US2134526A - Metalworking apparatus and method - Google Patents

Description

J. J. MOLAUGHLIN METALWORKING APPARATUS AND METHOD Oct. 25, 1938.

9 Sheets-Sheet 2 Filed May 15, 1936 a b [112,5 v 11.5. 1 A a .Z a .Z b f6 5 AJEZIZ' ZZQM &

K FWHHHHU A'FTORNEY Gm I Oct. 25, 1938. J. J. McLAUGHLlN 2,134,526

METALWORKING APPARATUS AND METHOD INVENTOR Jam .x/wzm mk v METALWORKING APPARATUS AND METHOD Filed May 13, 1936 9 Sheets-Sheet 4 INVENTOR Y Y kfo/m Jfl finqyh BY ATTORNEY Oct. 25, 1938.

' J. J. McLAUGHLlN 2,134,526

METALWORKING APPARATUS AND METHOD I Filed May 13, 1936 9 Sheets-Sheet 5 ATTO RN EY Oct. 25, 1938. J. J. M LAUGHLlN 2,134,526

METALWORKING APPARATUS AND METHOD Filed May 13, 1936 9 She ets-Sheet 6 BY I! Z9 29 2&2,

Oct. 25, 1938. J. J. McLAUGHLIN 2,134,526

METALWORKING APPARATUS AND METHOD Filed May 13, 1936 9 Sheets-Sheet 7 INVENTOR.

Oct. 25, 1938. J; J. MCLAUGHLI N METALWORKING APPARATUS AND METHOD Filed May 15, 1936 9 Sheets-Sheet 8 INVENTOR Jnma wm IZTEETH .fofn BY ATTORNEY Oct. 25, 1938. J. J. MCLAUGHUN 2,134,526

METALWORKING APPARATUS ANIS METHOD FiledMay 15, 1956 9 Sheets-Sheet 9 20 TEETH x INVENTOR ATTORNEY Patented @ct. 25, 193% METALWORKING APPARATUS AND METHOD John J. McLaughlin, North Tonawanda, N. 121, assignor to Buffalo Bolt Company, North Tonawanda, N. Y., a corporation of New York Application May 13, 19a, Serial No. ram

22 Claims.

My present invention concerns notching edges of ordinary stock bars to form blanks. conforming more or less roughly to the size and shape of the nut to be punched therefrom. For hex nuts,

a opposite sides of each notch form faces for two adjacent blanks, and the notches are so spaced along opposite edges of the bar that four adjacent sides of four notches correspond to four of the six sides. in Methods of notching bars for such purposes have been well known for fifty years or more, and my knowledge of the practical requirements, is partly based on long experience in quantity production of nuts in machines utilizing conmiercially approved methods of notching the bars; that is, forming the notches two at a-time in oppositely registering pairs, by laterally reciproeating punches; or by edgewise reciprocating swages. In both cases the longitudinal spacing of the pairs of opposite notches], is effected by intermittent lengthwise feeding of the bar. These reciprocatory bar-notching mechanisms, are commonly synchronizedwith other reciprocatory mechanism whereby the blanks are punched and cut, oil from the end of the bar, as fast as they are formed.

Reciprocatory mechanism for notching is necessarilyslow as well as cumbersome, but slowness is permissible because the adjacent nut forming mechanism is also reciprocatory, cumbersome and slow,

My present invention concerns forming the notches by a pair of power driven rolls having teeth with registering tips and valleys which engage the bars edgewise to form hex peaks on the edges, corresponding to the valleys, and notches corresponding to the tooth peaks. Prior art'patents show that ever since 1854, great lengths of hex-blank bar could be formed in this way, much faster than by any reciprocatory punching mechanism. Prior art patents, however, show the great advantages that would result from production of these long bar blanks, if the rolls could be designed and operated so as to insure blanks of substantially uniform length and shape; and if said shape could be made to approximate the outline of a hexagon closely enough, so that the trimmings would not involve prohibitive waste of metal. But the present-day art, prior to my present invention shows sporadic attempts resulting in successive failures.

The 1854 patent and other patents appearing at long intervals thereafter, notably in 1874, in 1880, in 1888 and in 1906, taken with the fact that since 1906, no one has succeeded, nor, so far (on. 80-30) x as I am aware, has even attempted to make bars with uniform hex peak edges by means of tooth rolls; convincinglyindicate a succession of paper patents which if tested out, necessarily resulted in failure, followed by definite abandonment.

One of the reasons for these prior art failures to solve the problem of hex-peak rolling, is that when toothed rolls are used, the same tooth and the same movements of the teeth that determine the shapes of opposite pairs of notches, determine the longitudinal spacing of said pairs, and any changes in the character of the rotor, or in the angles or, shapes of the teeth for any one purpose, such as' varying the notch shape, or for easier penetration of the metal by the tooth, or changing longitudinal feed and spacing of the notches, will cause interdependent changes in one or all of the other effects,

A serious dimculty is that all of these complicated efi'ects take place in a small, practically invisible region of the apparatus; so that it is very difficult to discover the causes of defects in the resulting blank bars, and still more diflicult to find out how to cure one defect without introducing or magnifying some other defect. Consequently, cut-and-try experiments have been the basis of all prior attempts to get desired results.

In my case, experimentshave been followed by the final practicaLtest of quantity production for commercial purposes; and having got successful results, it is possible to figure out the important factors of my success as well as the causes of prior failures.

One thing that may account for prior failures is that so fanas I can discover, relatively soft, very .hot bars were always used, because of the great strain that results from initial'engagement of the front face of a tooth so nearly fiatwise with the edge of the bar. Also at least two of the prior art patents definitely recognize that such rolling results in edgewise curvature of the bar, during the rolling operation. Obviously,

when such edgew'ise curvature is rolled into the bar, subsequent straightening necessarily involves increasing the notch angle on one edge of the bar'and decreasing the notch angle on the other edge of the bar, so that the blanks will not be symmetrical, and if the edgewise curving is not uniform, straightening will result in variably changing symmetry as well as variablechanging lengths.

As contrasted with the above, one reason for my present success in getting the blanks symmetrically shaped on opposite edges of the bar, and of substantially uniform length throughout the bar, is that I begin with cold. rolling instead of the hot rolling that characterizes theprior art.

With cold rolling, the pressures required are variable, and ultimately that the changing temperatures are an important factor in causin unequal lengths of blanks in different bars, and in different parts of the same bar. Heating increases the diameter of the rolls; and this causes increased depth of penetration of the teeth in the notches; and this causes increased endwise lengthening of the bar and of the blanks. For small nuts and cold rolling, I discovered that this could be cured by applying a cooling medium directly to the bar and the rolls, thereby keeping down temperatures and keeping them near enough normal so that the length variations were negligible. v

Practically all prior United States patents indicate thatpractical failure was partly a result of being over-influenced by the objectives and requirements for making hex peaks by means of. reciprocating swages. That is to say, undue weight was given to the fact that in swaging the parts ofthe bar on either side of the notch may be wedged apart forwardly as well as rearwardly. So, in such patents, the rolls were designed with a view to having each pair of teeth begin and complete their endwise wedging. as-free as possible from interference of or cooperation with preceding or succeeding pairs of teeth.

These objectives resulted in the use of relatively small rolls, with few teeth per roll, so that each tooth extends through nearly 'all the peripheral are that is available for working on the tangentially moving bar. Consequently, at certain times, only one pair of teeth, and, at .Other times, only one wedge surface of each tooth, is really in effective working engagement with v the metal. Consequently, longitudinal bar-feeding effects are some times almost wholly by unbalanced wedging in one direction, andat other times almost wholly by unbalanced wedging in the opposite direction. This necessarily results in unsymmetrical pressure-shaping of the peaks of the serrations; also unequal feeding effects, which permit variations in lengths of the blanks and edgewise curvature of the bar; also enormous variations in torque load, from practically no load near or just before dead center of one pair of teeth, up to maximum load when a succeeding pair of teeth comes into flatwise engagement with the edges of the bar. Also, for a given size nut, few teeth per roll necessitate proportionally smalldiameter rolls, and proportionally small, springy shafts.

As contrasted with these United States failures, one factor of my success was discovery that by using larger rolls with more teeth, it is possible to have enough teeth simultaneously engaging the metal, so that without unduly limiting endwise wedging of the metal, longitudinal feeding pressures in one direction may be at all times effectively controlled by feeding pressures in the opposite direction; also the torque load will be relatively uniform. I have since learned that this feature is characteristic of certain foreign patents which have been abandoned as failures. for'the last fifty to seventy-five years. Thus one of my discoveries is that mere large diameter rolls, and mere multiplicity of/teeth, was not the cause of failure and abandonment.

For any given size of blank, I make the rolls of sumcient diameter so that I can put not less than one and one-half, and preferably two or three teeth, in the same peripheral are that is occupied by one tooth when the rolls are designed in accordance with the more recent United States patents; and I believe that 18 teeth is. about the lower limit for reasonable effectiveness in controlling feed, torque, springiness, etc., as above described. Even at this lower limit, the desired endwise lengthening of the bar is substantially all by rearward wedging, instead of by both forward and rearward wedging.

When the roll is too large andhas too many teeth, rearward cupping-in of the metal by the forward faces of succeeding pairs of teeth, beins before much metal has been wedged rearwardly and thereafter metal displaced by teeth nearer the dead center must flow considerable distances lengthwise of the bar, or else be squeezed out edgewise, both of which involve difficulties and are objectionable for many obvious reasons. Up to the present I have found that somewhere near 26 teeth, when shaped and used in accordance with other novel principles of\my invention, give the most satisfactory all around results.

I have discovered that when rolling the bar cold, or at any temperature below extreme plasticity, simple wedging of the metal by opposite faces of the same tooth ceases as soon as the base of that tooth reaches the dead center, that is, when its tip is a half tooth in advance of dead center. At that point the forwardly wedging front face of each tooth begins to have a base-to-tip rolling action on the rear face of the bar peak in front of it. When that base-to-tip rolling is completed at dead center, the rear surface of the same tooth begins to have a reverse, tip-tobase rolling action on the front face of the bar peak behind it. Such opposite rolling actions are simultaneously effected on opposite faces of the same bar peak by opposite faces of two adjacent teeth.

One point about this is that the more teeth there are, the flatter is the curvature of said rolling action, and the nearer the face of the notch will approximate the shape of the tooth face.

- the initial tendency is for the tip of the tooth to shear the metal until the rear face rotates enough to start rearward wedging. I have discovered as a result of experiments, that by sufficiently blunting or rounding the tip where it first engages the bar, shearing tendency may be changed to smooth wedging. Preferably, the rear face of the tip is also rounded, preferably symmetrically with respect to the front face.

Both surfaces may be parts of the same cylindrical-surface, and the radius of the cylinder may be as'much as inch for a S. A. E, inch nut blank but usually inch, more or less, is preferable. Such surface is a sort of universal cam,

aisatee having the peculiar quality that nearly all of its working surface is rearwardly wedging when the tip first enters the metal, and it becomes less rearwardly wedging and more forwardly wedging, all the way to dead center. At dead center, where the greatest pressure is applied, it acts as a rotating cylindrical bearing surface to prevent the tip from digging into the bottom of the notch. 'lhis is also of importance at times when one faceof the tooth is rolling on one face of the bar notch and is not balanced by equal and opposite pressure of an adjacent tooth face.

The rounding has another very important advantage in connection with registry of opposite teeth. All prior patentees have understood that. the tips of opposite teeth should register vertically, but I have discovered that success in cold rolling hex peaks on the edge of the bar, depends upon having the registry very exact; far-more exact than has been realized, or has been possible heretofore. When cold rolling with few teeth and particularly with sharp teeth, the above described torque load variations havebeen so wide and rapid that variable back lash in the driving gears has permitted opposite teeth of the rolls to yield unequally; and very slight shift whether due-to back lash, or to unequal rotary or lateral springing or yielding of the shafts or;

housings, will permit the tips of opposite teeth to get out of registry enough to cause serious defects in the bar blank. That is to say, when opposite tips are in exact transverse registry, all transverse components of their wedging pressures are equal and opposite in all planes at right angles to the axis of the bar; and their longitudinal stretching and forward feeding components are the same for-both edges 'ofthe bar. But when the tip of one tooth lags behind the other, lines between corresponding areas of the wedging surfaces, slant across the transverse planes, becoming nearer perpendicular to the rearward surface of the lagging tooth and less perpendicular to the leading tooth, and thishas many bad effects. For instance, out of registry tends to be selfmagnifying, because the unbalance due to the slant favors the leading tooth and the leading tooth is more favored anyway, because it is nearer the dead center where less power is required to do the work. Also the transverse unbalance afiords a rotary component tending to curve the bar edgewise. Also, most. obvious of all, the opposite notches are out of registry so that the effective useful length and area of each blank is correspondingly. decreased.

Rounding the tips of the opposite teeth causes them to present larger surfaces, more nearly parallel with the length of the bar. and decreases I out-of-registry eflects, particularly near the dead center where the tips are nearest, so that it is possible to practically eliminate them, by 'hereinafter described means for adjusting the rolls for exact initial registry of the tips and gearing the rolls so as to practically eliminate back lash.

While-rounding of the tips presents a cylinder for first engagement of the tooth tip with the edge of the bar and enables the teeth to slidably clamp the bar before digging into the metal, I

have found that this eifect may be increased, by

transversely convexing the edge of thebar so that the first contact of the cylindrical tip is point contact instead of line contact. I Convexing the-edge of the bar is also useful whBre the tips are sharp edged instead of rounded.

In cold rolling hex peak edges, the teeth tend to squeeze out metal and make the bar thicker along the notches than along the central body portion, instead of vice versa, as in hot rolling; but. I have discovered this can be controlled so as to make the bar substantially as thick, or even thicker, along the, body portion, than along the notched edges. This may-be accomplished by slightly concave grooving the tooth faces, from base to tip. Such slight concavity is important at the tip, where the grooves meet, particularly how the tip and valley angles of the roll teeth may be varied; and, if desired, designed so as to roll notch surfaces apprommating a straightline, 120 hex outline, as nearly as may be desired.

The above and other features of my invention may be more fully understood from the following description in connection with the accompanying drawings, in which- Fig. l is a perspective view of a pair of tooth rolls, a bar being notch vrolled therein and a pair of flattening rolls driven by the notched bar which is forced forward by the notching rolls;

Fig. 2 is a section along the axis of the driven notching roll, on the line 2-2, Fig. Lgshowing also means for cooling rollbearings and independently and directly cooling the notching rolls, for cold rolling in accordance with my novel method herein set forth;

Fig. 3 is a transverse section on the line 3-3, Fig. 2, showing the means for screw indexing the driven notch and roll, for precise registry of the tips of its teeth, with the teeth of a cooperating roll from which it is driven;

Fig. 4 is a sectional detail on line 4-4, Fig. 2, showing on a greatly enlarged scale a desirable form of tooth; and part of a valley'between teeth; and Fig. 4a is a section on the line le -4a, Fig.

4, showing the transverse convexityof the edge of the stock bar and corresponding concavity of the face of the tooth shown in Fig. 4:

Fig. 5 is a side elevation of a bar showing dimensions and shapes suitable for makings. A. E. hexagonal nuts;

Figs. 6, 6a and Share. cross-sectional views of the bar shown in Fig. 5 on lines H, Ga-Ba and Gb-Gb, respectively;

Figs. 7, 7a, 7b and '70, Sheet 3, are all sectional detail views in a plane parallel with and midway between the lateral faces of a 26-tooth roll, including the operating part of periphery and bar; they are diagrammatically related vn the sheet to show the important successive phases and relations of the rolling operation;

Fig. 7d is a diagram showing how the peak and of the teeth in the critical region where the wedging faces of the teeth rotate across the'dead center; i

Fig. 8 is a side elevation showing roll and bar phase, relations corresponding to those of Fig. 70, but Fig. 8 shows operating peripheries of both rolls and both edges of the bar between them; Fig. 9 is a perspective view showing the front,

.oneside and front end of a horizontal roll mashowing on a larger scale, the means for variably limiting downward the adjustment of the upper rolls;

Figs. 13, 13a, 13b and 130, on Sheet '7, are diagrammatically related views like those for the 26-tooth roll on Sheet 3, but showing phase relations for a 36-tooth roll;

Figs. 14, 14a, 14b and 140, on Sheet 8, are

- corresponding diagrammatically related views 'showing phase relations for a 12-tooth wheel with sharp teeth; and 'how the tips may be rounded to embody one feature of my present invention;

Figs. 15, a, 15b and 15c, on Sheet 9, are corresponding diagrammatically related views showing phase relations for a -tooth 'roll, with narrower valleys between teeth; a

Fig. 16 is a face view of part of a roll showing rolled teeth convex from base to tip; and

Fig. 1'7 is a similar view of a bar 'having concave face notches rolled by the teeth shown in Fig. 16.

In these drawings, Figs. 1, 2 and 3 show the essential moving parts of a machine which was primarily designed for cold rolling notch bar blanks, suitable for S. A. E. hex nuts. In these flgures the housings and other stationary parts are not shown, except that Fig. 2 shows water cooled bearings and external cooling jets employed for practice of my cold rolling method.

As shown in Figs. 1 and 2, each notching roll is an annular ring 4, having its internal, shaftengaging surface of a radius two or three times the radial thickness of the ring. Preferably the side faces of the ring are coned, so that its shaft engaging surface is about twice as wide as the notching teeth. This gives greater strength and permits greater'shaft diameter and strength. Each notching ring- 4 is clamped between massive collars 5, 5a, coned to flt the ring. On both rolls, the collars project radially, equal distances beyond the tips of the teeth; and are formed with substantially parallel rubbing surfaces adapted to limit lateral thickening of the bar along its notched edge portions, under pressure of the teeth. As shown in Fig. 2, the stock bar is thinner than the distance between the rubbing faces. For S. A. E. inch blanks, the bar may be three or four hundredths of an inch thinner. See Fig. 411. When cold rolled, with plane faced teeth, the thickening is along the notched edges; consequently, the frictional grip of rubbing faces, is substantially the same for the opposite edges; so that said rubbing faces apply substantially balanced edgewise bending stresses as they rotate oppositely away from the line of feed of the bar, without causing any edgewise curvature thereof.

An important characteristic is that the journals 6, 6a, below and above ring 4 are massive, each being of diameter almost equal to that of the ring and. each of length substantially greater than its diameter. One object of this is to insure heretofore unknown rigidity of the shaft and prevent any springing apart of the rolls under the transverse stresses attendant on the rolling operation, particularly cold rolling. Another important objective, is to have great area for the journal surfaces, for dissipation of the heat generated by friction due to the deep working and flow of the cold steel, which is great at the edges of the bar, and also longitudinally along the middle portion of the bar. The heat effect due to rubbing of the side faces of the bar against the confining faces of the collars, is also substantial. As before explained, I have discovered that however rigid the roll shaft may be,

these heating effects are great enough to cause substantial variations in the length of the blank, primarily by varying the diameter of the rolls, bringing the tips of opposite teeth closer together, making deeper notches, displacing more metal, thereby increasing the length of the bar and varying the length of individual blanks. Aiso I have discovered that such variations can be minimized and made practically negligible by controlling the heat variations.

The great surface area of the journal makes it possible to carry off a substantial part of this heat by affording great area for water cooled bearings I, In. which are diagrammatically indicated by dotted lines in Fig. 2. I have discovered, however, that the heating effect is so closely localized in the bar and on the notching teeth that for cold rolling it can only be controlled by discharging cooling medium at proper rates and temperatures, directly against the bar and against the roll at the point where it engages the bar, as by nozzles 8, 8; and supplementing these by nozzles, as M, distributed at such intervals as may be found necessary, around the entire periphery of the roll.

As shown in Fig. 1, there are two pairs of rolls arranged in tandem. One is the pair of notching rolls 1, I, and the other the pair of flattening rolls 2, 2'. With plane face teeth,'the rolls operate to squeeze out more or less metal toward the flat faces of the bar Xa so as to make the bar thicker along the notched edges than along the center, and the second pairs of rolls operates to smooth out these deformities and make the thickness of the bar more nearly uniform. In the prior art, and in my own rolling of heavier bars as hereinafter described in connection with Figs. 9-12, the deformitieshave always been such that the notched bars could not be flattened sufflciently except by giving them a second pass through power driven. flattening rolls. In such cases, flattening must be by a second pass, instead of tandem, .because the feeding rate of power driven flattening rolls cannot be synchro- .nized accurately enough with the rate of the notching rolls to avoid varying push and/or pull on the notching rolls, and this causes variations of the spacing and shaping of the notches and in the resulting lengths of the blanks, thereby spoiling the bar for making nuts. In the present case, this variation in spacing is obviated, by disconnecting the power drive for the flattening rolls and rotating them only by the driving enamuse tening roll resistance. However, a very substantial backward push or forward pull may be applied, provided it' is constant in amount and not too great. So while some non-positive pull, as by a weak electric motor; or similar push may be applied by friction on the flattening rolls, I prefer to utilize the work done in flattening the bar as a source of sufficiently constant, and sufficiently small rearward push, as above described. As the rearward pressure thus applied, must be very uniform in order not to interfere with uniform feed, stretch, and notch surface rolling by the notching rolls, I prefer to have the bar engaging faces of rolls 2, 2' knurled, as shown. In this way, slight irregularities in the face of the bar may be regularized sufliciently for practical purposes, with less work, less slip, and less variations in' back pressure, than where smoothsurface cylinders are used.

The next important point is roll tooth registry. Though both rolls are provided with driving pods 9, M, as is customary in ordinary bar rolling machines, only the lefthand roll has its pod 9 connected to a power driving shaft 00, and this roll has a gear M rigidly keyed to it and meshing with gear lid on the other shaft. The pitch line of these gears is as near as possible the same distance from the axis as the axis of the bar at dead center. Hence any circumferential springing or back-lash is 100% efiective in rotating the tips of the notching teeth out of registry.

This shows the'above described importance of multiplying the number of teeth deeply working in the metal to smooth out and render more uniform the torque load on the tooth wheels and gears, but the present points of importance are: First, I employ relatively expensive, accurately made herringbone gears, which minimize unequal springing in response to the varying loads. It will be understood that a herringbone gear .is fundamentally two helical gears each having helically slanting teeth which are identically the same, but of opposite pitch. The spacing and the angle. of slant of each helical gear is usually 'such that three or four adjacent helices on said gear, overlap longitudinally, with the result that each helical gear maintainsthree or four points of simultaneous, endwise-progressing engagement wits its mate. This type of gear transmits torque very uniformly throughout the entire 360 of its rotation. Furthermore, when two such helical gears of opposite pitch are combined in herringbone relation, the endwise resultant of the slantingly engaged teeth is balanced-and the number of simultaneously engaged teeth is doubled.

It follows that except for the difflculties of getting correct phase relation, two separate oppositely pitched helical gears can be employed. of if the endwise resultant is properly counterbalanced in some other way, a single helical gear can be used; In either case, the progressive slanting engagement of the teeth has advantage over so-called spur gears, the form of the latter being such that slight wear, or slight change of radial adjustment of the shaft axes, will develop back-lash aswell as non-uniform torque trans-- mission.

The next important thing is providing for initial registry of the tips, asperfectly accurate as possible. This is accomplished by providing a -'rotary screw adjustment for the gear I id of the driven roll, a specific means employed being as follows:

The construction is the same as for the driving roll, except that in the drlving'roll the gear H is splined to prevent any rotation, while for the driven roll the gear Ha is rotatable and is provided with the means for rotatably adjusting the shaft and notching roll teeth, with reference to the gears and teeth of the driving roll.

In both of them (see Fig. 2) the lower journal portion 6, is integral with the collar- 5, which forms one clamp for the' notching ring 4, and is also integral with the main shaft element E3. The ring d, the internal diameter of which is only slightly greater than the shaft is, is nonrotatably secured to said.shaft by splines ta, preferably four in number. The opposite clamping collar bat for the notching rings is integral with bearing sleeve ta, which is slidable along shaft it. In both cases the endwise movable parts are clamped together by nuts, as ita, lib, igrewed down on threaded ends We, lid of shafts Differences are that driven gear lid is integral with a sleeve it formed with a disc portion Ma having opposite lugs i lc, each having radial with a sleeve it, which is splined to the end of the shaft E3. The circumferential wedge shaped spaces between adjacent faces of projections Mo and i502 are sufficient to permit rotation at least as great and preferably considerably greater than the distance between tips of teeth on the notching ring and, within this range they are adjusted and solidly locked against rotation .by

screws l5b threaded in projections 85a, lid, and similar screws Md in lugs Me. For adjusting, two adjacent screws Ma, I5b are suflicient to apply equal and opposite thrust between three adjacent lugs, thereby completely locking the parts against rotation. The opposite screws Md, i5b are employed for more symmetrically and securely clamping the lugs in the adjusted position.

By using gears with an odd number of teeth, in combination with notching rolls havingan even number of teeth or vice versa, a vernierlike relation may be established which greatly simplifies getting accurate tip registry. In a particular case I used herringbone gears having 19 teeth, and 4-pitch, in. connection with a notching ring having .26 teeth. Also, as shown in Fig. 1, keyed collar i5 may be .marked with ascale having its zero exactly in line with a tooth tip and in vemier relation with a scale on lug Mc.

The lug locking means for the torque-applying gear Ila. being at the end of the shaft l3 and remote from the notching ring 4,-which resists the with the back face of a collar which is integral with the shaft like present-collar 5, and, like the Fig. 11.

latter, located near and used as a clamping collar for the notching ring; all as will be explained in connection with Figs. 9 to 12 which show a machine primarily designed for notch rolling heavier bars.

The housing and other necessary supporting means for the vertical rolls shown in Figs. 1 and 2, may be in a general way, somewhat like that for the horizontal machine shown in said Figs. 9 to 12 inclusive, which was primarily designed and used for rolling heavy blank bars such as required for U. S. S. 1%"nuts and in which the rolls are horizontal instead of vertical.

As will be evident from all the figures, the housings, as well as the roll and gear assemblies, are very massive and all dimensions includin diameters of notching ring, shaft and gears may be approximately three times those required for the smallermachine shown in Figs. 1 and 2.

As before noted, ,the roll assemblies are slightly different from, and have certain advantages over the arrangement shown in said smaller machine. increases in high' ratio, with the increase of diameter, it became possible to make both journals 6m, 6:: integral with the shaft l3z, as shown in The toothed ring 4:: keyed by four splines, 41/, is clamped by two collars, 512, 5:0, neither of which is integral with the shaft, both being endwise slidable into. position. The clamping collar 5x abuts against the face of an annusaid notching ring, by a ring nut 5a. The oppo jtion are clearly shown in the drawings.

lar enlargement 1:, which is integral with the shaft I33, like collar. 5 of, Fig. '1, and the cooperating collar liv, on the other faceof the notching ring is forced into engagement with site face of said integral shaft enlargement x, is formed with segmental lugs My, extending parallel with the axis. The cooperating segmental lugs l5z are integral with the relatively rotatable sleeve I511 which carries gear H3. The adjusting and locking screws I42, I52, two of which are shown in Fig. 10, are arranged as in Fig. 3. The rotatable lug sleeve I511 is held in As the transverse strength of the shaft place by a ring nut I31! screwed on the other end of the shaft l3z. The stock bar X2 and the notch rolled portions X3 are shown in Figs. 9 and 10.

The housings, journal boxes and adjusting means for supporting the rolls in operative rela- The housing comprises a pair of similar end frames which are secured on heavy foundation channels l8; and spaced by tubular thrust members It against which they are tensioned by bolts Ha. Each frame comprises a lower section 20 havin a vertical slot 20a inwhich the bearing boxes ix, W, are removably fitted. Each slot is closed in at the top by a massive horizontal member 2|, the ends of which have lateral locking en-. gagement with the outside faces of the lower section 20, as shown at Ma. The top section is gl'ampeddownward by massive bolts 22 and nuts The upper journal boxes In: are each clamped downward by large diameter thrust screws 24 engaging plates 25, arranged to distribute the pressure on the top of journal box Ix; and each is clamped in the adjusted position by hand operated lock nut 24a.

The journal boxes and.journals are flooded with cooling liquid through pipe 26 having three branches, as shown in Figs. 9 and 10. As shown in Fig. 11, branch 28a discharges through a duct, directly onto the Journal Bar. The water from the'journal boxes, or other cooling fluid may be discharged on thecrolls and/or on the bar as in Fig. 2.

The very small range of vertical adjustment, permissible by the use of the herringbone gears may be accurately predetermined by having the bearing boxes Ia: of the upper roll, each very rigidly supported by a wedge 21 cooperating with a wedge surface 210 on the bottom of the box. Wedge 21 is accurately adjustable, and rigidly held, as shown in Fig. 12, by nut 28 on a retracting screw 28a, which is threaded in the wedge and has sliding engagement with bridge bar 28b, which is clamped to the housing by bolts 29; in combination with a pair of thrust screws 30 threaded in the bridge bar on opposite sides of retracting screw 28a.

In this larger machine both of the shafts are provided with the tooth-registering lug and screw adjusting means, having small-clearance between their peripheries and in order that the ends of the adjusting screws may not project far enough to interfere, the screw holes are countersunk, as shown in Fig. 10. Similar adjusting means for both rolls has the advantage that the paths for conduction of heat from the toothed rolls will be identical, thus making it easier to keep the rolls at the same temperature.

While the problems involved are much the same for both machines, it seems better to continue explanation of the other novel principles of my invention, in connection with rolling the smaller blanks required for the S. A. E. hex nuts for which the machine of Figs. 1, 2 and 3 was originally designed, it being understood that the basic principles described are more or less applicable to rolling heavier bars and that many of the discoveries in connection with cold rollfiig, are of advantage where the initial bar tem-,

perature is moderate enough so that the result- I ing plasticity of the steel is not great enough to introduce new difficulties and problems.

Figs. 5, 6, 6a and 6b show the dimensions, proportions, and, in a general way, the shapes of the stock bar xa, notched blanks Xb, and knurled blanks Xc, as actually produced in the tandem machine shown in Fig. 1. In this case, the rolls lied the longitudinally concaved faces described above, but the stock bar used, and hex-edge outline, is substantially the same as when plane faced teeth are used; but in the latter case, the knurling is localized along the thickened portions near the notches, instead of near the middle.

In this particular case, Fig. 6 shows that in the original stock bar the width at xu was about twice the thickness and that the upper and lower edges were slightly convex. The outline of the blanks produced so far as it can be shownin side elevation and on the scale used in Fig. 5, would be the same-if edges of the bar had been flat;- also if the teeth on the roll had been formed on transversely straight lines instead of being concaved. However, though shown four times enlarged in Fig. 4a, teeth 4a had their tips Id circularly rounded on a radius of and the bottoms of the notches at Xd, are similarly rounded. The tips of the blanks show short, relatively flat surfaces at Xe, because, in cold rolling hex peaks, practically all shaping of the notches is accompanied by depression of the metal, and the peak is depressed by adjacent metal, so that it does not completely fill the notch of the roll. There are other slight departures from hex outline, the reasons for which will be ex to the metal gathering effect of the groove;

and is the reverse of the effect produced by cold rolling with plane faced teeth operating on straight-edge bars. In the latter case, the tendency is for the teeth to squeeze out metal laterally and thicken the bar along the edges of the notches, rather. than along the center, particularly where the bars are cold rolled.

In this special case, the notching ring was made of diameter just large enough to accommodate 26 teeth of size and spacing suitable for blanks for nuts. In generaLthe 360 circumference of the ring, divided by the number of teeth, gives degrees of tqoth are, that is, the peripheral extent of each tooth; and this tooth arc plus the peak'angle always equals the valley angle. This interdependence of tooth are and valley angle is one of the complications making it diflicult to intelligently change the roll-diameter, or the number of teeth, or the tooth peak and valley angles. For any number of teeth or any given value of peak and valley angles, increase or decrease of either causes correlative decrease or increase of the other.

From the above, it follows that, for the 26- tooth roll illustrated on Sheet3,

which plus peak angle 106%=val1ey angle 120?. In a general way, it may be said that whatever the virtues or defects of this roll for this size of nut, they are the .same for any other size of roll having the same number of teeth and the same peak and valley angles, except that when thenuts are very large or very thick there may be certain practical difliculties, but these will be referred to later. Consequently, for convenienc'e in study of the action of the roll on the blank, the drawings may be and are enlarged proportionally. Sheet 3 is a 4 to 1 enlargement, and a roll so enlarged would make blanks about the size required for 1%," nuts.

Sheet 3 shows four sectionalviews, Figs. 7,70, 7b, 70, all in a plane parallel with and midway between the lateral faces of the rolls. These four views taken together, show the roll'and the blank at all important phases of the rolling operation. lowed by taking the first row of each view diagonally down the sheet, these being marked A, B,"

C; D; then the second row, E, F, G, H; then the third row, I, J,.K.

Figs. 7e and 7 are diagrammatic 8 to 1 enlargements of one "tooth arc". showing more' precisely the rolling action in the critical region where the wedging faces of the tooth cross the dead center.

These views show that in this 26-tooth, hexvalley roll, all tooth faces presented in the direction of the feed, are wedging the metal forwardly throughout all contact therewith; and, correlatively, all faces presented in the opposite direction are wedging the metal rearwardly. In the case of the first tooth, in positions A, B, C, D, this rearward wedging takes effect on the incoming unrestrained part of the stock bar and slows up the speed of its forward movement until the next .tooth has measured 'oif a new The successive phases can be fola substantial faces correlatively decreasetheir wedging angle.

These equal and opposite variations in the angles of the two sets of wedging faces become important near the dead center, where they are traveling approximately parallel with the direction of movement-of the bar. The reason for this is that each forwardly wedging face completes its rolling on the rearward face of a bar notch while its tip is approaching the dead'center and is still penetrating downward substantial distances in the metal as well as increasing angle; while a rearwardly wedging face completes its rolling. on a rear bar notch face, after its tip has passed dead center and is'rolling upward out of the bar notch.

A general idea of the peculiarities of the successive phases may be had by following them down the sheet, tooth A, Fig. 7, is about to enthat would be desired'for the hex'notch, but the 5 upper or valley end of this 30 face is already close-to the dead center where its downward penetration into the bar is approaching zero; while its lower or tip end is still penetrating downward in the metal at a substantial rate. result is that for one-half tooth arc of rotation. that is, 3&3, this forward wedging surface of F rocks from top to bottom on the face of the bar notch, and by the time F reaches position G, the 30 base of the bar notch has increased its 'angle by the half tooth arc angle, 6+8", so that the average or over-all angle is increased a quarter tooth arc, to 335?. The diagram, Fig. 7], shows this rolling action for. four positions of the forward wedging surface.

Therearwardwedging surface, beginning at 57% at position B, correlatively decreases its angles through the above positions C, D, E and F,.but it has to go through 4 more positions, to J, Fig. 7a, before it gets down to the 30 angle. During most of the rotation to phase G the incoming stock bar was freely wedged backward by said rearward wedging face. Thereafter, the rearward wedging is constrained by the forward wedging face of tooth C. There is nothing. critical until it reaches position H. where its rearward wedging angle is about 36 At this point the lower end of this rearward wedging face begins to move across the dead center and begins to move upward out of the bar, but at the same time it is decreasing angle, so as shown'at I, the lower part of the face is relieving pressure at the bottom of the notch face, while the upper end is still crushing downward. The result is a. rolling action through .a half tooth arc angle, 3&2", which leaves the rearwardly wedged face of the notch at an over-all rearward angle of 33J;, the same as forward angle rolled on the other face of the same bar peak.

Thus the forward wedging face rolls from; top

to bottom of the bar, notch while it is approach 75" The ing dead center, beginning when its lower tip end is nearly one-half tooth are before dead center, where said tip is penetrating downward in -the metal; and its upper valley end is beyond dead center and lifting away from the metal. Reversely, the rearwardly wedging face rolls correlatively from bottom to top of the bar notch when its lower tooth-tip end is moving away from dead center and lifting out of the notch and its upper end is approaching dead centerand still flattening downward into the metal.

The net result is that the notch angle is onehalf tooth arc angle, 6H,-less than the 120 required for bar. faces parallel with the hex blank to be punched therefrom. This discrepancy amounts to 3&2 for each face of the notch and theoretically an over-all correction could be had as indicated in Fig. 711, that is, by flattening the valley 6H, that is, one-half tooth arc angle. For each face, the flattening is 33. This gives a good approximation of a hex notch for the bar. This over-all correction would involve flattening the angle of initial engagement of the forward wedging tooth, from 12% to about 9 6. This may be useful for rolling very soft, plastic metal,..

but for cold rolling steel, such relatively flat engagement increases the danger of shearing the metal even when the tip of the tooth is cylindrical. Moreover, it tends to compress the metal transversely of the bar to such an extent that the top of the peak on the bar, which is always depressed and flattened to some extent, is still further depressed; thereby making the flat still longer, and of still less altitude. So for cold rolling hex peaks on the edges of "steel bars, it is found that a higher, sharper peak can-be made I and, in actual practice, more metal can be saved,

where the roll-valley angle is 120 as shown in Figs. 7e, 71, even though the resulting bar notch face is steeper than the regular-hex. So prac-- tically,-for cold rolling, the theoretical correction is undesirable and as a, matter of fact an [even better peak can be rolled and more metal saved, where the roll-valley angle is even less flat, say, 110 instead of 120, and its tooth angle is correlatively sharper, say, 96 instead of 106.

It may be noted that for too few teeth or too sharp tip and valley angles, the convexing may become serious, in which case it could be corrected and a, straight line notch face generated by convexing the respective straight tooth faces along the curves corresponding to the curves each of them generated by rolling on its bar notch face.

Tip rounding: This has the advantage of presenting the tooth tip to the metal as a circle even at position A, where it first touches. So opposite teeth flatly and slidably clamp the bar before digging into the metal. Then, for the first small penetration, the arc of the cylinder gives equal forward and rearward 'wedging. Then it begins to wedge or cam the metal rea ardly more than forwardly, but its rearwardly presented surface 3 shifts to forward surface, and at deacL center position H they are equal and opposite. Thus the tip is a cylindrical bearing, rocking in the bottom of the notch, and this is particularly important at and near dead center, where it receives its final form. For about one-eighth tooth arc in advance of dead center, downward penetration of the tip of the tooth H was practically completed and the forwardly wedging surface had rotated out of contact with the bar, except atthe very bottom of the notch where the cylindrical tip presents half of its surface forwardly to take .the pressure and prevent marring the bottom of the notch; and its whole surface acts as a cylindrical bearing for the angular rotation of the tip relatively to the bottom of the notch. This rotation smooths and even polishes the bottom of the notch.

'As will be more fully explained in connection with Fig, 14, Sheet 8, a tip at A, Fig. 7, measures 01f and cups in between itself and tooth E, a length of metal which, measured parallel with the bar, is then less than when the valley reaches dead center position at F, J. Consequently, there is a stretching effect that tends to increase the space available for the bar hex projection. At the same time, downward,movement of the valley as a whole, creates pressure tending to expel the metal from the valley. These forces cause relative longitudinal, angular and vertical shifting and flow of the metal adjacent the tip, and the cylindrical curvature of the tip tends to facilitate and smooth such shifting and flowing as compared with theabrupt action of a sharp tip.

These tendencies, and the values of the tip convexing will be more evident from consideration of Sheet 8, where'Figs. 14-14c inclusive show phases of a roll having 12 sharp teeth. When the tooth E is at dead center, the preceding tooth A is only justdigging into the bar. At this time the bar is being fed by tooth E at the full peripheral speed of the roll, but the tip of tooth A is moving in the direction of line a-b, which is at an angle of 30 to the direction of feed, and the component of its movement, lengthwise of the bar, is much slower'than the bar movement. This is shown on Fig. 14 where the vertical distance to the bottom of a notch is indicated by line a-c; the corresponding horizontal feeding component by line 6-1); and the primary 30 direction and The difference between lengths of lines 0-4; and b'c shows that the feeding speed of tooth A is about one-eighth less than the peripheral speed; hence one-eighth less than the 100% effective feeding speed of tooth E measured by the much longer-line w-b. Thus, when the sharp tip engages the bar at A, its speed lengthwise of the bar is one-eighth slower than that ofthe bar. Consequently, the rear surface of tooth A acts like a ratchet to-slow down the bar and jam the face of the already completed bottom of a notch. backward against the frontface of tip E, thereby deforming the notch. Under these conditions, it is easy to see the value of tip rounding; first, rounding the tip of tooth A to smooth the ratchet 'efl'ect by permitting momentary rearward cammingof the metal by the cylindrical surfaces; andsecond, rounding the tip of E, so that whatever remains of rearward Jamming tendency by tooth A, it will take effect on thesurface of a cylinder. Another advantage of rounding is that just before position E the rearward thrust of E was not counterbalanced by anything except by the merest tip of the forward face pivoting in thebottom of the notch. "The cylindrical tip helps to spread this pressure, and prevent gouging at the bottom of the notch.

The value of rounding is also finportant in conerally about 2, but even this small-angle lag, makes the thrust between opposite tips, along line 0.1: slant at an angle of about 10 to the dead center line 0. Also, for theopposite tooth surfaces between vertical lines 0 and 0', those 'of tooth H are rearwardly wedging while those of lagging tooth H are forwardly wedging; and their opposite pressures, exerted out of line, at different wedging angles, and along opposite edges of the bar, constitute a rotating couple introducing the previously described rotary eifects, such as tendency to increase their out-of-registry lead and lag, edgewise curving, etc.; in addition to decreasing the useful length of each blank, by an amount equal to twice the distance of lengthwise displacement shown by vertical lines 0, 0-.

Number of teeth: As above explained, an important object of my invention is to control and regularize the rolling and endwise feeding action of teeth in the critical region at and near the dead center by having enough other tooth surfaces deeply engaging the metal at points in ad- Vance of said critical region; but not so many that the desired lengthening of the bar by one pair of rearward wedging surfaces will be unduly interfered with by a succeeding pair of forwardly face, and its 27 degree, larger-area, forwardly wedging face; and in practice, the resultants will vary with variations in resistance of the metal, friction on the bar, and particularly with any variations in vertical registry of the tips.

For a short are beyond C, the areas of oppositely wedging contact surfaces begin to equalize, but before reaching D, base-to-tip rolling has begun and the forward wedging face has already rolled'out of effective pressure relation with the bar notch surfaces. Between D and E just before the nexttooth A engages the bar, the forward wedging face exerts pressure on the notch face, only at the very tip and this pressure on this vanishing tip area must be suflicient to bal- 1 and at this instant the tooth A comes in with the above described ratchet detent action on the bar, checking the speed at which E was feeding the bar and thus working in combination with said unbalanced fulcrum leverage to Jam the notch face back against the tip of E.

Rounding the tip can greatly reduce the damage, but it does not change the above described condition that feed at C and D is dependent solely on the action of a single tooth, and that between D and E the pressure is almost wholly rearward by one tooth and is balanced onlyby digging in of the tip.

The unequal lengths of thebianks and the bad shape of the notched face has resulted in proposal to reduce a number of teeth to 8 and try to control the lengths of the blanks by using a bar feeding roll intandem with and geared to the toothed roll; but as may beinferred from the fact that while the bar-feeding speed of a rotary cyltooth and the cylinder, applying reversing torque loads and causing slip, both fatal to uniform bar feed and notch form.

As contrasted with this, my invention recog-,- nizes the feeding effect of my notching rolls is paramount; and that the unequal feeding tenden-. cies of teeth throughout the half tooth arc in advance and the half tooth arc beyond dead center can only be controlled by using enough other teeth, in advance of such critical region, to co operate with and afford an internal balance between the opposite faces of all teeth.

Comparison of the diagram showing for the l2-tooth roll on Sheet 8, with the similar shows ing for the 26-tooth roll on Sheet 3, shows that at the critical point G, Fig. 7b (which corresponds to D, Fig. 140), I have not only a half of a forward wedging tooth still deeply in the, metal at dead center, but I also have the tooth C deeply and eifectively engaging the bar and cooperating. to control the feeding speedand the resulting lengths of the blanks. Then, at dead center position H, corresponding to E, Fig. 14, said cooperating tooth C has penetrated still further to position D, so that the otherwise unbalanced pressure on the rear face of H, is modified and controlled by pressure on the forward face of D.

On Sheet 3 this control of forward feedby two surfaces outside of the critical tooth arc is accomplished without any restraint on the rearward wedging of the bar up to the point E, and very little restraint up to and slightly beyond F, where the base-to-tip rolling of .the forward surface be'gin Sheet '7, showing similar phases for a 36-tooth roll shows how control of the feed and blank length by teeth outside of the critical region, can be still further increased and balanced by increasing the number of teeth up to the point where rearward flow of metal from tooth G, is

seriously interfered with, and then practically prohibited, by interposition of the forwardly wedging surface of tooth D.

Figs. '15, 15a, 15b, 15c, diagrammed on Sheet 9, show a 20-tooth roll actually used in the larger machine Figs. 9-12, for notch-rolling, heavy bars, suitable for 1%," U. S. S. hex nut-blanks, These views show how a commercially usable blank bar may be rolled where considerable variations are made in some features of my invention, and others are omitted. In this case the number of teeth is very considerably below the preferred number of 26, and is approaching the minimum limit. The rolls are set too far apart so thatthe bar notches are correspondingly shallow andtips A, B, C and D are at least one full phase higher out of the bar than when the roll setting is normal, as in all the other views. high position of the roll and sharpening of the tooth peak angle from-106 down to 92, where tried because of the great strain involved in cold This' rolling these heavy bars, which were 1 thick the benefit of all its novel features, including means for practically perfect registry of opposite teeth; also, elimination of back-lash or other yield; also the tips of the teeth were rounded even though on a relatively small radius. Thus with several features of my invention perfectly utilized'and two more of them utilized in principle but not in suiiicientdegree, blanks of uniform shape and length were produced.

,Figs. 16 and 17 show respectively, a 26-tooth roll 4k having longitudinally convex tooth faces, 417, 4r, and a bar X'I having correspondingly concave faces Xp, Xr. Such bars are useful where the blanks are pierced to internally expand the metal and thereby fill out the concavity of the bar peaks. These bars also had blanks of commercially uniform length and perfectly uniform shape because the rolls had enough teeth, the teeth had rounded tips; the rolling operation had to the benefit of all novel features of my apparatus; and the temperature was controlled in accordance with my method.

As concerns my cold rolling method, it is to be emphasized that the cooling medium is preferably a liquid, preferably water or a water-like fluid, because the heat generated by the rolling is enough to render the bar hot, within the meaning of hot as applied to water; and the specific heat of water is so great, and the latent heat at boiling is also so great, that there is no difficulty in limiting the temperature variations to a relatively narrow range, usually substantially 'below the boiling point of said water.

On the other hand, in hot rolling, the bars are initially many hundreds of degrees above ordinary temperatures. Consequently, they are quickly cooled hundreds of degrees by the atmosphere, and by the rolls, and by any .other parts of the apparatus "with which they come in contact. The resulting temperature variations are obviously very great; and applying water or other effective cooling medium to such hot bars, can only accelerate heat losses, and thereby greatly, magnify variations in volume, length and plasticity.

So rolling bars that are initially hot presents conditions entirely different from those of a bar which is initially at ordinary temperatures and has practically no temperature variation during the time it approaches the notching roll; and the only heat that causes the objectionable temperature variation, is the relatively moderate heat generated in and by the rolling operation; and cooling water used as above described is an extremely effective means for limiting such heat variations to a very narrow range.

I claim: I

1. A method of shaping stock bars by means of toothed rolls, to formconnected hex nut blanks having substantially uniform shape and dimensions throughout, and substantially without edgewise bending of the bars, which method includes edge rolling of the bars while they are approximately cold; thus generating excessive heat by edge rolling of the bars to form notches having faces approximating the angle of two adjacent hexagons, while said bars are approximately cold; thus generating excessive heat by the great friction between the teeth and the cold metal of the bars, and by the violent compression and excessive flow of the cold metal within the bars; while limiting the effects of heat, by applying cooling medium directly to the rolls and to the metal in and near the rolls, in quantities and at temperatures suiiicient to prevent substantial changes in volume or plasticity of said metal.

3. A method of edge-rolling stock bars to form hex nut blanks of uniform shape and dimensions, which method includes rolling cold bars by toothed rolls to form notches having faces approximating the 120 angle of two adjacent'hexagons and applying cooling medium to the metal being rolled, in quantities and at temperatures sufiiciently to limit rise of temperature due to friction and flow of the metal while being rolled, thereby maintaining substantially uniform plasticity, temperature, and lengthening of the metal during the rolling.

4. A method of edge-rolling stock bars to form hex nut blanks of uniform shape and dimensions, which method includes transversely convexing the edges of the bar and then passing it through the notching rolls.

5. A method of edge-rolling stock bars to form hex nut blanks of uniform shape and dimensions, which method includes transversely convexing the edges of the bars and rolling the.

notches by toothed rolls having their wedging surfaces transversely concave on an arc of less curvature than that of said convex edge of the bar.

6. A method of edge-rolling stock bars to form hex nutblanks of uniform shape and dimensions; which method includes transversely convexing the edges of the bars and rolling the notches by toothed rolls having their wedging surfaces transversely concave on an arc of less curvature than that of said convex edge of the bar, and applying cooling medium to the metal being rolled, in quantities and at temperatures sufiiciently to limit rise of temperature due to friction and flow of the metal while being rolled, thereby maintaining uniform plasticityand length of the metal during the rolling.

.7. A machine for edge-notching bars to form nut blanks, including parallel shafts with similar meshing gears and similar toothed notching rings rigidly secured on the respective shafts; the gears being herringbone gears of width substantially greater than their radius. Y

8. A machine as specified in claim 7, with the further feature that the herringbone gears have many more teeth than there are notches in the tooth roll; and each tooth of each gear slants circumferentially through an are greater than one tooth arc of the roll which it drives.

9. A- machine for edge-notching bars to form nut blanks, including parallel shafts with similar meshing gears and similar toothed notching rings rigidly secured on the respective shafts; the tips of the teeth having cylindrical surfaces to which their wedging surfaces are approximately tangent; means for driving one of said shafts, thereby driving the other shaft through said meshing gears, the gears being herringbone gears of width substantially greater than their radius.

10. A machine for edge -notching bars to form .nut blanks, including parallel shafts with similar meshing gears and similar toothed notching rigidly secured on the respective shafts; each ring having not less than 18 teeth with tip angles not less than 90; the tips of the teeth having cylindrical surfaces to which their wedging surfaces are approximately tangent; 'the diameter of each shaft being more than twice the minimum radial thickness of its ring; means for driving one of said shafts, thereby driving the other shaft through said meshing gears; the gears being herringbone gears of width substantially greater than their radius; and the securing means for one of the gears including screw means for relative rotary adjustment of its shaft and tooth tips into exact registry with the tooth tips of the other shaft.

12. A machine for edge-notching bars to form nut blanks, including parallel shafts with similar meshing gears and similar toothed notching rings rigidly secured on the respective shafts; each ring having not less than 18 teeth with tip angles not less than 90, the tips of the teeth having rounded surfaces for initial contact with the bar; the diameter of each shaft being more than twice the minimum radial thickness of its ring; means for driving one of said shafts, thereby driving the other shaft-through said meshingg'ears; the gears being herringbone gears of width substantially greater than their radius; and the securing means for one of the gears including screw means for relative rotary adjustment of its shaft and tooth tips into exact registry with the tooth'tips of the other shaft.

13. A machine for rolling hex peak edges on bars to form hex nut blanks, including parallel shafts with similar intermeshing gears and similar toothed notching rolls rigidly secured on the respective shafts with opposite teeth in registering pairs, each roll having enough teeth, set close enough, so that when the wedge surfaces of one pair are near dead center, a succeeding pair will be deeply engaging the bar; in combination with rolls for flattening the notched bar, ar-' ranged in tandem relation to the notching rolls, and the peripheral speed of said flattening rolls being governed by the endwise speed at which the notched bar is pushed forward by said notching rolls.

14. A machine as specified in claim 13 and in which the flattening rolls have knurled-surfaces.

15. A machine as specified in claim 13 and wherein the teeth of the notching rolls have rounded tips. I

16. A machine as specified in claim 13 and wherein the flattening rolls are driven by the notched bar.

17. A machine as specified in claim 13 and wherein the intermeshing gears are helical gears and the means whereby one of said gears is rigidly secured to its shaft includes a screw whereby it may be accurately adjusted for getting registry of opposite teeth of opposite notching rolls.

18. A machine for notching bars to form nut blanks, including means for rotatably supporting similar sets of notching teethv with opposite teeth in position for engaging opposite edges of the bar, the tips of the teeth being formed as cylindrical surfaces to which their wedging surfaces are approximately tangent.

19. A machine for notching bars to form nut blanks, including means for rotatably supporting similar sets of notching teeth with opposite teeth in position for engaging opposite edges of the bar;

said teeth having their wedging surfaces transversely concave.

20. A machine for notching bars to form nut blanks, including means for rotatably supporting similar sets of notching teeth with opposite 'teeth in position for engaging opposite edges of thebar, said teeth having their wedging surfaces transversely concave and their tips rounded.

21. A machine for edge-notching bars to form 'nut blanks, including parallel shafts with similar meshing gears and similar toothed notching rolls rigidly secured on the respective shafts with upposite teeth in registering pairs, each roll having enough teeth, set close enough, so that when the wedge surfaces of one pair are near dead center, a suceeding pair will be deeply engaging the bar; .the tooth-arcs being not substantially more than 18 nor substantially less than 10, the

teeth-having valley angles approximating hex notch angles and having their wedging faces longitudinally convexed on curves such that when near and crossing dead center, they will roll substantiallyj plane bar-notch faces.

22. A machine for edge-notching bars to form nut blanks,.including parallel shafts with similar meshing gears and similar toothed notching rolls rigidly secured on the respective shafts with opposite teeth in registering pairs, each roll having enough teeth, set close enough, so that when the wedge surfaces of one pair are near dead center, a succeeding pair will be deeply engaging the bar; the tooth arcs being not substantially more than 18 nor substantially less than 10, the teeth having valley angles approximately hex notch angles and having their wedging faces lon-' gitudinally convexed on curves such that when near and crossing dead center, they will roll ap-

Images