Apparatus for and Method of Induction-Hardening Machine Components

The present invention relates generally to the technology of induction heating and more particularly to the use of induction heating for case-hardening of machine components such as gears.

Machine components such as gears, splined shafts and sprockets are frequenly subjected to high torque loads, frictional wear and impact loading. The gears in a power transmission, for example, will encounter each of these forces during normal operation. In the typical gear production facility, the machining of gear teeth is followed by heat treatment to harden them. Heat-treating gears can involve many different types of operations, all of which have the common purpose or singular objective of producing a microstructure with certain optimum properties. The hardening process, however, often distorts the gear teeth resulting in reduced and variable quality.

In order to avoid these problems associated with conventional heat-treating and to improve the ability of the machine component (gear) to withstand the aforementioned loads and wear forces, the base metal is given a hardened outer case by selective hardening. In this manner, it is only the outer surface which is altered and the base metal retains its desirable properties such as strength and ductility.

One technique for the selective hardening of this outer case on such machine components as gears is to induction-harden the gear teeth individually. Another hardening technique which is also selective is a process referred to as selective carburizing. Single-tooth induction hardening is performed with a shaped intensifier that oscillates back and forth in the gear tooth space. This is usually done with the gear submerged in the quench. The process is relatively slow because only one gear tooth is processed at a time. Selective carburizing is most widely used and the process involves covering the surfaces to be protected against carburizing with a material that prevents the passage of active carbon during the furnace operation. The most widely used method to stop carbon activity is copper plating. A gear is copper plated on all surfaces except the teeth, then carburized. The part is then copper stripped, finish machined, re-copper plated all over, furnace-hardened, and quenched.

The difficulties and expense of the carburizing process have prompted companies to consider alternative techniques such as induction heating for selective case hardening, but to do so on a larger scale as opposed to the single-tooth method. U.S. Pat. No. 4,675,488, which issued June 23, 1987, to Mucha et al., discloses a variation on the single-tooth process described above, wherein the process involves inductively heating and then quench-hardening a few teeth at a time while the rest of the teeth are cooled for the purpose of preventing drawback of previously hardened teeth (column 1, lines 55-65). While all of the teeth are ultimately induction-hardened, the inductors are extremely complex and expensive. The Mucha et al. patent also mentions the attempt by others for several years to devise a means for induction heating the outer peripheral surfaces of gears by using an encircling inductor so that the gears can be treated by the inductor and then quench-hardened immediately thereafter in order to create the desired case hardening on the outer surface of the gear. The solution suggested by the Mucha et al. patent is to provide two induction heating coils with the workpiece located concentric in the first induction heating coil. This first coil is energized with the first alternating frequency current for a fixed period of time. Once deenergized, the workpiece experiences a time delay period and thereafter the first induction heating coil is reenergized with a second alternating frequency for another fixed period of time, though substantially less than the first period of time with the first alternating frequency. At the end of this second period of time, the workpiece is immediately transferred into the second induction heating coil in a concentric manner and experiences a second delay time. Following this step, the second induction heating coil is energized with a radio frequency current for a third time period and immediately quenching the outer surfaces by quenching liquid sprayed against the surfaces while the workpiece is in the second induction heating coil.

Several years ago, a dual-frequency arrangement for induction heating was described wherein a low-frequency current would be used for preheating the gear teeth and then a high-frequency (radio frequency) current could be used for final heating prior to quench hardening. This dual-frequency arrangement is employed to some extent by the Mucha et al. patent as just described. This dual-frequency concept was also recently described by the present inventors in their article entitled "Induction Gear Hardening by the Dual Frequency Method" which appeared in Heat Treating magazine, Volume 19, No. 6, published in June, 1987, As they explain in their article, the principle of dual-frequency heating employs both high- and low-frequency heat sources. The gear is first heated with a relatively low-frequency source (3-10 KHz), providing the energy required to preheat the mass of the gear teeth. This step is followed immediately by heating with a high-frequency source which will range from 100-300 KHz depending on the gear size and diametral pitch. The high-frequency source will rapidly final heat the entire tooth contour surface to a hardening temperature. The gear is then quenched to a desired hardness and tempered.

Dual-frequency heating is the fastest known way of heating a gear. Heating times range from 0.14 to 2.0 seconds. This compares, for example, with 4-5 minutes required for a laser to scan a gear, tooth by tooth. In dual-frequency heating, the spinning workpiece is preheated while riding on a spindle centering fixture. Then a quick "pulse" achieves optimum final heat. Next the piece indexes into a water-based quench, for a total process time of approximately 30 seconds. Dual frequency is unique among gear-hardening methods in that it allows competing specifications to coexist. That is, for a given case depth requirement and distortion limitation, with conventional hardening methods one requirement tends to consume the other. Because dual-frequency hardening puts only the necessary amount of heat into the part (2-3 times less energy than conventional induction), case depth requirements and distortion specifications can both be met, precisely.

With any induction heating process whether dual- or single-frequency, and regardless of the type of part and its material, the part characteristics dictate the optimum design of both the induction heating coil or coils and the most appropriate machine settings. Only with the properly designed coil and the appropriate machine setting is it possible to achieve the contour and case hardening specifications deemed to be most appropriate from the standpoint of wear and load resistance while still retaining overall part strength, material ductility and part specifications. A gear which is too brittle will fail prematurely, often by a tooth cracking or breaking of the gear blank body.

Traditionally, a fixed coil design has been used for a wide range of different parts and machine settings were made on a "best guess" basis by the induction machine operator. By fixing the coil, one variable is eliminated and the operator attempts to zero in on an acceptable final part by trial and error procedures. The more experienced the operator, hopefully the greater the number and variety of parts he will have experienced and to the extent that he is able to draw on that experience, he may be able to come close to an acceptable part, but only after repeated attempts.

Since this entire approach is not scientific, the best one can hope for is to reach an acceptable part but not an optimum part. This problem is magnified when applying induction heating to irregularly shaped objects such as gears. Heretofore, there has been no attempt to try an derive a set of formulae to precisely determine the most optimal coil specifications and induction machine settings for a given part and which is repeatable, part after part, regardless of the size, shape, material or other characteristics. Instead, gross parameters are selected for the coil based on general part size and then machine settings manipulated until the combination of variables comes close to something that can be accepted.

In order to avoid the uncertainty in coil specifications and machine settings, and to enable induction hardening in a precise and optimum manner, regardless of the type of machine component part or part geometry and features, the present invention provides a machine structure and a method of induction hardening using a series of formulae for establishing coil specifications and machine settings which formulae are based on component part size and features. This process or scientifically calculating the specifications for a unique coil and the machine variables (settings) based on individual part characteristics enables predictable and uniform results for the induction hardening of the part in an orderly and repeatable fashion.

Previously, any calculating which may have been done was rudimentary at best, based only on surface area and depth of penetration. The series of formulae allow the coil and machine variables to be set scientifically rather than by guesswork and the needless trial and error attepts are eliminated while at the same time improving part quality from merely an acceptable or tolerable level to an optimum level.

With the unscientific and haphazard method of prior approaches, there was no control over what variables would be altered by the machine operator and as previously mentioned, altering of one may have an effect on the other such that a suitable combination of variables might never be achieved. With the present invention, although there is some feedback and the possibility if required of adjustment of fine-tuning the settings, only one variable, power, is varied. Further, the feedback and possible need for adjustment is solely to machine component failures or machine parameters going out of tolerance, not because the series of formulae and equations (mathematical algorithm) are imprecise or too loose.

A method of induction hardening of machine components with the aid of a mathematical algorithm in order to compute induction coil specifications and induction machine settings according to one embodiment of the present invention comprises the steps of providing part specification data pertaining to the machine component, calculating induction coil specifications for the machine component according to the mathematical algorithm, determining induction machine settings according to the mathematical algorithm, constructing the specified induction coil, loading the induction coil on the induction machine, making the determined induction machine settings on the machine, heating the machine component by induction by way of the induction machine and quenching the machine component.

One object of the present invention is to provide an improved method of induction hardening of the machine components.

Related objects and advantages of the present invention will be apparent from the following description.

FIG. 1 is a block diagram of the induction machine components for use in induction hardening according to a typical embodiment of the present invention.

FIG. 2 is a diagrammatic flow chart of the main steps to be followed in the induction hardening of machine components according to a typical embodiment of the present invention.

FIG. 3 is a flow chart illustrating the more detailed steps and feedback loops associated with the method set forth in FIG. 2.

FIG. 3A is a continuation of the FIG. 3 flow chart.

FIG. 4 is a front elevation view of a low-frequency induction coil suitable for use in practicing the method of the present invention.

FIG. 4A is a top plan view of the FIG. 4 low-frequency induction coil.

FIG. 4B is an end elevation view of the FIG. 4 low-frequency induction coil.

FIG. 5 is a front elevation view of a high-frequency induction coil suitable for use in practicing the method of the present invention.

FIG. 5A is a top plan view of the FIG. 5 high-frequency induction coil.

FIG. 5B is an end elevation view of the FIG. 5 high-frequency induction coil.

FIG. 6 is a diagrammatic comparison between the conventional induction coil relative to the part and the induction coil according to a typical embodiment of the present invention.

FIG. 7 is a diagrammatic illustration of the arrangement of induction coils and workpiece comprising part of the work station.

For the purposes of promoting an understanding of the principles of the invention, reference will now be made to the embodiment illustrated in the drawings and specific language will be used to describe the same. It will nevertheless be understood that no limitation of the scope of the invention is thereby intended, such alterations and further modifications in the illustrated device, and such further applications of the principles of the invention as illustrated therein being contemplated as would normally occur to one skilled in the art to which the invention relates.

Referring to FIG. 1, there is illustrated a block diagram 10 of an induction machine for use in the induction case hardening of machine components such as gears. The work station 11 includes all part-handling components, spindles, slides, some of the electrical components such as transformers, the component part (workpiece), induction coil(s) and the quench station, as represented by block 12. As will be described hereinafter the component part to be induction hardened should dictate the particular process steps to be followed in order to optimize the microstructure of the finished, induction-hardened part. In some instances, both high (R.F.) and low (L.F.) frequencies will be applied, while for other parts only R.F. will be used. To this end, FIG. 1 diagrammatically represents the presence of both an R.F. generator 13 and an L.F. generator 14 which are coupled to the work station. The actual coupling of these R.F. and L.F. generators is to their corresponding induction coils, either an R.F. coil singularly, or both an R.F. and L.F. coil used in combination and positioned at the workstation. The control block 15 is representative of the operator controls and interface electronics with the remainder of the machine. These controls will interface both with activities at the workstation as well as with the R.F. and L.F. generators.

In the preferred embodiment, the induction machine is integrated with computer/microprocessor controls enabling the machine to be run automatically by means of machine settings and commands which are input to the microprocessor by a data entry means such as a floppy disk which is loaded into disk drive 15a which is connected to the microprocessor. With the floppy disk prepared with data as to each process time and sequence, power levels, part movement data and so forth, the selected and programmed machine process runs automatically by way of the microprocessor. Due to the critical sequence of events and the importance of precisely controlled times, precise time intervals, powder levels and temperature, computer controls fully integrated with the induction machine settings and machine variables is believed to be essential, and the only means to provide the optimum means of contour-hardening the component part once the induction coil specifications and machine settings are derived.

Referring to FIG. 2, a flow chart 20 is illustrated which discloses the sequence of main steps involve in some of the typical processes which may be employed in the induction hardening method of the present invention using the mathematical or computer algorithm represented by block 21. Initial component part data, represented by block 21A, is taken from the component part blueprint or other part specification data sheets which may exist. Alternatively, the part may be inspected and certain data derived from the physical item. The nature of the part will dictate the particular process selected, represented by block 21B. This component part data and process selection is then inputted into the mathematical (computer) algorithm which is run via computer software and used to calculate the preferred coil design parameters (block 22) and to derive the appropriate machine setting (block 23) for the induction machine which will receive the designed coil(s) (block 22a) and the component part.

In the event both L.F. and R.F. coils are going to be used, depending on the selected process sequence, then the coil design parameters calculated at block 22 will be compiled for both coils. The mathematical algorithm which is structured into computer software to be run on a PC or similar system consists of a series of formulae and equations which utilize component part data which will vary with different parts as the part details vary, thereby enabling the generation of induction coil specifications and machine settings and controls which are precisely tailored to the specific component part and which can be readily altered and reset as the selected component part changes. The machine settings are programmed onto a floppy disk (block 25) which is loaded into the microprocessor via disk drive 15a which is then input into the induction machine (block 03).

Information such as the number of teeth, diametral pitch, pressure angle and face width are just some of the component part data (for gears) which is identified from the print of specification sheet and utilized in the calculation of induction coil parameters and induction machine settings. The computer program which is used to run the component part data is set forth at the end of the specification and is identified as "Listing A." In this program, statements 140 through 280 identify the selected component part data for induction-hardening of a gear.

Obviously different component part variables would be selected for different types of parts, but this particular disclosure focuses principally on gears which would constitute the majority of machine component parts requiring case hardening. The primary input data from the part (block 21A) is in turn used to calculate related data (block 24) regarding the gear or related component part. If the induction hardening machine is equipped with NC or microprocessor type controls which are compatible with floppy disk or similar data storage means as is the case in the preferred embodiment, the derived machine settings can be placed on such a floppy disk (block 25) or on another suitable form of data entry and input into the machine (at block 30) in order to automatically set, control and adjust machine variables.

Once the component part data is derived, the calculations made, the specifically designed coil(s) fabricated and mounted onto the machine at the workstation, and the floppy disk programmed and loaded, the induction hardening process is initiated. Described in FIG. 2 are four different process sequences which have proven to be suitable depending upon particular part geometry and specific part data. Process "A" (block 26) incorporates dual-frequency (R.F. and L.F.) induction wherein the part is first subjected to a low-frequency preheat via a first coil and then a high-frequency (R.F.) final heating step via a second coil. As noted, the tempering step is also achieved by low-frequency induction. In process "B" (block 27), the preheat step and final heat step are both achieved by high-frequency induction. The preheating is achieved by pulses of the high-frequency while the final heating is achieved by application of a continuous high frequency. Process "B," as was true in Process "A," employs a low-frequency temper.

Block 28 illustrates the step for Process "C" which includes a high-frequency pulsed preheat followed by high-frequency final heating. The only difference between Process "B" and Process "C" is that Process "C" uses pulsed high-frequency induction for the tempering step. The fourth and final process selected as an example of some of the process variations which can be utilized with the present invention is set forth at block 29 and identified as Process "D." The only difference between Process "D" and Process "C" is that Process "D" incorporates an additional preheat step, prior to the high-frequency final heat. This additional step to the initial preheating step is performed by high-frequency pulsed induction. The selection of a particular process is governed by part configuration and related aspects such as diaimetral pitch for gears.

In the preferred embodiment, the component part to be induction-hardened is positioned relative to the coil(s) at workstation 11. The component part will be rotated while the coil remains stationary. It is possible to secure the component part between vertical centers (see FIG. 6) or on top or bottom mounting centers or fixtures. The part begins at a top or uppermost position and step by step is indexed downwardly and axially from the first coil, into a second coil, if used, and downwardly again to the quench location. Tempering is performed by one of the coils used for preheat or final heat. To achieve the desired induction heating, the coil must be positioned around and closely to the outer surface of the selected component part. Consequently, the entirety of the part size and contour must be considered in order to decide if the part can physically pass through the preheat coil in order to move downwardly and axially to the final heating coil or on to quench. If any portion of the part is larger than the surface to be hardened, there is no ability to have the workpiece travel through the coil from one position to another. If this part movement is restricted, then Process "C" would be selected since there is no second coil and all heating is done by a single high-frequency induction coil. It is important to recognize that axial symmetry and concentric positioning of the coils with the centers and quench location is needed to maintain precise and accurate part movement throughout the process.

Other considerations involved in the selection of a particular process include the diametral pitch of the gear and the available generator power. For example, for a gear with a diametral pitch of 4-10, Process "A" is preferred if the component part shape permits. For diametral pitch of 10.1 to 24, either process "B" or Process "C" is acceptable, again assuming the component part shape permits. For a diametral pitch of between 3 and 6, Process "D" may be used in lieu of Process "A," depending upon the available generator and its power level. Once the process is selected, a number of part variables are identified to be run via "Listing A" such as number of teeth and face width.

It is preferred to perform low-frequency tempering (Processes "A" and "B") as well as low-frequency preheating and thus processes with these settings are initially preferred. If a part is not receiving sufficient generator power because of the wattage of the generator, then the extra preheat step in Process "D" will achieve an acceptable part without the need to purchase a new, higher-wattage generator.

The low-frequency generator 14 has a frequency range of approximately 3-10 KHz and the heating generated by this low-frequency induction starts deep within the part and works outwardly toward the outer surface of the part. It is possible to select a special frequency within the 3-10 KHz range rather than use a variable range, but a fixed frequency is really only an option when there is a single component part to be hardened and not a variety of parts. The wattage of the low-frequency generator is fixed at 100 Kw using a solid state inverter.

The high-frequency (R.F.) generator 13 has a frequency range of approximately 100-300 KHz. The wattage of this high-frequency generator is a variable between 200 and 900 Kw.

Although four different process sequences have been suggested, it is possible to mix steps from the four processes in order to create some further permutation of process steps beyond what is suggested by Processes "A" through "D".

Referring to FIGS. 3 and 3A, the process steps of Process "A" in FIG. 2 incorporating feedback logic is described in detail relative to parts which are initially run according to the series of formulae of the computer algorithm and the particular process. The calculation of coil design parameters and induction machine settings for the particular part has been made by the series of formulae, the machine variables are set (block 32) and the floppy disk is loaded in the induction machine which is computer controlled or computer controllable. The part is next loaded on to the machine and the specially fabricated induction coils are positioned at the workstation.

In Process "A," the fist step is a low-frequency induction preheat at block 33. Following this step, there is a final heating stage at block 39, quench at block 44 and a low-frequency temper at block 56. The numerous blocks in broken-line form and the connecting broken-line arrows are representative of monitoring, checking and feedback functions. However, these functions are only an option, and would not be needed until after several hundred or thousand cycles are run on a new machine. Following the preheat step, the part should be at a predetermined temperature based upon the material of the part and the requisites for proper heat treatment. The part temperature is automatically checked at block 34 and depending on that temperature, three options are available as the next step. If the part temperature is low, the power to the induction coil is increased, logic path 35, and the next part is run. Similarly, if the part temperature is high, the power is reduced, logic path 36, and the next part is run. These power adjustments may be made as many times as desired (X times), until the proper power level is determined. However, if more than two or three attempts are required in order to try and achieve the proper power setting, then there is probably some problem or difficulty with the equipment and the machine should be stopped in order to check the low-frequency generator or sensor.

The third logic path 37 is followed whenever the part temperature is within tolerance to the predetermined level. That particular temperature is recorded and the part goes onto final heating (block 39).

Following the final heating step (high frequency) at block 39, the part temperature checking resumes. The same three logic paths, 40, 41, and 42, exist and the number of repeat tries is set as a variable "X." If the part temperature is low, the power is increased via logic path 40. If the temperature is too high, the power is decreased via logic path 41. In either case, another part is run through the final heating stage and the part temperature checked again until the power level is determined and the part temperature within tolerance. An acceptable part, based on part temperature, passes via logic path 42 to an intermediate stage 43 prior to quench (block 44). At stage 43, the energy monitor is checked and if any reading either above or below the selected level is found, the part serial number is recorded (block 45). All parts arriving at block 46 pass on to block 47 (either directly or via block 45) and from there to quench.

The quench step, illustrated at block 44, includes both pressure and temperature concerns. If the pressure is low following quench, the pump is checked (logic path 50) and if there is a problem at that point, the machine is stopped. If the pressure is high, the filters A and B are switched (logic path 51) so that one filter can be cleaned while the other filter is used in the process. An acceptable pressure reading at block 52 results in a temperature chedk with two adjustments for a temperature which is either high or low. An acceptable temperature reading at block 53 routes the part on to temper after the pressure and temperature readings are recorded (block 54). By providing the quench station with two filter networks which are switchable and cooperatively arranged with the pump, there is a backup if one filter becomes clogged.

The temper step (block 56) is illustrated with two feedback paths 57 and 58. The part temperature is checked at block 59 and the power is either increased or decreased accordingly and the next part run. If too many attempts are necessary to zero in on the right power level to achieve the desired part temperature, the machine is stopped and the generator or sensor checked. An acceptable temperature is recorded (block 60) and the next part is run. Related logic flow paths are followed for processes B, C and D, though with different steps at the preheat stage.

Although the foregoing description may suggest a trial and error approach, it is unlikely that the numerous feedback loops will be required for several hundred if not several thousand cycles of the machine in view of the accuracy of the derived formulae in determining the induction coil parameters and machine settings for any selected part and the repeatability of the machine performance. Only after numerous repeated cycles is there any chance that machine components may experience a degradation in performance. It is only after such numerous cycles that use of the feedback loops need be considered. Further, the monitoring and checking which is performed in the feedback loops is done automatically via computer controls and the computer disk software. Even though each differently styles part which is run via the series of formulae and equations constituting the mathematical algorithm would conceivably result in a different set of coil and machine data, the formulae are capable of treating virtually any type of part. The component part variables are selected and run through the formulae such that optimum case hardening results accompanied by the most desirable set of material properties for the base metal. Further, while there are numerous variables for the coil design and numerous machine settings, the only adjusting which is done, and this after thousands of cycles via the feedback loops, is the power. This is an adjustment which can be automatically made. The validity of the series of formulae used to calculate the coil parameters and machine settings is so high and the results so precise that all other settings are left intact and the minimal amount of tweaking is done solely by a power variation.

The series of computation formulae are arranged into a computer program (Program A) for ease of data entry and calculating. Manual calculating is of course an option by following the computer program logic and equation, but manual calculating is neither as fast nor is convenient as the use of the computer program. Descriptive text has been included as part of the program listing directly in the body of the program in order to assist in an understanding of the variables and the various calculating steps which must be followed. The flow, logic and actual equations and formulae are clearly presented and the resultant calculating based on component part parameters and features is fully described by following the program steps and data entry.

When this main program ("Listing A") is run for a particular part, two supplemental printouts can be generated. These two printouts are also provided at the end of this description and prior to the claims. The first printout is entitled "Listing B" and gives input data, gear data and coil data via three separate sublistings. For a hypothetical company and part, the main program has been run and actual data generated and is listed for illustrative purposes only. The coil data for both low-frequency (L.F.) and high-frequency (R.F.) coils is listed. This coil data is applied to the design variables which are shown for the low-frequency coil 70 (FIGS. 4, 4A and 4B) and the high-frequency coil 71 (FIGS. 5, 5A and 5B). By substituting the actually derived numerical values for these coil variables (A, B, C, etc.) as found on the respective figures there is sufficient design data and dimensional information to fabricate the required induction coil (s). The designed coils 70, 71 are then fabricated and as a result of the precise algorithm which is used, there is a unique match of the coil(s) to the specific part.

The second listing generated from the main program is entitled "Listing C" and prints all machine variables which must be set. This listing also provides step-by-step instructions as to what machine settings must be made, when and in what sequence. These settings are all done precisely and automatically by loading of the generated floppy disk into the disk drive which is connected to the microprocessor.

Another feature of the present invention is the use of an inclined or angled high-frqeuency induction coil 71 which is placed around the part so as to create the visual effect of a single spiral pitch on a cylinder. The part is rotated at approximately 600 rpm while the coil remains stationary. This coil configuration is illustrated in FIG. 6 and is diagrammatically shown relative to a conventional coil 72 on the same part 73. In the traditional or prior art approach, the induction coil is generally a straight cylinder or cylindrical sleeve which has an axial height that is approximately the same as the part area to be hardened. The part is rotated at a slower rate, approximately 50 rpm. These conditions result in poor quality case hardening which penetrates deep into the base metal of the part. In the case of gears and related tooth components, these teeth become too brittle and may break or crack. By the use of the angled induction coil 74 and a much higher RPM to the part, 600 RPM as opposed to approximately 50 in prior at devices, the case hardening is extremely uniform and truly provides a hardened case of uniform thickness over the base metal which retains its desirable strength, material ductility and part specifications. These prior art approaches and the configuration of the present invention with this angled induction coil approach are compared in FIG. 6.

Referring to FIG. 7, one arrangement for some of the key components of the present invention is illustrated in diagrammatic form. As indicated, the induction hardening begins with the workpiece 73, in the preferred embodiment a gear, mounted between vertical centers 80 and 81 (shown in broken line form). An arrow 82 denotes the direction of axial movement for the gear which is initially positioned within low-frequency induction coil 70. The next step is to pass the gear to the high-frequency induction coil 71 and from there to the quench station 83. If a low-frequency temper is called for, the gear oves axially back to the low-frequency coil. Rotary drive member 84 imparts the desired rotary motion to the gear.

The angled induction coil is best described as having an annular ring shape which is spiralled with straight and substantially parallel inner and outer walls 75 and 76, respectively.

While the invention has been illustrated and described in detail in the drawings and foregoing description, the same is to be considered as illustrative and not restrictive in character, it being understood that only the preferred embodiment has been shown and described and that all changes and modifications that come within the spirit of the invention are desired to be protected. ##SPC1##