US2205851A - Temperature cycling - Google Patents

Description

c. w. HANSELL TEMPERATURE CYCLING Filed April 1. 1938 INVENTQR. HARE/v HANSELL ATTORNEY.

Patented June 25, 1940 UNITED STATE signor to Radio Corporation of America, a

corporation of Delaware I Application April 1, 1938, Serial No. 199,422

2 Claims.

The present invention relates to circuits where extremely constant dimensions are required, such as concentric resonant'lines, and has for its primary object to relieve the internal stresses in the essential mechanical parts of said circuits, whereby substantial constancy of dimensions of the resonant line is insured.

As is well known, a concentric resonant line consists of an inner conductor and an outer conductor having such dimensions and being so coupled as to form a tuned circuit for a predetcrmined frequency. The conductors of the line have substantially uniformly distributed induct ance and capacitance, and because of their construction form a low loss circuit. Concentric resonant lines have been used wherever there is need for a tuned circuit; for example, as the frequency stabilizing element of an oscillator, as an interstage coupling element, and as a filter. For a more detailed description of the general oper-- ation of such line, reference may be had to the following publications: United States Reissue Patent No. 20,189, granted to Roosenstein December 1, 1936; "Electrical Engineering" for August, 1935, pages 852 to 857, article by Hansell; "Proceedings of the Institute of Radio Engineers" for April, 1936, article by Hansell and Carter; "Proceedings of the Institute of Radio Engineers. for November, 1931, article by Conklin, Finch and Hansel]; Electrical Engineerin for July, 1934, article by Terman; United States Patent No. 2,106,776, granted to 'Trevor et' al. February 1, 1936; and United States Patent No. 2,104,554, granted to Conklin January 4, 1938.

I have found that the dimensions of the concentric resonant line change from time to time over along period of time (usually several years), thus requiring a continual adjustment of the line circuit. These changes in dimensions are not very large in physical magnitude but they cause difficulty in radio transmitters whose frequency is determined by the physical dimensions of the line and where a frequency change of even as little as one part in a million is often a matter for serious concern. I believe this to be due to the release of internal stresses in the metallic elements of the resonant line, which exist due to Welding, brazing, soldering and machining when the line is manufactured. The release of internal stresses is often caused, or accelerated by changes in temperature, or temperature distribution, occurring during operation. These stresses in materials are described and illustrated in a manher which will assist the reader to understand them, in an article by J. P. Den Hartog which appears in Journal of the Franklin Institute" for August, 1936. His Case II Circular hot spot in plate, and his Fig. 12 represent fairly well the stresses set up during fabrication of the end plate and inner conductor assembly of a quarter wave resonant line section such as is illustrated in the attached drawing. Since most industrial materials are fabricated at relatively high temperatures and retain stresses which increase as the temperature is lowered, I propose to release the internal stresses in thematerials from which the concentric resonant line is made, and to enable said materials to attain their final dimensions in a relatively short period of time, preferably as part of the manufacturing process and before the lines are placed in service.

Briefly stated, I propose to age the materials by temperature cycling the elements at high and low temperatures, after the parts which go to make up the resonant line are constructed and assembled but before they are used. More specifically, I propose to speed up the aging process by subjecting partially fabricated parts, as well as the finally assembled whole, repeatedly to very low temperatures, provided the low temperatures are not applied too rapidly or in a manner to cause new undesired stresses at normal temperatures. By applying unequal temperatures in various portions of the mechanical parts, I can increase the existing undesired internal stresses beyond the elastic limit of the material so that, when uniform temperatures are reached again the internal stresses will be lessened. By correctly carrying out the process of temperature cycling, there results far more constant and permanent mechanical dimensions much sooner than if cycling is not used.

In the single figure of the drawing, I have shown a simplified drawing, incross section, of a quarter wave resonant line such as may be used to stabilize the frequency of a radio transmitter. This resonant line includes an outer conductor l and an inner conductor 2. Both conductors l and 2 are usually made of copper but they may also be made of other materials, such as steel, coated with a good conductor such as copper or silver.

Outer cylindrical conductor I is closed at both ends by means of circular plates 3, 4 of the same material as the cylinder. The plates and the end of the cylinder are so shaped as to assure contact where their inner surfaces come together. Tht's is accomplished by tapering back the end surfaces of the outer cylindrical conductor, and/or the surface of the plates, as shown in an 'Tlie"'erfd"'plates 3 and 4 -may enough and elastic enough to maintain continresonant frequency.

uous contact under all conditions of temperature. Preferably the screws should have the same temperature coefficient of expansion as the materials which they join together since then they will be less liable to loosening due to temperature changes.

The end plates 3 and Lshould be dished to some'extent, as indicated in the drawing, to insure that an expansion of material near the center, due to a rise in temperature there, will always buckle the plates in one direction so that a compensation can be made for the effect upon the Flat plates are undesirable because they may buckle in either direction unpredictably when the center rises in temperature due to high frequency power losses when the line is being used in a high frequency transmitter.

Attached to the center of one of the end plates is the cylindrical inner conductor 2 which is preferably a quarter wave in equivalent electrical length, at the operating frequency. This requires a physical length a little less than a quarter wave, to compensate for end effects. The ratio of diameters between the inner surface of the outer cylindrical conductor and the outer surface of the inner cylindrical conductor should preferably be about 3.6 or 3.7 to obtain an optimum ratio of circulating energy to power loss, as taught in Franklin'U. S. Patent No. 1,937,559, and in British Patent No. 284,005. At the open end of the inner cylindrical conductor there is provided a obtained by welding, brazing,

section of metal bellows I with'the aid of which an Invarrod 8 may be used to obtain an exact frequency adjustment and to compensate for temperature variations to obtain a nearly constant frequency, as taught in the references previously cited.

The inner cylindrical conductor 2 is mounted on one of the end plates 4, at its center, preferably by means of an accurately fitted mechanical joint finished with silver solder. Other joints used. During the process of soldering, brazing or welding, the center of the end plate is usually raised to a relatively high temperature which tries to expand the material against the restraining force of the outer portion of the end plate. As" a result of these forces, which can be shown to be of very great magnitude, the material near the center of the plate flows under compression and relieves at least a portion of the stress. Later, when the whole structure cools off, the center portion of the disc tends to shrink and in doing so reverses the stresses so that instead of being under compression the center portion is under great tension.

A portion of the tension stress may be removed immediately, if the material is soft enough to flow, but another large portion remains and this remaining portion may cause later dimension changes which change the resonant frequency of the line in a manner for which compensation cannot be provided by means of the Invar rod 8.

If the end plate is of a relativelysoft material,

cause more or less sudden etc. may also be.

such as copper or brass, there will be a slow flow of material at a gradually diminishing rate, extending over a long period of time, even if we could hold constant temperatures in all parts of the line. If the end plate is of harder material with a relatively sharply defined elastic limit, such as steel, the slow release of stress at constant temperature may be very small but an uncommon drop in temperature or a drop in temperature followed by a rapid increase in temperature, which results in lower temperatures at the center of the plate than at the rim, may flow of material and a jump in the resonant frequency of the line.

The purpose of my'present invention is to remove the conditions which can result in slow or tion, I may artificially cool the inner cylindrical conductor 2 and the center of the end plate 4 of the line shown in the drawing by any desired means, such as dipping in a cold liquid, or by dropping the center conductor through a hole in a refrigerated chamber. The resultant shrinking of the material at the center of the plate will increase the internal stresses and cause a strain or flow of material which will result in a reduction of stresses when the parts are brought back to a normal temperature distribution again. The effect of low temperature at the center may be increased, if at the same time we apply heating to the outer part of the end plate to increase the stresses through expansion of the outer part. Thus, one procedure which I contemplate using in the practice of my invention is the deliberate establishment of temperature distributions which are the reverse of those established by soldering or welding in order to cause flow of material with consequent lessening of internal stresses under normal conditions of temperature distribution.

Another way of applying my invention is to place the completed line, or any appropriate parts of it, within a chamber where low and high temperatures may be established rather quickly, or to place the line or its parts, in first a cold and then a warm chamber, over and over again. In principle, I subject the line to a greater range of temperatures, and to lower andhigher temperatures than those to which it will be subjected while it is in service in a radio transmitter. Furthermore, in the process, I subject the line to temperature changes which are more rapid than those to which the line will be subjected in practice. By this means I cause a release of internal stresses in advance of use in a radio transmitter so that there will be little if any further release of stresses in service and therefore little if any drift of resonant frequency with time.

It is desirable to start temperature conditioning the line by using maximum extremes of low and high temperature and then to reduce the range of temperature swing gradually toward the normal operating temperature. The cycling of temperature causes a release of stresses of a secondary order which could cause relatively small jumping and drifting of resonant frequency over a long period of operation.

aeoaeu.

As a refinement, I may apply and remove high frequency energy, or set up heating within the line to simulate the heat production in operation in a transmitter due to high frequency losses. By cycling these power losses on and off we may speed up the aging and attainment of final dimensions to a few hours, or days, whereas it would perhaps require years of time to reach final dimensions in some transmitters which are not started or stopped frequently.

It will be noted that, duringoperation, the high frequency power losses cause the inner cylindrical conductor 2 and the center of the end plate 4 to operate at atemperature which is higher than that of the outer cylindrical conductor I and the rim or outer part of the end plate 4. It is therefore desirable that the internal stresses of the end plate 4 pass through a minimum when the temperature difference between center and the rim of the end plate is about half the maximum operating value. This condition gives least probability of dimensional changes as a result of operation. It tends to be obtained automatically when the line is condesirable to hold the rim of the plate in a jig which will prevent change in rim dimensions due to the stresses set up by the heating.

It is conceivable that lines may be so constructed. and operated under such conditions, that a final equilibrium of stress and strain and final dimensions almost never are reached, or they are reached only after gross dimensional changes have taken place. Instances of continuous dimensionalchanges are common in poorly heated insulated steel plates on furnaces where the plates are subjected to large temperature differences throughout their area. My invention cannot overcome this sort of phenomenon for which poor design and gross overheating of the line in service might be the cause. In such cases improved design, usually larger line dimensions, artificial cooling, etc., is the proper remedy.

What is claimed is:

1. The method of reducing dimensional changes of a tuned circuit in the form of a metallic resonator, which includes the steps of repeatedly subjecting, artificially, the resonator as a whole to extremes of high and low temperatures beyond the limits which said resonator may meet during actual service. I

2. The process of relieving internal stresses in a concentric resonant line having an inner and an outer conductor connected together at one of their adjacent ends by an end plate, which comprises heating the center of said plate to a temperature beyond the limit which it may meet during actual service and to such an extent that stresses are developed therein which are a minimum when the temperature difference between the center and the rim of said plate is approximately half the normal value when said line is in actual service.

CLARENCE W. HANSELL.

Images