hrs. per ton........... 38.9% 
 
 The two methods of reasoning lead then to 298 Kw. hrs. 
 and to 243 Kw. hrs. per ton of copper or efficiencies of 65% 
 and 80% respectively, an agreement that is fairly satisfactory 
 considering the shght amount of data available. 
 
 The analogy between steel furnace and copper furnace can 
 perhaps be carried further to give an idea of what can be 
 expected in foundry practice. It is often true, at least I know 
 of several specific instances, that night moulding is so ineffi- 
 cient compared with day moulding that one can practically 
 
14 
 
 double up on melting costs, fixed charges, etc., incident to pro- 
 duction on a three shift basis by confining the foundry to one 
 day shift per 24 hours and still gain in production cost. 
 
 Under such conditions it is advisable to pour metal only be- 
 tween 7 A. M. and 5 P. M. The steel furnace above referred 
 to has, by starting a furnace crew at midnight, poured two 
 ton heats at 7:30 A. M., 12 M. and 4:30 P. M. with an energy 
 consumption averaging 710 Kw. hrs. per ton. On the same 
 basis of three heats per day, and for copper the furnace would 
 only have to be operated eight or nine hours, the energy con- 
 sumption on a 12,000 lb. per day output should not exceed 315 
 Kw. hrs. per ton—the furnace being started cold each morn- 
 ing. 
 
 Brass Melting in the Electric Furnace. 
 
 If we assume the same furnace efficiency in melting brass 
 that we arrived at for copper, and that assumption is an emin- 
 ently safe one since melting brass is here the lower temperature 
 operation, we find that we should be able to melt an 80% 
 copper 20% zine brass in the Weeks furnace as follows: 
 
 (1) Cold Furnace and Cold Charge. 
 
 1 ton brass charge, 530 Kw. hrs. per ton. 
 2 tons brass charge, 350 Kw. hrs. per ton. 
 
 (2) Hot Furnace and Cold Charge. 
 
 1 ton brass charge, 280 Kw. hrs. per ton. 
 2 tons brass charge, 205 Kw. hrs. per ton. 
 
 (3) Three Heats Per Day. 
 2 ton charges, 265 Kw. hrs. per ton. 
 
 It is of general interest, although not of any particular 
 value, to show that the oil burned in an oil fired furnace would 
 if burned in a modern oil engine supply more than enough 
 electrical energy to melt the same amount of metal in the 
 electric furnace. 
 
 The theoretical equivalent of 315 Kw. hrs per ton neces- 
 sary to turn out 12,000 Ibs. copper per day is 0.385 gallons oil 
 
15 
 
 per 100 lbs. copper melted as compared with 2.14 to 4.07 oal- 
 lons used in oil fired furnaces. Practically, 1.24 gallons of oil 
 burned in a modern oil engine would be required to furnish 
 the electrical energy necessary to melt 100 Ibs. of copper under 
 the same conditions and this is still much below the oil con- 
 sumption in the oil fired furnaces. 
 
 rod 
 
 Discussion. 
 
 So far no attempt has been made to get any comparison of 
 fuel costs. With oil at 2.5e per gallon, power on a 3 heat, 
 12,000 lbs. per day output will have to sell for between 0.40c 
 and 0.82c per Kw. hr. to equal actual costs of oil and blast 
 in oil fired furnaces where energy to the blower is supplied at 
 the very nominal rate of le per Kw. hr. and where 1.78 to 
 3.46 gallons of oil are used in melting 100 lbs. of average red 
 brass. 
 
 Even the lower figure, 0.40c¢ Kw. hr. is guaranteed by some 
 of the oil engine concerns selling engines of the capacity neces- 
 sary to operate a 2-ton furnace, contingent on 2.5¢ per gallon 
 of just such an oil as was used in our Works. The higher 
 figure, 0.82c per Kw. hr. is obtainable for day loads from a 
 great many of the central power stations in the various cities 
 in this country. The electric melting of brass and copper is 
 therefore on a more attractive basis as regards energy costs 
 than is the electric melting of steel and the latter undoubtedly 
 has a wide commercial application in the making of castings. 
 
 Ag regards losses of metal, the one electric furnace test de- 
 tailed above does not compare favorably with the figures given 
 for the oil fired furnaces. The loss in the electric furnace 
 should, however, be attributed to seepage into the fresh un- 
 glazed lining of the furnace and to more or less skull being left 
 in the furnace since no provision was made for complete drain- 
 age of molten metal. In experiments with another type of 
 furnace, which was still less well adapted for copper melting, 
 there was a difference between weight of metal charged (5,000 
 Ibs.) and ingots cast of less than one per cent., and in the 
 latter case there was considerable volatilization of copper— 
 in fact all of the men in the building were seriously poisoned 
 
16 
 
 as a result of the copper fume. In the case of the Weeks 
 Furnace there was absolutely no fume and fume in an 
 electric furnace is really the only possible way of actually 
 losing metal unless slag coverings are used. No slag was used 
 in the Weeks Furnace—merely a shovel full of charcoal cover- 
 ing the metal. 
 
 Theoretically, the electric furnace has an immense advan- 
 tage over the ordinary oil fired furnace in that the electric 
 furnace may be practically sealed to prevent access of air— 
 which means that so long as the vapor pressure of the zine in 
 the charge does not exceed atmospheric pressure the maxi- 
 mum possible loss of zinc must be the furnace full of zine vapor. 
 
 At the usual pouring temperatures of a 20% zine brass each 
 cubic foot of gas saturated with zine vapor contains approx- 
 imately 0.025 Ibs. zine. In a perfectly closed electric furnace 
 of 150 cu. ft. capacity (Weeks’ Furnace capacity), heated 
 uniformly, the maximum loss due to volatilization would be 
 3.75 lbs. zine and this would be practically independent of 
 the total amount of brass in the furnace and also independent 
 of the length of time the metal were kept at that temperature. 
 With a 2-ton charge the loss would be 0.094%. Similarly 
 the loss of zine from a 40% zine brass at its usual pouring 
 temperature would be somewhere near 0.12%. The loss of 
 copper would be negligable so far as volatilization is con- 
 cerned in the Weeks or similar furnace. 
 
 On the other hand, we have seen that the fuel oil fired furnace 
 requires that some 10,000 cu. ft. of hot combustion pro- 
 ducts leave the furnace for each 100 lbs. of red brass melted. 
 If these gases left the furnace saturated with zine at the 
 final pouring temperature they would carry with them many 
 times the amount of zine originally charged. Fortunately the 
 gases do not by any means all leave the furnace saturated 
 with zine at the final pouring temperature, not even the 
 gases towards the end of the heat, since with a uniform gas 
 velocity the gas does not remain in the furnace for more 
 than a fraction of a second. But if brass is left in an oil 
 fired furnace under blast after it is ready to come out it does 
 
17 
 
 eertainly follow that the zine percentage in the melt will drop 
 off at an alarming rate. 
 
 The same arguments might be used in comparing oppor- 
 tunities for oxidation. 
 
 The losses of metal here reported for the oil fired furnaces 
 under test conditions—less than 2%—even if considered all 
 zinc—seem remarkable and are a credit to both furnaces and 
 management. It is folly to claim that the losses in ordinary 
 practice from all sources with the electric furnace would 
 average lower than this 2%, but it is certain that if equally 
 careful crews operate electric furnaces and oil fired furnaces 
 the former will have a decided advantage. 
 
 THE ELECTRIC BRASS FURNACE. 
 
 We have seen that the use of the electric furnace for melt- 
 ing brass is not at all a commercial impossibility. There 
 are, however, many view points that should be considered by 
 the designer of a furnace for this specific purpose. 
 
 In general there are four types of electric furnace, 1. e., 
 there are four rather different ways of applying electrical 
 energy to the heating of metals and some of them do at the 
 present stage seem inapplheable. 
 
 (a). The Induction Furnace. On account of the high elec- 
 trical conductivity of the molten copper alloys it has been 
 found difficult to obtain a continuots molten bath withont 
 the use of conducting crucibles and at the present time the 
 crucible manufacturers are having a lot of trouble in making 
 an article satisfactory for the purpose. 
 
 (b). The direct are furnaces, in which a slag covering is 
 essential and in which the ares are drawn to this slag, are 
 undoubtedly the simplest and therefore most attractive. 
 
 It has, however, been demonstrated in our own plant that 
 some copper is volatilized in this type of furnace, due to high 
 local temperatures—not perhaps enough to enter seriously into 
 the matter of costs but enough to endanger the health of the 
 workmen unless proper precautions are observed. Slag is 
 
18 
 
 also more or less of a nuisance in the case of the alloys con- 
 taining zine. Zine copper alloys, as has been shown, have high 
 vapor pressures. even at their melting points and the margin 
 between pouring temperature and melting point must be made 
 as small as possible to prevenx .:cessive zine losses and to 
 insure metal remaining quiet in the moulds. This condition 
 makes it imperative that the metal be either poured directly 
 from the furnace into the moulds or that a reasonable heat 
 capacity be provided in the way of hot—not merely warm— 
 ladles. In those foundries with which I am familiar crucibles 
 are provided for the transfer of hot brass to the moulds and 
 these crucibles are always heated in pit furnaces to tempera- 
 tures well above that of the metal being poured. The man- 
 agers of these foundries do not consider pouring directly from 
 the furnace into the moulds practicable. 
 
 My own experience with tapping furnaces has been unsatis- 
 factory so far as the removal of small portions of a furnace 
 charge of steel and copper are concerned. Pouring over the 
 lip or spout of a tilting furnace, when a slag covering is used, 
 is always more or less accompanied with tedious delays due 
 to the necessity of skimming the metal in the ladles or pots.* 
 If the slag is removed altogether from the furnace before 
 pouring begins then heat can no longer be applied in the direct 
 are type of furnace and the metal soon gets too cold to pour 
 decently. A satisfactory solution can undoubtedly be worked 
 out but the question of transfer of melted metal must be 
 seriously considered in any design involving this type of furn- 
 ace as applied to brass castings. Copper has been poured into 
 ingot moulds just as we ordinarily pour steel—through val- 
 vular or bottom pour ladles—and the results were quite satis- 
 factory. The ladle must, however, be heated more thoroughly 
 than is ordinarily the case for steel, otherwise nasty skulls 
 are formed. 
 
 (c) The Indirect Are Furnaces such as the Stassano Furnace 
 or such as the Weeks’ Furnace herein described, are very simple 
 and the heating is perfectly independent of either slag or 
 
 
 
 
 
 *Note—It is assumed that bottom i is i 
 f pouring from th i 
 where brass is being cast in green sand raeilds: S¢ MS ols Hint aces 
 
te 
 
 metal. This is undoubtedly a great advantage. Such a furnace 
 is however quite certain to cost more in repairs to lining, etc., 
 than the direct are furnace. 
 
 The indirect are furnace as applied to steel has also one 
 advantage (many disadvantages) over the same furnace 
 applied to brass, namely the effect of the presence of intensely 
 hot basic slag on the are itself. Such a slag promotes the 
 formation and maintenance of a steady are. Zine vapor on 
 the other hand is an extremely poor electrical conductor and 
 the are in a zine furnace or in a brass melting furnace is 
 snappy and rather unsteady. This is not at all fatal for in 
 Mr. Weeks furnace we kept an are going for 42 hours con- 
 tinuously in an atmosphere of practically pure zine vapor. The 
 electrodes were not adjusted once during this period, in fact 
 they were soldered in place, so to speak, by zine which had 
 condensed and frozen in the electrode openings so that the 
 electrodes could not be adjusted. This condition of unsteadi- 
 ness of the are can, however, be remedied to a certain extent 
 by providing suitably designed electrical apparatus. The 
 sticking of the electrodes is also mentioned merely as an inci- 
 dent, that also need occasion no serious worry. 
 
 (d). The fourth type of furnace is the resistance furnace 
 in which heat is developed in some portion of the lining, the 
 lining itself acting as resistance or some resistor being im- 
 bedded in the lining. Fitzgerald designed an ingenious fur- 
 nace with a conducting arch or roof from which heat is 
 reflected to the metal underneath and although he’ referred 
 to it as an unsuccessful furnace it seems to offer possibilities 
 for this rather low temperature class of work. A crucible fur- 
 nace has been designed in which the crucible serves as risist- 
 ance. It is said that if an ordinary graphite pot is heated and 
 the graphite allowed to burn out of the inner side of the walls 
 then the clay lining left in the pot is sufficient to keep the 
 metal from seriously short circuiting the crucible resistance. 
 If this is the case the crucible furnace should be useful, par- 
 ticularly for small batches of special metals. In general, 
 
 
 
 8Trans. Am. Electrochem Soc. XIX 273, 1911. 
 "Trans. Am, Electrochem. Soc. XX 281, 1911. 
 
an 
 
 however, I should personally prefer to work with real and 
 fairly large capacity furnaces. . 
 
 20 
 
 Finally the furnace heated partly with fuel and finally with 
 current has frequently been proposed. Many have tried to 
 apply this principle but so far as I know no attempt has been 
 very successful. A successful fuel fired furnace must be 
 designed for fuel combustion, and a successful electric furnace 
 must be designed with reference to applying electrical energy 
 simply and efficiently. The two designs have fundamentally 
 different requirements and a furnace designed for the com- 
 bination is quite certain to be both compheated and inefficient. 
 
 On the other hand, while it is perfectly practicable to trans- 
 fer 5-10 or 20 ton steel heats from a fuel fired furnace to 
 an electric furnace, it seems rather poor economy to transfer 
 small batches, and this would be even more the case with 
 copper zine alloys. Personally, I should use either a fuel fired 
 furnace or an electric furnace, not both. 
 
 Research Laboratory, 
 General Electric Co., 
 April 3, 1912. Schenectady. N. Y. 
 
 (Written discussion of this paper is invited, and may be sent to the Secretary, 
 
 W. M. Corse, care of Lumen Bearing Company, Buffalo, N. Y.)