Home Radio Citation Keck, Henry Chapman (1947) Home Radio. Master's thesis, California Institute of Technology. doi:10.7907/jtz5-g327. https://resolver.caltech.edu/CaltechTHESIS:06062025-181159664 Abstract No abstract. Item Type: Thesis (Master's thesis) Subject Keywords: (Industrial Design) Degree Grantor: California Institute of Technology Division: Engineering and Applied Science Major Option: Engineering Thesis Availability: Public (worldwide access) Research Advisor(s): Unknown, Unknown Thesis Committee: Unknown, Unknown Defense Date: 1 January 1947 Record Number: CaltechTHESIS:06062025-181159664 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:06062025-181159664 DOI: 10.7907/jtz5-g327 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 17416 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 26 Jun 2025 18:17 Last Modified: 26 Jun 2025 18:38 Thesis Files PDF - Final Version See Usage Policy. 38MB Repository Staff Only: item control page

H O THESIS THE M E R A D I IN PARTIAL FULFILLMENT o:p PROFESSIONAL 0 THE REQUIREMENTS DEGREE OF INDUSTRIAL CALIFORNIA INSTITUTE OF TECHNOLOGY FOR DESIGNER AT THE PASADENA CALIFORNIA HENRY KE C K 1 4 9 7 o E C O Ji 1 E N T ij PAGE TITLE PAUE 1 I.NTROLJUl_;TlOl\J REt>U L TS OJ:t' '1'.tiESlS 2 D I 4 MARKET A.ND ~01\Jt>U.MER REt>EAR~.ti SEVTION 25 I I I ~ttOl~E lY_ o~· LJE0IGN 71 DIRE~'l'lOl\J OF FU'l'URE RbS~AR~.ti Al\JD DEVELOPMl:!:1\JT 87 I N T R O D U C T I O N REQUI REMENT To design a radio, table or floor model, which has increased utility over existing radios. CHOICE OF REQUIREMENT After careful study of the radio field it was found that little attention had been directed to radios small enough for the small home, yet fine enough in desi gn and tonal quality to be used as the main family radio. For this reason it was felt that an engineering and design contribution could be made on a radio of this type. CONSECUT ION Market research was carried on to determine the probable extent of the radio market. In conjunction with this consumer research was done to ascertain the needs and desires of consumers in radio buying. Engineering research was done in acoustics to determine what had be en done and what could be done in the field of improving the acoustic qualities of a small set. In conjunction with this a study of the human ear was made to determine its peculiarities in translating air ·wave motion into sound. Production facilities of the company with which the work was carr ied on were stu died. They were considered as one of the basic challenges of the desi gn. The extent of the chang e from existing models that would be immediately practicable for the company was discussed at leng th with company executives and engineers. A working model of the proposed radio was made. Its acoust ic properties were tested and recorded according t o recognized standards . Su ggested changes on the working model were made in accordance wi th facts found. A program of additional desi gn and acoustic research was sugg ested. R E S UL T S I 0 F T HE S I S W ORK A comprehensive market and consumer surv ey of the radio field was made which is of value to the radio industry and to radio designers in building sets that are more utilitarian than those available at the present time. II . A radio different in conception and design than anything previously done was developed using a standard superheterodyne circuit. III An acoustical system was developed for this radio which it is believed is superi or to any system existing at the present time in its size class. IV A radio easily adaptable to the mass production techniques of a particular company of medium production capacity was designed. This is of particular worth in a product design since it brings the design closer to the pragmatic test of industrial design: to g ive the ultimate consumer a higher living standard as expeditiously as possible. V A direction for futu re fruitful research was indicated. MA2KET. AND CONS UMER RESEARC H SECTI:ON MA R K E T A N D C O N S U ME R R E S E AR C H C ON T E NT S PAGE MARKET RES EARCH SU1v1MARY OF' RESULTS EST I MATED POSTWAR RADI O SALES GRAPH OF U.S. TOTAL RADI O PRODUC'I1ION AND U.S. INCOME PAYMENTS 1929-1946 RADIO SATURATION HOW THE MARKET IS DIVIDED RADIO SALES HISTORY BIBLIOGRAPHY ( OF THIS SECTION) CONSUMER RESEARCH sm,rnARY OF RESULTS RESULTS OF DEALER INTERVIEWS OWNERSHI P OF RADIOS I N THE U.S. (CURTIS PUBLISHI NG COMPANY REPORT ) CONSUMER SURVEY (SYLVANIA ELECTRIC PRODUCTS CORP.) CONSU1'EER SURVEY (MCCANN.,,.ERICKSON AGENCY) CONSUMER SURVEY (ELECTRONIC CORP OF A:M.ERI CA ) BIBLI OGRAPHY ( OF THI S SECTION) 4 6 7 8 9 10 11 12 14 17 19 21 23 24 Market MA R K E T Research - 4 - R E S E A R C H S U MMA R Y OBJECT OF RESEARCH To determine the number of radios and the type of radios that can be sold in the United States in 1947. PROC EDURE Sales history of the radio market was checked as far back as 1922 for total number of sales and for total retail value of sales. Two economists were consulted to help in determining the trend of sales of the radio industry as a whole. The percentage of electric consumers in the Mountain and Pacific States as compared to the country as a whole was determined to be used as a basis for determining the approximate potential market of companies distributing in those areas. Purchasing power of the dollar, Total income payments and other government indices were checked to help ciiBtermine the relative ability of families to buy as compared to pre-war standards. RESULTS 1. 6,141,000 radio set~ will be sold in the United States in 1947. This is the estimate of the Research Department of Electrical Merchandising. It has been made by projecting the combined predictions of 82 major utilities in the U.S. to the total number of electric customers in the U.S. 2. 655,000 radio sets will be sold in the Pacific States in 1947. 184,000 radio sets will be sold in the Mountain States in 1947. These estimates are based upon the percentage- of the total number of electrical customers in the United States that these areas represent. Regional variations have been neglected. 3. Straight table sets, console combination sets and table combination sets will sell best in 1947 in that order by dollar volume. This estimate is based upon sales histories of these types in 1940, 1941,1942 and 1946. Only 79,047 straight consoles were sold in 1946. This compared to the 714,324 console combination sets sold in the same period (despite acute cabinet shortages) shows the undesireability of producing that type of set. Market Research - 5 - RESULTS (continued) 4. The radio industry is selling to a practically saturated market in this country. New sets sold are almost entirely replacements or additional radios in the home. DISCUSSION OF RESULTS Caution should be used in utilizing the statistics mentioned. Predictions of the type made in statements 1 and 2 are at best approximate even though they represent an almost complete coverage of the market. Too many variables are involved , however, to use the figures as more than rough gauges as to production. Many families in this country ape in a position to spend more money on consumer goods than before the war for two reasons: 1. Total income payments in the u.s. have risen to such an extent that they more than offset the decreased purchasing power of the dollar. 2. The number of persons employed per family has incl"'eased giving the family unit greater buying power. Since radios are on the borderline between luxury and necessity these two facts should have a healthy effect on the radio industry. Reference to the graph on page 7 shows this borderline situation clearly since in almost every instance when Total Income Payments go up, radio sales go up proportionately. On the other hand when Total Income Payments go down radio sales go down faster proportionately. A relatively unexploited market that may prove . fruitful is the market represented by the constant demand of hotels and restaurants for radios. The annual sale of radios to these sources amounts to $4,000,000 in the Unites States in a normal year. 28% of the country's hotel business is concentrated in the Mountain and Pacific states (17 .041s in the Pacific States and 10.96% in the Mountain States.) Orders from chain hotels and restaurants are of course particularly helpful. The pages following present the detailed information from which the rrresults" were deduced . Mark et Research - 6 - ESTIMATED POSTWAR RADI O SALES I N DOLLlL"RS IN THE MOUNTAIN AND PACIFIC STATES Furniture Store Radio Sales State Montana Colorado Wyoming Ne w Mexico Arizon a Utah Nevada Washing ton Oreg on California No . of Stores 76 176 46 '7 8 79 81 16 294 182 1,413 Postwar Sales in Dollars 'f. 405,000 540,000 135 , 000 1~'5 5, 000 4 05,000 540,000 not ava:Elable 1 , 755,000 1,080,000 7,560,000 ~ Department Store Radio Sales Montana Colorado Wyoming New Mexico Arizona Utah Nevada Washing ton Oreg on California 45 56 15 24 29 36 9 88 56 28 4 460,000 805,000 1,035,000 1,955,000 4 , 025,000 690,000 805,000 3,220,000 2 , 185,000 2 , 875 , 000 Note: These fi gures are part of th e results of the Radio and Television Survey published in 1946, which was made by the ma g azine Home Furnishings. Although they were not directly mentioned in the results of the Mark et Research they are includ e d since it is felt that the b r eakdowns indica t e d are valuable to manufactu rers d istribu ting in t h ese areas . ,, t ..... l/ ! •11 : • : - ,--· _· •. :· : • ; !'(rJ, ~:,.,; ! ,.__, :-·· -···f-!!.(~ I ·::TR')i:fllj : t4r11·n<'•( •. -2.1.hnq. · j__ I ·. I · , i T i .J_±._;~ J Market Resear ch - 8 - R A D I 0 s A T U R A T I O N January 194 States %Radio Total families . Owners Radio Families Mountain States Montana Idaho Wyoming Colorado New Mexico Arizona Utah Nevada L211,_ooo 148,400 141,700 73,100 332,100 133,400 169,800 166,700 45,800 88.9 92.3 92.4 91.4 91.4 73.5 82.4 95.9 89.3 l,_077,.000 137,000 131,000 66,800 303,600 98,000 139,900 159,800 40,900 Pacific States Washington Oreg on California 328631,000 667,000 412,000 2,784,000 95.5 94.5 93.7 96.0 3,!6892000 630,000 386,000 2,673,000 u.s. TOTAL FIGURES 37,600,000 90.4 33,998,000 The above figures are broken down in subtotals for Urban, Rural-Nonfarm and Rural-Farm as follows: Mountain States Rural-Nonfarm Rural-Farm 606,000 362, 0 00 243,000 93.9 85.6 81.5 569,000 310,000 198,000 Pacific States Urban Rural-Nonfarm Rural-Farm 2,711,000 787,000 365,000 96.9 92.9 90.4 2,628 , 000 731,000 330,000 23,748,000 7,733,000 6,319,000 95.2 87.4 76.2 22,597,000 6,586,000 4,815,000 Urb.a.n u.s. Total Urban Rural-Nonfarm Rural-Farm Source: Radio Families USA 1946, published by Broadcast Measurement Bureau, Inc., New York, New York Market H OW T H E January 1947 MA R K E T I S Re search - 9 - D I V I D E D Residential and Rural Electric Customers Percentage of U.S. Total Mountain States 927,349 2.99 Montana Idaho Wyoming Colorado New Mexico Arizona Utah Nevada 111,654 130,263 49,624 263,628 77,538 114,755 148,872 31,015 .36 .42 .85 .25 .37 .48 .10 Pacific States 3,293,793 10.62 Washington Oregon California 558,270 331,860 2 ,403,663 1.80 1.07 7.75 .lo Note: These figures show the relative importance of the Mountain States and the Pacific States, as compared to each other and as compared to the United States as a whole, as a potential radio market. The importance of the single state California in this market is of particular interest to manufacturers in this area. Source: Market Analysis Dept. of the magazine Electrical Merchandising. Fi gures compiled by the Edison Electric Institute and the National Electrical Manufacturers' Association. Market R A D I O S AL E S Retail sales Year 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942-X· 1943 1944 1945 1946 Research - 10 - H I S T OR Y Number of Radios Sold 4,428,000 3,827,800 3,420,000 3,000,000 3,806,000 4,084,000 6,026,800 8,248,000 8,064,000 7,100,000 10,538,000 11,150,000 13,100,000 4,400,000 -------------~ --- --------- 14,031,000 Retail Value $600,000,000 300,000,000 225,000,000 140,000,000 230,099,000 270,000,000 330,192,000 450,000,000 450,000,000 250,000,000 355,000,000 380,000,000 460,000,000 154,000,000 ------------------------------701,000,000 Note: It is of interest to note the fluctuations in the radios sold each year compared to the U.S. Total Income Payments compiled by. the Department of Commerce. The graph on page 7 shows the extreme close relationship between radio sal e s and U.S. Total Income Payments. ~:-civilian production of radios terminated April 22, 1942 by order of the War Production Board. All figures above are at retail value. The figures for the years 1935-1942 include t h e value of the tubes in t h e receivers. Totals sho¥m include exports to · other countries. Source: Compiled for the 1943 Broadcasting Yearbook by Radio Today. The figures for 1946 are from Electrical Merchandising, January 1, 1947. They are compiled from sales records of association executives and leading manufacturers. Market Research - 11 - B I B L I OGR A P HY 1. How the Market is Divided, Electrical Merchandising, January 1, 1947 2. Radio Saturation, Electrical Merchandising, January 1, 1947 3. B.MB Introduces Itself, Advertising and Selling, August 1946 4. LESTER, BERNARD :Market Research as a Basis for Effective Sales Organization, American Management Association, Marketing Series 38-49 5. Selling £nters a New Era, American Management Association, Marketing Series 57 6. Bigg.m- Farms, Wall Street Journal, February 4, 1947 7. Management Poll, Fortune, February 1947 8. Statistics, Survey of Current Business, December 1946 9. WOYTI NSKY , W.S. Relationship Between Consumers' Expenditures, Saving s and Disposabie Income, The Review of Economic Statistics, February 1946 10. 32,500,00 Radio Owning Families, Printer's Ink, May 3, 1946 11. Report on European Industry, Radio News, May 1946 12. Estimated Major Appliance Sales in 1947 by Reg ions, Electrical Merchandi sing, January 1, 1 94 7 13. Discussi on on t h e radio market with Mr. Gaffey, economist of U.S. Dept. of Foreign and Domestic Commerce, Los Angeles; Discussion on radio market with Professor Horace N. Gilbert, economist, California Institute of Technology 14. Radio Buying Habits, communication from the research dept. of t h e Curtis Publishing Company 15. Selling to Restaurants and Hotels, Ahrens Publish ing Company, November 1946 Note: This bibliography includes only references used for Market Research . Similar bibliographies will be f ound after Consumer Research and Engineering Research. Consumer Research C O N S U ME R R E S E AR C H - 12 SUMMA R Y OBJECT OF RESEARCH To find out what the public desires and what the pu blic will buy in the radio market of 1947. PROCEDURE E:xtensive reading was done in current periodicals pertaining to the radio field. Requests ~Nere sent to the companies that had made radio surveys recently to obtain more direct information. A check was made of the leading department stores and radio stores in the Los Angeles area to ascertain opinion in this area and to check the findings of the national surveys. The response of movie audiences to various types of interiors was noted. RESULTS l • . A console model that does not have a phonograph has practically no market. 2. Frequency Modulation or provision for its addition to the set is a very important feature and should be emphasized particularly in the larger models whose o-ctput and fidelity warrant its use. 3. Short wave is a feature which helps to sell radios even though it is little used, 4. Push button tuning in general is unsatisfactory since too much adjustment and servicing is required. 5. In consoles a split top or a top which opens up is very undesirable due to the inconvenience occasioned when a vase or picture must be moved every time the set is used. 6. Tone , reception (sensitivity, clarity) and cabinet styling are the most important factors in deciding the purchase of a home radio in the order :given. ( The intelligence of this choice gives some refu!B.tion to the generally held conception that the public consists only of morons . ) Consumer Research - 13 - RESULTS (con tinu ed) 7 . Known brands of radios sell best, However , among the well known nationally advertised makes there is little brand loyalty . That is, pers ons buy ing radios seldom insist upon buying the same make that they own or previously owned. DISCUS SI ON OF RESULTS On the following pag es are the sources from which the results are derived. In interpeting these results it should be constantly borne in mind that the results of any survey are based upon what people say as opposed to what they do . Therefore , complete reliance should not be placed upon statistical surveys of the public's opinion. The results should be considered more from the standpoint of indicating a trend. It will be noted that brand name is extremely important in radio sales. When nationally advertised brands are available it is very difficult to sell lesser known brands. Ther efore new names in the field should stimulate sales by: 1. Giving a better radio for the same price as nationally knovm brands. 2. Adver~ise intensively in the areas in which they distribute. 3. Design a radio which from appearance, servic ing and ease of use standpoint is superior to nationally advertised makes. In conjunction with this a reg ional sales feature such as local stations marked on the dial would be a sales stimulant. The reason for check ing the response of' movie audiences to variou s types of interiors is that the movie industry receives a constant flood of mail indicating audience reaction to the interiors in the movies shown. It has been found that interiors with many earmarks of familiarity (su ch as modified colonial style) combined wi th modern features such as larg e windows and fireplaces and simple furniture are the most popular . This reaction is an excellent means of cataloguing public taste . This means that t he public is gradually being won over to greater simplicity in desi gn . This is turn means that the potential market for radios of simple f u nctional desi gn will become increasingly imp ortant. Note: For further details of t h e consumer research work pleas e turn to t h e pag es follo wing this. Consumer Research R E S UL T S OF S P OT - 14 - C HE C K S UR VE Y PROCEDURE Survey was made by personal intePviews with sales managers and sales personnel in the radio departments of five major department stores in the Los Angeles area and five major radio stores in the same area. RESULTS 1. (Qu estion) Of the radios you sell what brand and what price in that brand do you prefer to sell the most? 60% preferred Magnavox selling at $250 to $400 40% preferred .other brands Comments: Great preference expressed for Magnavox was due mainly to excellent cabinet work, tone and styling of the Magnavox. Preferences for other brands was inconclusive. This question was ased to determine indirectly what people desire in a new radio and further to determine the dealers point of view with regard to the currently available sets. 2. (,Q,Uestion) In your .opinion who is most important in buying or directing the decision to buy a radio, men or women? 90% said women were most important 10% didn't know Comments: This is strong corroboration for the importance of paying attention to the desires of the women in the home in designing radios. 3. (Question) v\fhat features in o!Pder of prefer·ence are most desired in radios by women? by men? Women 1. ~tyling (like furniture) 2. Tone 3 . .B:ase of use Men 1. Gadgets 2. Tone 3. Styling {appearance) Comments: overwhelming opinion of sales personnel was that women lik e furniture styling in their radios that will conform somewhat to their other furnishings. Hidden knobs and grilles were also preferred by women. Althou gh price was not listed it was the opinion of those interviewed that in 1947 price would become of prime importance. Consumer Research - 15 - RESULTS (continued) 4. (Question) Vvhat materials are most preferred in radio cabinets in order of preference? First choice wood Second choice Plastics Third choice Metal Comments: Wood was first choice of all those interviewed because of its tonal qualities, tactile qualities and relative infrangibility. Plastic cabinets were disliked since once cracked a whole new cabinet was necessary. Also plastics connote cheapness and the tonal qualities in plastic sets were described as "tinny". Metal too was indesirable tonally. 5. ( Question) What are the most popular colors in the radio sets you sell? Wood F'irst choice Mahogany Second choice Walnut Third choice Blonde Plastics First choice Ivory Second choice Brown Third choice Black Comments: It is of interest that even in Southern California which is noted for its advanced ideas the old standby Mahogany still ranks .first. Ivory, a cream white is much preferred in comp.arison to dead white which looks lifeless. 6. ( Qu estion) Is there any trouble brought on in your store by problems of shipping and storage? Comments: Majority of persons interviewed were satisfied with shipping and storage of their sets. One larg e department store sales manager expressed opinion that crates strong enough to be piled one on top of the othe;rwould help his storage problem. On the other hand sets too well packed were mentioned as too expensive to unpack. Unsolicited comments of interest made by sales personnel: 1. Stromberg-Carlson 112PF console (34 11 high, 35n wide and 17 11 deep) was considered by one major department store an excellent standard console combining features everyone was looking for. It had dials hidden by door, automatic record changer that pulls out from cabinet, record cabinet, top that does not open up and is not split into two parts. Consumer Re search - 16 - RESULTS (continu ed) 2. Four of the dealers interviewed felt that higher wattag e output would be desirable in the rad ios they sold since standard test of cu stomers seems to be to turn radio on f u ll blast momentarily to see how it sounds. If set doesn't vibrate and the sound doesn't seem too distorted it's considered a good set. Note: It should be mentioned at this point that the most effective way of increasing the cost of a radio chassis is t o increase the wattage outpu t it is capable of delivering. 3. Three of the persons interviewed expressed opinion that a record cabinet as optional equipment in matching style to the radio set is a very desirable selling feature. Consumer 0 WN E R S H I P OF Radio ownership !?;L. type Without Phonograph With Phonograph 1Ul families in TJ. S. 86.6% 19.6% Research R AD I OS West (Mt. and Pacific States) 83.4% 20.6% Families planning to buy radios Without Phonograph 14.4% With Phonograph 14.5% RADIO V\iITH PHONOGRAPH All families in U.S. Planning this Purchase Time of Action Within two years after item available 60.4% Within two years after war 3.5% More than two years 16.7% Indefinite(didn't know) 19.4~ 100.0% Estimated Expenditure Less than $50 ~S50-99 100-149 150-199 200-299 300 and over Don't know - 17 - 1.1% 11.9% 14.9% 14 ..6% 1 6 • 7°i ,o 8.6% 32.2% 100.6% 12.5J& 19.0% west 53.2% 2.5'.6 25.3% 19.0% 100.0% o .o~t 15. oft 15.0% 12. 5Js 12.5% 11.2% 33. s·si 100.0% <!4P·153 • Median Price RADIO WITHOUT PHONOGRAPH All frunilies in U.S. Planning this Purchase Time of Action Within two years aft er item available Within two years 4.7% after war :More than two years 10.7% Indefinite(don't know) 24.3% 100.0% west 65.4% 6.1 % 14.2% 14.3~1a 100.0% Consumer Research - 18 - RADIO WITHOUT PHONOGRAPH (continued) All fa.-rJJ.ilies in u.s. Planhing •tnis Estimated Expenditure Purchase Less than $25 5.2% o1 6 • g JO ~6 25-49 $50-74 15.0% 3)75-99 7.5% $100-149 17.2% $150 and over 13.2% Don 1 t know 35.0% 100.0% Median Price west -rr:s% 3.9% 17.3% 3.9% l T. :5% 17.2% 32. 6fb 100. O~b $93. In reviewing these figures it will be noted that the median price for families planning to buy radios without phonographs and for families planning to buy radios with phonographs is very high. This is an indication of the optimism with which people announce plans to buy ·when the sale isn 1 t immediate. This survey was made in 1944 when the possibility of buying a radio was remote and when incomes in general were inflated. A review of the average prices paid for radios in the past shows great divergence from thes e figures; a divergence much greater than would be expected by the difference between median and average methods. Source: Research Department, The Curtis Publishing Company, Urban Ho1;1sing Surve¥, November "l946. Survey weighted to families having incomes of ;,/1)1500 to ~isooo. Approximately 70% of the interviews were made to families in this income spacing. Consumer Research - 19 - C O N S U ME R S U R V E Y ( This survey was made by the Sylvania Electric Products Company in 1945. It was made nationally by personal interviews to persons in a carefully se lected cross section of the buying public.) RESULTS 1. How much shopping does a buyer do when he buys a new radio? 70.9% bought in first store 15.8% shopped in two or more stores 6.6% received as a gift 4.2% bought second hand 2.5% bought from miscellaneous sources 2. Brand loyalty 15.4% purchased same brand as previous set 84.6% bought brand that was different from previous set Of the families with three radios in their homes: 5.8% had all the same brand 31. 2J$ two radios of the same brand 63.0% had each radio of a different brand 3. Desirability of short wave 52.0% of persons 4uestioned had short wave in their radio s e ts Of those who have short wave: 9.9% use it frequently 16.9% use it occasionally 73.2% use it seldom or never Note: On a recheck of those who have short wave it was found that the majority of people who now have short wave want it despite the fact that they don't use it. • 4 . Push-button tuning 31% of those interviewed have it 22% of those who have it report that it is unsatisfactory 5. Frequency Modulation 91.3% say they will want F.M. when they buy a new set 70.1% of t hose who want F.M. are willing to pay $5.00 extra for it. Note: The mere addition of Frequency Modulation circuits to a radio set means little unless the audio amplifiers and the loud- Consumer Research - 20 - RESUL'J.1S (cor t l nued) speaker are improved to such an extent that they can reproduce the added fidelity afforded by F.M. reception. Therefore an additional $5.00 means little. It is not enough to cover the cost of improving the rest of the radio set. Source: Communication from the Sylvania Electric Products Corp. 500 Fifth Avenue, New York 18, New York. Consumer C O N S U ME R Re search - 21 - S U R V E Y { This survey was made by the Mccann-Erickson Advertising Agency for Stromberg-Carlson Radio Corp.) RESUL'I'S 1. 45% of those interviewed planned to buy a radiophonograph combination 2. 27% expected to buy table models 3. 15% wanted radio consoles Note: Company expects this to be pared down since in actual practice it has been found that buyers paying price of console want phonograph included. 4. Portables are likely to provide strenuous competition to table models as secondary home radios. Prices 1. 30% of those interviewed expected to pay more than $150 for model of their choice. 2. 35% expected to pay iW75- $150 3. 35% expected to pay. less than $ 75 4. 2% of those interviewed o~med Stromberg-Carlson radios but 6% expressed preference for it due to advertising and quality reputation of the company. NEW COMPANY POLICY 1. New nationwide distributorship set-up of 6560 incorporated as opposed to the prewar twenty s·t ate distributorship of 2000 dealers. 2. New emphasis on lower priced models. F'ive sets will now be sold at prices ranging from $ 32.50 to $60.00 with next set at $120. Prewar price scale was one aet at $ 25, one at $35 and next set at $150. Top price now is $ 750. Source: Magazine Business Week, September 7, 1946 This survey is included specifically since it shows the direction one major company is taking. It is of interest Consumer Research - 22 - to check the prices people will pay for radios as compared to the prices mentioned by the survey of the Crowell Publishing Company. Consumer Research - 23 C O N S U ME R S UR VE Y ( This survey was made by the Electronic Corp. of America. The results are based on returns from a questionaire directed at 5000 recent* small set buyers.) RESULTS 1. Majority of persons questioned did not feel that snnll set was the ultimate set they would buy. Note: Electronic Corp. of America feels this presages growth in larger more expensive set business. However, this investigator feels that this isn't necessarily the case since almost any person buying a radio would prefer to buy a largar and better one, but cannot for reasons of price. 2. 900 of the 3000 questioned indicated intention of buying another small set within a year for another room in the house. 3. 1800 of those questioned did not consider their table model set their ultimate buy for the next two years. 4. The small sets were bought for the most part for rooms other than the living room. 5. Less than 100 of the persons questioned expressed any intention of buying a television set at the present time. 6. Frequency Modulation was desired by 32% of the persons questioned for their large living room sets. ~:-Note: Survey was made in 1946. Source: The magazine Electrical Merchandising, December 15, 1946. Consumer Research - 24 - B I B L I OGR AP HY 16. Urban Housing Survey, Curtis Publishing Company, November 1946 17. Postwar Radio and What the Public Will Buy, Advertising and Selling, May 1945 18. Five Steps for Conducting Market Research, Printer's Ink, September 13, 1946 19 . Radio Firm Broadens Market, Business Week, September 7, 1946 20. Four City Comparison of Buyer Preference, Printer's Ink, September 20, 1946 21. Los Angeles Times, Interview with Mr. Selzer of the Research Department 22. Designing the Postwar Receiver, Radio News, October 1946 23. Patterns in Selling Radio Service, Radio News, July 1946 24 . Short term Forecast ,of Hou se Beautiful, House Beautiful, December 1946 25. Vv'hy Buyers Wait for Brand name Radios when others are Plentiful, Sales Management, November 20, 1946 26. How to determine size of a survey Sample, Printer's Ink, September 6, 1946 27. Your best markets may be surprises, Advertising and Selling, November 1946 28. Radios are Furniture, House Beautiful , Nov ember 1946 29. Designing the Postwar Receiver, Radio News, October 1946 30. Attitudes of Rural People toward Radio Service, Bureau of Agricultural Economics, U.S. Dept. of Agriculture, January 1946 31. Caldwell-Clements Study Distribution and Promotion Areas and Quotas, Printer's Ink, March 29, 1946 ENGINEERING II ENGINEERING 25 - R ES E ARCH S E C T I ON PAGE THE ACOUSTIC PROBLEM 26 DISTORTION IN LOUDSPEAKERS 32 THE HUMAN EAR 36 ACOUSTIC SYSTEMS 40 LOUDSPEAKER CHAMBER 42 TEST PROCEDURE 48 PRODUCTION F'ACILITIES 64 Engineering T H E A C OU S T I C - 26 - P R O B L E M The faithful reproduction of music and speech by a radio is a question of integrating many variables in a judicious way. The most important factors control- ing the sound which is eventually heard from amplitude modulated wave forms~~ are the following: 1. The fidelity of the microphone and the transmitting circuits of the broadcasting station. In general the fidelity of the frequencies broadcast is high. 2. The band spread of the radio station. In the broadcast band of 550 to 1550 kilocycles the band spread allotted to each radio station is (with certain exceptions) 10 kilocycles. Thus, as has been shown by Fourier, any modulation about the median radio frequency of the station greater than 5 kilocycles will interfere with the radio stations which are 10 kilocycles above and 10 kilocycles below the station in question. In order to give a margin of safety to this limitation the Federal Communications Commission has specified that broadcasting station may not ~krnpl itude modulated wave forms are mentioned as distinct f rom frequency modulated waves since in the latter case there is no rEmon for imperfectly transmitted audio modulation due to station overlap. On frequency modulation the Federal Communications Commission has provided a band spread wide enough to cover the entire audio spectrum. Eng ineering - 27 - modulate their radio frequencies more than 4.9 kilocycles. With express permission of the F.c.c. a modulation of 7.5 kilocycles is sometimes permitted but this is the exception not the rule. The point of this is that 4.9 kilocycles is approximately one third of the audio spectrum. Thus automatically two thirds of the range of sounds the ear is capable of hearing are cut off. In practical terms this means that most of the overtones that give distinctness to musical instruments will be cut out and that the overtones in the human voice that give each voice its particular quality are cut out to a large extent. However, despite this great ha.Jidicap the fact that the one third of the audio spectrum that is broadcast lies in the region where most of the energy of music and the human voice is concentrated makes it possible to have reasonably faithful radio reception. 3. Atmospheric conditions caused by sun spots and lightning. The phenomenon known as "fading" is mainly due to the disturbances in the ionosphere which it is believed are caused by sunspots. Lightning of course results in severe static for stations in the broadcast band. Frequency modulation reception is not effected by atmospheric static of t his type. Engineering - 28 - Manmade conditions such as power stations nearby, X-ray machines diathermy machines, electric motors and fluorescent lamps turning on cause disturbance to radio reception. 4. The radio receiver is the next source of distortion. It is possible to achieve nearly linear amplification in modern radio sets. However, this is an expensive achievement and is seldom attained even on costly sets. On cheaper sets the main reason for distortion is cheap intermediate frequency cvil::o ana. ou tput transformer. The I.F. coils on these sets are usually undercoupled. adjustment. 'rhis obv iates precise Furthermore they have the unfortu- nate tendency of "peaking" at close to three thousand cycles. This of cou rse tends to dis- tort and overemphasize frequencies in that part of the audio spectrum. The output transformer of the smaller radio sets is invar:iroly too small. This means that its inductance is t _oo small sinc e inductance is a direct function of the number of turns. This in turn causes fur- ther distortion~:-. ~:-Note: The plate current of the final audio amplifier usually passes through ' the primary of the output transformer polarizing the iron stampings of the transformer; this reduces the magnetic permeability which in effect reduces the inductance of the transformer. This means that the bass register will be reduced if the transformer is not large enough. The reason for this is that inductive reactance is directly proportional to frequency . Engineering - 29 - 5. rhe loudspeaker and cabinet in which it is 1 enclosed is most important to faithful reproduction of sound. It will be treated separately on pages following. 6. The room in which the set is played has an important bearing on the quality of the sound heard . The reverberation period* of the room can be so long that speech will sound garbled amd music will sound fine. On the other hand if the reverberation period is too short ~ech will sound clear and crisp but music will sound "tinny". A further effect of the room on sound is that different materials have different coefficients of absorption. # This means that almost every room will have different characteristics not only in reverberation period but in the absorption and reflection of different frequencies. A good example of this is the -::•Note: The reverberation period of a room has been arbitrarily defined by Sabine as the time for the energy density in a room to fall to the minimum audible value from an initial value 10 6 times as great, i.e. a range of 60 decibels. For small rooms of medium absorption the reverberation period is: Tis time in seconds T = .050 y_ where Vis volume in cubic feet A A is total absorption in Sabins. #Note: Material.shave different average coefficients of absorption between each other and also each material has different coefficients of absorption for different frequencies. Engineering - 30 - contrast always found between singing in the shower and singing in the living room. In the former case the tile walls abs~~ b very little sound so there is a build up of intensity which gives the person singing a false sense of his vocal abilities. In the living room on the other hand, where there is greater volume and where there are many absorbing materials this is not the case. The extreme example of absorption is found in the anechoic sound chambers of Harvard, UCLA and other institiutions. In t h ese chambers you literally can hardly hear yourself talk. 7. The human ear itself is a source of sound distortion. As will be shown its characteristic at low intensity levels is to cut off low frequencies and high frequencies accentuating the middle range of 3000 - 5000 cycles where it is most sensitive. It should be mentioned that despite all these handicaps to fidelity it is poasible with fre quency modulation system and with powerfu l narrow band spread speakers properly located in a large acoustically treated room to hear a symphony with greater fidelity than the person sitting in the concert hall. This brief summary of the major factors involved in distortion of sound shows the great complexity of achieving faithful reproduction and the consequent difficulties Eng ineering - 31 - when cost and size of equ ipment are added limiting factors. After making a careful survey of the problems involved it was determined that a definite advance could be made in i mproving the acoustic qualities of a radio set through careful design of the loudspeaker chamber. has been done on large sets with some success. This However, on smaller sets it was found that little had been attempted or attained. It was thus decided work on a s maller set might prove worthwhile in this respect. The first step in this work involved the study of loudspeakers themselves. The chapter following gives a re- view of the work done in this regard. Engineering D I S T OR T I ON - 32 - I N L OUDS P E A K E R S Nonlinear distortion occurs in loudspeakers when there is a nonlinear element in the vibrating system. A de- scription of the most important non-linear elements follows: 1. The outside suspension system of the lou dspeaker is a non-linear element in a direct radiator loudspeaker. The stiffness is not constant but is a function of the amplitude; in general increasing with increasing amplitude. 29 This type of distortion may be eliminated by placing the fundamental resonant frequency of the loudspeaker at the lower limit of the reproduction range. Above the fundamental resonant frequency the velocity of the cone is not appreciabJy affected by the suspension system because the mechanical reactance due to the compliance of the suspension system is small compared t o the mechanical impedance of the rest of the system. 30 2. Inhomogeneity of the flux density through which the cone travels is anothe r souce of distortion. This can be eliminated by making the flux density associated with each tµrn independent of the amplitude. This requirement can be satisfied by making the voice coil large and slightly longer than the air gap. Engineering - 33 - 3 . Another non-linear element is the cone in the region from 100 to 1000 cycles. One effect of this is the production of subhermonics. Since the range from 100 to 200 cycles contains the most power in both speech and music 31 it is very important that distortion be reduced to a minimum in this range. Subharmonics are par- ticularly objectionable since they are noticed by the ear more readilu than are higher harmonics. The reason for this appears to be that it is more difficult to mask a low tone. This means that subharmonics of very low intensity can be easily notived. One means of reducing subharmonic distortion in the low frequencies mentioned is to increase the rigidity of the cone29. In experimental work done by RC A it was shown that by increasing the cone thickness two and one half times the rigidity factor was increased fifteen times. This reduced distort- ion but the added weight of the cone necessitated the use of a heavy voice coil (25 grams). This is about 25 times the mass of the voice coil used in console type radio loudspeakers . It is thus obvious that this method of reducing harmonics is out of the question from a cost standpoint on loudspeakers to b~ used in low priced radio sets. Engineering - 34 - 4. Distortion of the ordinary overload type is one of the most disturbing types of infidelity. 33 It occurs in small radio sets when speech and music are produced at high levels. Since a relatively small cone is driven through large amplitudes distortion occurs beyond the region in which Hooke's Law applies. Larger speakers are not subject to this trouble so much since they can produce adequate power output without large e~cursions of the cone. 5. Many small loudspeakers have unstable frequency regions. When a signal passes through these regions extraneous frequencies are produced. These frequencies are manifested as grunting sounds as the loudspeaker jumps from one stable frequency to another. 6. When the dimensions of a loudspeaker cone become comparable to a wave length of sound the system becomes directional. As the wave length becomes shorter, i.e. the frequ ency goes up the system becomes more directional. At the lower frequencies the directional effects are much less evident. This phenomenon has been explain- ed theoretically by Huyghens and has been proven experimentally. It is difficult to rectify this condition simply by redesign of the cone since the factors controlling the radiation angle are the same as those which experience has shown con- Engineering - 35 - trol the frequency range. 34 The factors which cause the pattern to be widen intrinsically limit the frequency range. Therefore cones should be designed specifically for the range of frequencies desired or external means should be used to spread the higher frequencies. Some of the factors which cause distortion in loudspeakers can be reduced in importance by proper design of· the mounting of the loudspeaker. Commercially sever- al cabinets have been designed improve speaker response. These commercial types will be briefly discussed on pages following as prologue to the acoustic design that was conceived for the purposes of the design described in this thesis. Engineering T H E HU MA N - 36 - E AR As the final transducer of wave energy into sound the ear is of fundamental importance in any work involving acoustics. For this reason the most important character- istics of the ear affecting loudspeaking systems were explored. RESPONSE FREQUENCY CHARACTERISTIC Equal loudness level curves35 show an extremely important characteristic of the human ear. Loud- ness-ii- is a function of frequency and intensity as is shown by these curves. At the threshold of hearing it has been shown experimentally that at 100 cycles a note nru.st be 38 decibels higher than a n ote of the same lou dness at 1000 cycles to be heard. FREQUENCY AND VOLUME RANGES I N SPEECH AND ll'TIJSIC The reproduction of speech with perfect fidelity re quires a fre quency range of 100 to 8000 cycles and a volume range of 40 decibels. Orchestral music on the other hand requires a frequency range of 40 to 14,000 cycles and a volume range of 70 decibels. Since the normal ear is sensitive to the frequency range 20 to 15 , 000 cycles it more * Note: Loudness of a sound is the ma gitude of the auditory sensation produced by the sound. A loudness scale must be so constructed that when loudness units {phons) are doubled the sensation will be doubled . Up to 8000 cycles the difference in loudness of pure tones is small for the same pressure. Above this frequency loudness decreases as f e equency increases . Engineering - 37 - than fulfills the needs of speech and music. of course is to be expected. This An important limit- ing factor, however, is the fact that a good radio receiver does not cover a volume range of more than 50 decibels. This means that some form of volume compression must be used. It is also in- dicative of the need of having the radio turnedlup to a given volume level in order to hear speech and music faithfully. This brings ou t t b e point that it is best to have the loudspeaker at some distance from the e.ar since the loudness level close to the speaker is higher than that in the rest of the room and is likely to be uncomfortable to the person next to the radio while it is just right for the rest of the people in the room. AUDITORY LOCALIZATION The ear can determine the direction from which a sound is coming with great accuracy. A given sound arrives at the two ears at slightly different time intervals producing a difference in phase. This effect plus the effect of the difference in intensity of the sound at the ears (due to diffraction) gives the hearing mechanism its accuracy of localization.3 7 , 38 Below 1000 cycles the difference in intensity due to diffraction is small. , At the high- er frequencies it becomes increasingly noticable. The characteristic of the ear to be a directional Engineering - 38 - collecting system is a corollary to its localization characteristics. This is fumportant to the radio designer since it means that the ear will tend to discriminate against reverberation and will be most sensitive to direct sounds from the radio speaker. It is therefore important that the speaker have a wide nngle of distribution of all frequencies so that listeners in every part of a room can hear as much direct sound as possible. MASKING The reduction of the ability of a listener to hear one sound in the presence of another is known as masking. Masking audi:grams for single frequency tones39 show the exact relationship that exists between frequency and masking quality in decibels. These audiograms show that it is always easier to mask a high tone with a low tone than the reverse. This means in practical terms that distortion in the lower frequencies will have a harsher effect on the :: ear than distortion in the high frequencies since it is more difficult to mask the low frequencies. NONLINEARITY OF THE EAR The ear has been shown to be a nonlinear system since when a:;pure tone of high enough intensity is impressed upon the ear harmonics of the original tone are heard. Furthermore when two loud tones Engineering - 39 are sounded together a group of tones is heard, These tones consist of sum and difference frequencies and harmonics. Th is non-linearity in- fer s that the elastic restoring force of the elements of the vibrating system 'iS , not proportional to the first power of the displacement. If a second power term is added the element is asymetrical the restoring force being different in magnitude for positive and ne gative displacements .* ------------------- ------ - ----•:(· Note: Assume the displacement of the non-linear element may be expressed a s follo ws: p equals actuating force y = ap + bp 2 where a equals a constant b equals a constant Suppose two harmonic forces are impressed upon the system: Total force on the nonlinear element is: p = p 0 cosw1 t +- p 0 cosw2 t Substituting this in the original equation: y: ap 0 cosw1t ~ ap 0 cosw2 t-+ ½bp 0 2cos2w1 t -t ½bp 0 2 cos2w2t -+ -+ bp 0 2 cos '(w1 2 )t -+ w bp 0 2 cos(w1 - w2 )t-+ bp 0 2 Six different frequencies will thus be _heard by the ear. They are: w1 /2n, w2 /2n etc. Engineering A C OU S T I C - 40 - S Y S T E MS There have been attempts to improve the sound output of loudspeakers through the design of the cabinet in which they are encased. Some of these have been mod- erately satisfactory. However, for application to the design discussed in this thesis they were unsatisfactory. The systems studued were the The Magic Voice 40 , The Acoustic Line 4 1, The Acoustical Labyrinth 42 , The Compound Horn Loudspeaker 4 3, The Bass Reflex System44 , The Helmholtz Resonator 45 , and The Low Frequency Horn46 . Since the defects of these systems with respect to the design in question fall into general classifications it is unnecessary to explain each design mentioned. The references indicated will suffice if the reader is interested in detailed study • . The main defects of these systems are: 1. They depend upon some form of cabinet resonance. 2. With the exception of the bass reflex system they are not amenable to the small dimensions desired in this thesis design. The most objectionable loss in fidelity from loudspe~kers is due to troubles in the low frequency region. (These troubles are explained elsewhere in this work.) To offset the loss in response which is inherent in the low frequency regions the systems mentioned depend on reinforcement of the sound waves in a given low frequency region below the re gi on where response normally would fall off. However , the disadvantages of this ap- proach are that resonance will be different for each Engineering - 41 cabinet due to changes in the structure of the cabinet and to inadequate dimensional dependability on the part of wood. This of course means expense to the manufactur- er in adjusting each cabinet to respond properly. Another unfortunate quality of resonant reinforcement is that it occurs over such a narrow frequency band. It thus has a tendency to cause a bump -ci!.n the response curve. This means that it accentuates some of the low fi'·equency notes but not others. Therefore amplitude distortion occurs. 11 hangover 11 is another undesired quality which frequently occurs in the systems mentioned. This is a tendency of the notes in the resonant region to die out too gradually causing a blurring which is especially noticable in male speech. Inherently the systems mentioned require loudspeaker chambers of large dimensions. This is true since large volumes are necessary for low frequency air resonance or for simple low frequency reinforcement. Since a loudspeaker chamber of small dimension was desired in this design, this was the final factor in discarding the known acqustic systems. Engineering L OU DS P E AK E R - 42 - C H A MB E R In order to overcome the difficulties inherent in the existing loudspeaker chambers it was decided to do two things: 1. Use a completely enclosed chamber for the loudspeaker. 2. Damp the natural resonant frequency of the cone of the loudspeaker. The reasons for choosing a completely enclosed loudspeaker chamber are the following: 1. It was most important to the design to have a small speaker chamber. The reason for this was economy in manufacture and ease of handling. A small chamber that is open at the back is subject to a variety of defects acoustically. The most important of these is the cancellation effects that occur due to the back wave from the loudspeaker. The back wave is of course 180° out of phase with the front wave. Therefore at specific frequencies, depending on the size of the chamber there will be a wave of sound from the back of the loudspeaker which will be out of phase with the wave from the front of the speaker. cause amplitude distortion, This will On the other hand there will be some frequencies where just the opposite effect will occur, i.e. there will be additive effects which will also cause amplitude distortion. For example in the chamber used the dimensions are approximately 16 inches wide by Engineering 10 inches deep by 12½ inches high. - 43 - Assuming that the back wave travels by the shortest path to the front of the speaker it will travel 32 inches. When 32 inches represent~ wave length or an odd multiple of ½ wave length of sound there will be additive effects caused by the back wave, which is in phase with the front wave under these conditions. This will cause acer- tain group of low frequencies to be accentuated. Similarly when 32 inches represent a wave length or an even multiple of a wave length there will be cancellation effects caused by the front wave and the back wave being out of phase. At the low frequencies where these effects are most noticed this means thRt frequencies in the region of 210 cycles will be overaccentuated and frequencies in the neighborhood of 420 cycles will be underaccentuated. This effect has been veri- fied experimentally by several observers. The result of t h is open back then on a small loud-:speaker chamber is a response frequency curve t h at has alternate peaks and valleys. This ef- fect is eliminated by the closed back loudspeaker . 2. Another advantage of the closed back loudspeaker is that its low frequency response does not f all off as rapidly as that of an open back loud- Eng ineering speaker. - 41: - This is due to the fact that it acts as a simple source whose power output is proportional to the square of the frequency as shown by the follow equation:32 P: 2ffpck 2A2: Average power in ergs per second p - density of the medium in grams per cubic centimeter c - velocity of sound in centimeters per second k - 2/1 1 - wave length in centimeters A - _SEo 4 TT S • Area of the surface of the source in square uentimeters E0 = Maximum velocity in centimeters per second over the surface S . On the other hand the open back speaker acts as an acoustic doublet whose power output is proportional to the fourth power of the frequency. 'rhe power output of such a source is represented by this equation:32 p = 2 npck4A2: Average power in ergs per 3 second As a consequence of this fourth power proportionality the power output of a direct radiator loudspeaker falls off rapidly with frequency when the dimensions of the baffle are small compared to the wavelength. On the other hand the back enclosed speaker has Engineering - 45 - a power output that is independent of the frequenc y provided the system is mass controlled"'~-. Thus it can be understood that the response of the back enclosed speaker will be superior to that of the open back. point is important. However, one additional That is the fundamental resonant period of the cone. It must be low e- nough to take advantage of the properties of the system mentioned. In general for small loudspeakers this is not the case. 3. Another important feature of the completely enclosed case is that the sound coming from it is less dependent on the position of the case in the room than is the sound coming from an open case. The quality of sound coming from an open backed case changes appreciably for instance when it is placed near a corner as compared to a more central location in the room. It can now be understood that the completely enclosed speaker chamber acts as an infinite baffle; however, it does not, of course, have the bulk of a large baffle or its immobility. DAMPING THE LOUDSPEAKER As has been stated previously one of the most annoying forms of distortion that occur in loudspeakers is the *Note: The systen1 is Mass Controlled if the velocity of the cone is inversely proportional to the frequency. Engineering - 46 - low frequency boom that occurs due to the large excursions the cone makes at the low fre quencies and more precisely to the fact that t h e natural resonant frequency of small loud speaker cones is usually in the important low frequency region of 100 to 120 cycles. To damp this resonance then is important if the response frequency of the system is expected to be flat in the low frequencies mentioned. This can be done by acoustical means, namely by placing damping material behind the loudspeaker. If this material is placed in close proximity to the speaker it will be most effective since there the movement of the displaced air is greatest. The determination, however, of the exact thickness and weight of material to use involves so many variables that it is best done empirically. 52 This was done as is mentioned in the test procedure discussion. In addition to the damping material behind the loudspeaker additional damping material was placed a short distance from the back of the loudspeaker chamber to damp the wave motion which comes through the hemisphere of material behind the loudspeaker cone. Engineering T E S T - 48 - P R OC E DUR E The testing of the loudspeaker and the loudspeaker chamber was carried out in three major steps: 1. Response frequency* measurements and directional measurements were made with the loudspeaker mounted in a large baffle. 2. Response frequency measurements and directional measurements were taken with the speaker mounted in the designed ch~~ber with the back and the proposed insulation arrangement removed. 3. Response frequency measurements and directional measurements were taken with the speaker mounted in the chamber with the back of the chamber enclosed and with the insulation arranged as specified in the design. The work was done on the roof of a building with the test equipment mounted at all times eight feet above the roof. This practically obviated the possibility of reflections from other structures interfering with the true patterns of the loudspeaker output. The response frequency curves plotted are comparable to each other and to the curves of other investigators working under similar test conditions . The curves definitely show the effect of changing one element at a time in the system. The reason for this was of course to determine if the stipulated approach to the problem of damping the low frequency resonant hump of the loudspeaker was correct. ANALYSIS OF RESULTS The curves of Figure 2 show the response of the loud -i:•Note: The response of a loudspeaker is a measure of the sound produced at a designated point in the medium with the electrical input , frequency and acoustic conditions specified. Engineering - 49 - speaker mounted in a baffle of dimensions eight feet by eight feet. As mounted under the test conditions mentioned above the true response of the . loudspeaker is sho~m b~ these curves. It will be noted, as predicted previously(Engineering, Distortion in Loudspeakers), that there is a sharp rise.in response at close to 100 cycles per second. Eliminating this rise is one of the major points of the chamber designed. Figure 2 shows that the loudspeaker tested has a verzpoor response characteristic compared to absolute atandards of response. The ideal speaker would have a flat response of from 30 to 15,000 cycles. However, this is too much to expect from a speaker of this size and cost. Figure 3 shows the response of the loudspeaker with the input(.4 volt) applied directly to the voice coil of the loudspeaker. The otject of this test was to determine the limiting characteristics of the speaker transformer. A comparison of this curve with the 20 volt input curve of Figure 2 shows that the speaker transformer, when not polarized by the plate current of the final output tube in the radio, does not distort the output appreciably. For this reason and for the reason that the loudspeaker was originally designed to be used with its associated transformer all tests were conducted Engineering - 50 - by applying the constant voltage inputs to the speaker transformer. Figure 4 shows the response of the loudspeaker when mounted in the chamber designed with the back of the chamber and the proposed insulation arrangement removed. It will be n oted that the greater part of the resonant hump of the Figure 2 seems to have disappeared. However, the curves of Figures 8 and 9 show the large volume of back radiation from the chamber under the conditions specified. This back radiation in the ordinary living room would be reflected from the surface toward which the back of the speaker chamber was facing causing a rise very similar to that shown in Figure 2 near 100 cycles. Note also the rise in the curve of Figure 4 at 300 cycles. Tb.is is due to the back radiation of the l oud- speaker adding to the front radiation. This phe- nomenon is mentioned in more detail in the discussion of the Loudspeaker Chamber. Figure 5 shows the resp ons e frequency characteristics of the loudspeaker chamber with the back enclosed and with the insulation placed as specified by the design. Note the complete elimination of the hump at 300 cycles. It will be noted that there is still a rise at 800 and 4000 cycles and a sharp dip at 1000 cycles. However, these variances are not as important in speech or music energy distri- Engineering bution as the frequencies corrected. - 51 - Furthermore, they are best corrected electronically not mechanically. A study of Figures 6 through 11 shows the tendency of the high frequencies to form a beam of energy rather than a uniform spread. As has been mention- ed (see Distortion in Loudspeakers) this is a result of the inherent efficiency considerations in loudspeaker design and is not to be considered a fault unique in this particular speaker. For fur- ther discussion of this beaming tendency of; loudspeakers and its elimination see Direction of Future Research. CONCLUSION 1ro TES'r PROCEDURE In the time available for this work it would be foolish to claim that all the wrinkles of the system had been worked out. However, the curves sho'W:'. the great improvement in response which occurs when using the system desi gned in this work. It is recognized that the true test of a loudspeaker is not a matter of graphs; it is a matter of listening. However, the human ear is too insensitive to recognize small improvements which indicate a direction of approach. Thus instrumentation is neces- sary. Please turn to the pages following (Figures 2 through 11) for a graphic review of the results of the test work. Engineering - 52 - ACKNOWLEDGEMENT In carrying out the testing procedure great assistance was rendered by Earle R. Bunker and Jack H. Slaton, graduate students in Electrical Engineering at the California Institute of Technology. W. H. Pickering,Ph.D., Associate Professor of Electrical Engineering at the California Institute of Technology was extremely helpful in procuring the necessary equipment for t he tests and in making suggestions for the most effective test procedure. En g ineering - 64 - PR OD U C T I ON F A C I L I T I E S The company with which the thesis work was done had t wo main produ ction factories: 1 . Cabinet construction 2 . Chassis assembly and final assembly The cabi n et plant was equipped to manufacture all types of plywood and filled wood cabinets inclu ding complete finishing operations . Space and machinery was available for very hi gh production . lities for forming plywood . However , there were no faci That type of work was done by a supplier of formed plywood parts . Thes e part s were preferred on smaller sets due to the economi es achieved by using fewer parts and by using preformed p arts wh ich speede d assembly time . Plywood had the disadvantag e of being h ard on production tools due to its resin content . Also the cost of formed plywood per piece is hi gher than a similar piece of filled wood(i . e . wood with a veneered surface). These disadvantag es precluded the u se of ply- wood on larg er sets . 1'he joints in practically all the cabinet work were g lued with a p h enolic resin . Thi s is standard practice in the mass produ ction of radio chassis . There were no electronic gl u ing facilities used for the total quick assembly of finished cabinets or subassemblies . Al l Gluing work was done by the slower hot press method . Eng ineering - 65 - The company was particulary interested in seeing a new radio developed which did not entail large set-up costs but which if sales warranted could be adapted to high production techniques readily. Chassis assembly and final assembly were done in the s ame plant. Chassis assembly was done by a group of women on a nonmoving line; each women did a standard group of operations. The movable line was not used since product~on did not warrant t he large capital outlay involved. The production of each chassis was broken down into standard groups of operations. Therefore a change to a new chassis entailed considerable expense. For this reason it was decided to use a standard chassis. This greatly increased the ease of adapting the design to the production methods of the company. If the chassis with minor changes could be adequately fitted into the design one important aspect of the industrial design ideal would be satisfied: using existing facilities and subassemblies in an inventive way to satisfy more fully the needs of the manufacturer and the ultimate consumer. Although for t he present the standard chassis was used it was suggested that future developments emphasize increasing compactness when this could be done without sacrificing ease of production and serviceability. Engineering - 66 - Specific recommendations were: 1. Lower the main tuning capacitor. This would reduce the overall height of the cabinet required to house the chassis. Electrically this would effect the set since the rotor is grounded. The stationary plates have a capacitive relationship to the rotating plates and to the chassis but the distance between the two sets of plates is so much less than the distance between stationary plates and the chassis that the capacitance and thus the reactance of stationary plates to the chassis is unimportant. 2. Eliminate the slide rule dial assembly with its troublesome pulley arrangement in favor of a round dial which would use a simple planetary gear arrangement to achieve the necessary g ear reduction to make fine tuning possible. Gears of Micarta (cloth impregnated with a phenolic resin) could be used for this purpose. 3. Turn power transformer on ' its side. stands upright. It now Another change that was suggested was made to increase the ease of using the set. It was: 1. Make the ON-OFF switch part of the volume control instead of part of t h e tone control as it is now. The quickest reaction of most people is to t u rn the dials at either end of the dial group since the central dials are usually less used. The volume control on this set is the left hand control. The tuning dial is the ri ght hand control and the tone control is the central knob. Eng ineering B - 67 - I B L I OG R A P HY This bibliography is a continuation of the bibliographies that appear after the market and consumer research sections. Many of the works included had indirect bearing on this thesis work but are included to aid future investigators in this field or related fields. 32. OLSEN, H.F. Elements of Acoustical Engineering, D. Van Nostrand Co. New York, 1940 33. Wide Range Loudspeakers, RCA Review, June 1946 No. 2 p. 155 34. Analysis of the Energy Distribution in Speech, Physical Review, March 1922 pp. 221-232 35. Subharmonics in Porced Oscillations in Dissipative Systems, Journal Acoustical Society of America, 1935 Vol. 6 No. 4 p. 227 and Vol. 7 No. l p. 64 36. The Action of A Direct Radiator Loudspeaker, Journal Acoustical Society of America, July 1944 Vol. 16 No. 1 p. 1 37. Conditions for Wide Angle Radiation from Conical Sound Radiators, Journal Acoustical Society of America, Vol. 15 No. 44 1943 38. Loudness, Its Definition, Measurement and Calcu- lation, Journal Acoustical Society of America, 1933 Vol. 5 No. 2 p. 82 39. A Self Contained Bridge for Measuring Impedances, Bell Laboratories Record, Vol. 12 No. 6 p. 314 1934 40 . The Function of Intensity and Phase in the Bi- naural Locati tm of 'I'ones, The Physical Review, 1920 Vol. 15 No. 5 p. 425 41. STEINBERG, J~C. and W.B. SNOW Physical Factors (wire transmission of symphonic music and its reproduction in auditory p~rspective.) The Bell System Technical Journal, Vol. '13 No. 2 p. 245 1934 42. American Tentative Standard Acoustical '11erminology, Journal Acoustical Society of America, 1937 Vol. 9 No. 1 Engineering - 68 - BIBLIOGRRPHY (continued) 43. CDUL'I'ON, C.V. , DICKEY , E. T. and S . V. PERRY The Magic Voice, Radio Engineering , October 1936 44. PHELPS, W. Acoustic Line Loudspeakers, Electronics , March 1940 45. OLNEY, B The Acoustical Labyrinth, Electronics, April 1937 46. OLSEN and MASSA The Compound Horn Loudspeaker, Journal Acoustical Society of America, July 1936 Vol. 8 p. 4 8 47. HOEKSTRA, C.E. Vented Speaker Enclosure, Electronics, March 1940 48. A Low Frequency Horn of ,Small Dimensions, Journa~ Acoustical Society of America , Vol. 14 p. 173 1943 49 . OLSEN and MASSA Applied Acoustics 50. PRESTON, WILLIAM The Orig in of Radio Fadeouts, The Physical Review , May 15, 1940 Vol. 57 No. 10 pp. 887- 894 51 . Technical Note No. 197, Forest Products Laboratory, U.S. Forest Service, Madison, Wisconsin 52. Conferred with Dr . Hugo Benioff, seismolog ist , Calif , Inst. of Tech. Dr. R. Vv . Leonard, physicist , U. of Cal. at L. A. Dr. Loye, Director of Acoustics Division, Electrical Research Products Corp . 6 601 Romaine Ave . Hollywood , California 53 . WOOD, ALEXANDER Acoustics, Blackie and Son Limited, London and Glascow, 1940 54 . MCLACHLAN, N.W . Elements of Loudspeak er Practice, Oxford University Press , London , 1935 55. WATSON, F.R . Sound, John Wiley and Sons, New York , 1935 Engineering - 69 - BIBLIOGRAPHY (continu ed) 56. STEWART, G. W and R. B. LINDSAY Acoustics, D. Van Nostrand Company, New York, 1930 57. RAYLEIGH, LORD rrhe Theory of Sound, MacMillan and Co. Limited, St. Martin's Street, London, 1926 (2 Volumes) 58. EASON, A.E. The Prevention of Vibration and Noise, Henry Frowde and Hodder & Stoughton, Strand w.c.2 1923 59. HELMHOLTZ, HERMANN Sensations of ~rone, Aberdeen University Press, 1895 60. DE BOER, J. The Efficiency of Loudspeakers, Philips Technical Review, Eindhoven, Holland, 1939 October pp. 301-308 61. Testing of Lou dspeakers, Philips Technical Review, Eindhoven, Holland, December 1939 pp.354-363 62. SMITH, F.L. The Relationship between the Power Output Stage and the Loudspeaker, The .L'unalgamated Wireless Technical Review, The Wireless Press, Sydney, Australia 63. SCHILLING, H.K. Acoustic Exp e riments in Fresnel Diffraction, The American Physics Teacher, American Institute of Physics, New York, October 1938, pp. 265-269 64. MASS, FRANK Acoustic Design Charts, Blakiston Co. Phila., 1942 65. HARTOG, DEN Mechanical Vibrations, McGraw-Hill, New York 1934 66. CRAllTDALL Theory of Vibrating Systems and Sound, D. Van Nostrand Co., :New York Engineering - 70 - BIBLIOGRAPHY (continued) 67. AIGNER and STRUTT The Physiological Effect of Several Sources of Sound of the Ear and Its Consequences in Architectural Acoustics, Journal Acoustical Society of AmeriC.§:.z_ January 1935, Vol. 6 68. STRUTT, M.J.O. On the Equivalent Mass of Driven Loudspeaker Cones, Philips LaboratO+Y, Eindhoven, Holland 1932. Desig n III CHO ICE OF - 71 - DES I G N The final choice of the desi g n was an attempt to inte g rate all the variables involve d in the most satisfactory way to the producer and t o t h e consumer. At the completion of the mark et and consu mer research many idea sketches were drawn with the aim being to incorporate t h e desirable features found b y the consumer surveys and to make the radio as practical as possible fr om the manu factu rer's standpoin t . These sk~hes were finally consolidated so that from them t h ree definite desi gns evolved, a table model radio, a me dium s ized _ two piece table or floor model, and a larg e t wo piece rad i o-phono g raph. 'l'hese desi gns were discu sse d with company execu tives a nd eng ineers . The larg e t wo piece radio-phon og raph was b e st li k Ed b y the company. Howev e r, it was felt t h at t h is desi gn shoµ ld be held in abeyance in favor of the medium sized two piece radio _._ Th e mEd.u m sized rad io woul d be more efficadou s as a mas s production pro du ct; tb e reason for thi s i s that due to its lower cost mOPe people cou ld. af ford to buy it a nd du e to its size it wou ld fit more easily into smaller homes , which ar e the trend in building. Furth ermore it was t h ought that the medium sized set would f i ll a g ap in t he radio i n du stry ; a set with fine enough tone a n d appearance to be placed in the livi n g r oom ye t low enough i n cost to b e purchas e d by f&.milies with moderate incomes . Desig n - 72 - In addition to this there has been little work done to improve the acoustics of radios of this size. In the larger more expensive sets where economy in production is not so important the acoustic problems pan be more easily solved. Furthermore as is explained (see Acoustic Systems) t h e very fact of the larger size aids the acoustical engineer. - 73 - TABLE MODEL RADIO FEATURES Prominent tuning knob eliminates fumbling for proper knob. Opening in back serves as handle without marring clean simplicity of the set. Eccentrically located speaker reduces tendency of the set to vibrate. - 75 - TWO PIECE TABLE OR FLOOR MODEL RADIO FEATURES: Large tuning dial, visible front and top. Speaker can be placed at some distance from the radio proper. The radio can be conveniently tuned sitting down. Versatility of the set is evident since it can be used together as shown or apart. It thus fits readily into many different environments. - 77 - TWO PI ECE RADI O-PHONO COMBI NAT I ON FEATURES Rad i o-phono unit separates from combined lou d sp eak e r a nd record cabinet enabling smaller u nit to be placed on a table beside a chair . Records weight do wn loudspeaker chambe r r e du cing tendency of chamber to vibrate at the low frequencies . Finger space under record albums facilitates t h eir removal from cabinet. Design - 79 - The two-piece idea is used for the following reasons: 1. Separation of the radio proper from the loudspeaker enables the user to have the radio controls close at hand yet suffer none of the adverse affects of having the speaker blasting into his ear. 2. The radio is more versatile in that it can be used together or a part. It can thus be adapted to the individual needs of purchasers more readily. 3. Repairs are more easily made on this set than on sets of similar qu ality since the whole radio unit can be conveniently taken to the dealer and tested. Further more it is possible to makenany repairs on the set wi thou t removing it from its cabine t due to the removable cover on the bottom of the cabinet. The attachment of the radio and loudspeaker electrically is accomplished by means of the extension cord of four wires. Two of t h ese are for the sig nal to the primary of the lou dspeaker transformer and two are us ed for the neg ative feooack circuit to the first audio amplifier as is shown in the electric circu it diag ram. An investi- gation was made of the possiblities of using a small transmitter to tune the main set as was done by Philco in one of their expensive sets. However, t h is was not used due to the difficulties which arise when th e next Desig n - 80 - door neighbor turns on his set or his electric razor. Frequently, enough signal is transmitted to turn on the remote tuned set. An additional reason for not using such a unit is that the development time which the company would be obliged to take to perfect it would make its actual use less probable. The physical attachment of the radio and loudspeaker chamber is accomplished by the combined handle, leg and frame braces of the loudspeaker chamber which control lateral motion and by the legs of the radio which fit over the loudspeaker chamber and thus control ba.ck and forth motion. Although means of clamping the two units more rigidly to g ether were considered they were not used in the design since control of the vertical relationship between the two •pieces added someth ing to the cost but little to the practicality of the design. ELECTRIC CIRCUIT DESIGN A standard superhetrodyne circu it of the company with which the work was done was used. (F'or photostat of this circuit see the following pag e.) A standard circuit was used since it can be readily adapted to the proposed two piece radio. Its use also obviates ex- pensive new produ ction and development requirements. The circuit is desi gned to drive a six, ei ght or ten inch speaker. '11 0 prop erly match the ou tput impedance of the set the speaker must have a voice coil impedance of 3.2 ohms. 'rhe power output of this set is l¼ watts. Design - 82 - The set and the chassis on which it was mounted constituted one of the basic challenges of the design since the dimensions of the chassis and the location of the components of the set on j_t were relatively fixed. CHOICE OF MATERIAL In the consumer research overwhelming preference for cabinets of wood was shown particularly for sets to be used in the living room. Therefore wood construction was of cou rse indicated. The exact type of wood was not s p ecified since the variety of wood used is so dependent on the wood available in the c u rrent "tight" wood market. For the purpose of the pilot model of this radio a s olid wood with v e neer was s uggested. Veneer construction has the advantage that cheaper woods can be u sed in the core of the piece. Furthermore veneer construction is less subject to shrinl{age cracks and to c h anges in dimension than solid wooct4 8 . In addition to this more care can be taken in t he selection of surface wood in veneered f urn iture. The manu facturer can dry veneer stock quickly and cheaply and he can utilize all of his wood to best advantage. - 83 - R ADI 0 '11 op unit of two piece radio that can be used tog eth er or apart . Radio and l oudspeaker connected electrically by means of a four wire extension cord . FEATURES Larg e easily read tuning dial with knob desi gnations embossed in it. Side openings ac t as handles and means of v entilating the set. Removable bottom p anel permits servicing with ou t removing set from cabinet . - 85 - LOUDSPEAKER CHAMBER Unit onto which t h e radio set on the p r evious pag e fits . FEATURES End pieces serve the fourfold function of stiffening the chamber, acting as legs for t h e chamber, acting as handles for the radio when it is u sed on the chamber. Speaker panel although easily removable is acou stically ti ght fit when in place. T Fr I I --- - ,-:; _ - - -·-- "~- G -- ---- - --- - - - - - ~ <~-·- ,- F·--- ---·-J - - - - - - --l -f , ----- - - - - -- ---:.·•~ I I J -- - - -- - - -- --- -- -- -- - -- - - ! ,,.__ ___ _ - - - - - - - - - --- l · ----- - [!, :-,, , - ,... r _, ; ri - '"r,.._·:z -J ,· l-' -- - - - - - ~ ~ - - - - - ~.r -il< ' T - - ---·::=-:,J l'r------~- D p r c:u~:i- ~-:. -•i.~ - '"' r~ -- £ • LJ "" 1. = .... i.c'=" , B II ~~ - -- - -I +-; - - - - - - - - - IG~ k~ = e-,l 1 I _\ l_ • · .• !:J " [.i._ 9 -4-- .._ L ; i-- :-_,- I I [ ~ ,.., " !;bt< 4 7 AL MOOEL - q - :,7 EYF'L~~u,,c_D V IEW ' --- -¢~d1 I ~ , - : - \\ - ' SPEAK!.R ; T □ i:i - C "- __________ r;r.' ' -- - --- ,- ::,~~~::,;::: e ,....._ : ;:;, r I "i i LH. tl10C PR N I: L r - -\_ --0 - , ----=----::: -- / ' J 7 ---- ?-- , I J I ' \ 1....0.!!L I I ' - ~ TYPB '"RRME V !EW [ NOTES: R5 NQ TEQ) -C - 87 - IV DI R ECTION OF FUTU R E RES EAR H A N D D E V E L O P ME N T One phase of improving the fid e lity of t h e sound output from the louspeaker has only been tou ched upon. This is a dispersing system for t h e high frequency beam which is always more directional than t h e low frequ enc y output. u. s. Patent 2,102,212 is one step in this direction but its cost and difficulty of mounting and bracing mak e it impractical for the low priced set. A better way of dispersing the hi gh fre quencies cou ld be achiev e d by a carefully desi g ned diffraction grating. 'I'he leng th, width and thickn ess of t h e slits in the g rating would have to be well calculated. Th e flat unit wou ld be ea s ily assemble d int o any c abin et. Fu rthermore it would have less tenden cy to load t k e speak er a nd to set up spurious vibra t ions than t he vertical fins u sed i n the patent mentioned. Desig n of the two p iece unit so that a record player coul d be later be bought and v. sed with t h e se t wou ld increase the utility of the desi g n . l?uthermore it would allow people to buy t he set g radually, one unit at a time.