(2)
 

Patterns for Cannons

    At Salisbury, as at any cannon foundry of the period, clay of a quality suited to modeling, molding, and the requisite firing was used for both cannon models and molds. It would have been worked into a paste, with sometimes a binder (sometimes temper) of lint from linen cloth added, along with fine sand, and ashes. Occasionally, burned salt mixed with iron filings (called by the early founders "lymmel of iron"), was introduced and thoroughly kneaded and mixed. Chopped straw might be added for a stiffer mix. Loam, dung, and powdered brick are other ingredients sometimes used.

    The cannon patternmaker formed his pattern about a smooth log mandrel, shaping, scraping, and adding patches and fillets of clay to build up his form. The series of steps are shown in the rather sanitized French drawings from Diderot.  English contemporaries first wrapped their mandrels with a tightly coiled rye-straw rope, whose function will be seen later (Fig. 3).

   Wet trowels and spatulas were used to blend the plastic medium and secure the requisite final shape and surface texture (Fig. 4). Longitudinal template boards helped maintain correct barrel profiles. After slow air-drying to prevent cracking, the pattern would then be thoroughly dried and cured over a long pit of embers, similar to the later step of drying the mold, illustrated in Fig.7-A

    For the relatively plain armaments of the Continental Army, little additional ornamentation was added - but at this juncture (Fig. 5-A) is when cannon founders, using removeable skewers and temporary iron pins, would add rings, arms, floral devices, mottoes, dolphins or other ornaments of ancient ordnance, and these were separately formed first either in wax or plaster (Fig. 5-B: Orig. retained by publication: No longer available. It showed many of these designs and devices...).  The finished pattern was next smeared with hog fat - or as we have seen, with the coveted black lead from New Haven - which acted as a separator or mold release to prevent the mold from sticking fast to the surface of the cannon pattern. .

    Middlebrook says this model or pattern was made about one-third longer than the cast barrel but this needs clarification. When the molds were finally assembled in the casting pit, this extension on the upper end acted as a reservoir of hot metal during the cast, greatly increasing hydrostatic fluid metal pressures in the lower or critical breech end, while simultaneously trapping impurities and aspirations as the molten iron surged into the cannon mold sprue. In post-casting operations, this extension or "head" was laboriously removed by sawing it free from the barrel.

    In the English foundries on this time level, it was often the practice to make a separate "head" pattern for this section, the mold from which was assembled onto the upper end of the barrel mold prior to casting. At other times, this "head" portion may have been integral with the barrel pattern (as Middlebrook suggests), which seems to be the case with the pattern upon which the worker with the mallet is apparently attaching separate trunnion patterns (Fig. 5-A). (That this is indeed a pattern, not a cast barrel, is indicated by the absence at left of a breech and cascabel and the visibly-protruding pattern mandrel). Note particularly the right end of this pattern where an integral "head" portion can clearly be seen extending beyond the muzzle face. Middlebrook also says that the patterns for cascabel and trunnions, being round, were turned of wood on lathes.     In this, he is almost certainly wrong. While wood is a preferred medium for reuseable patterns in split-mold casting (see below), it has no use in expendable patterns or pattern components which are completely invested by their mold media.

 Playing In The Sand Box

   Basic foundry practice in sand casting involves forming a cavity in sand with a "pattern" - a duplicate in wood or clay (today often metal or plastic) of the object to be cast. The pattern is "invested" in "green" sand (usually containing an organic binder), retained in a split "flask". When the pattern is completely covered and the sand rammed down hard, then the flask is opened along the split. The upper half ("cope") is set to one side, and the pattern is "drawn" from the bottom ("drag") half of the flask. Then the flask is reassembled; guide pins help exact realignment. The molten metal is then introduced into the sand cavity via an opening ("sprue").  When chilled, the flask is "shaken out" and the cast object retrieved...
 
When the patterns cannot be "drawn" from their molds (as was the case in cannon casting), then they must be melted or burned out ("lost wax" method) for smaller objects, or broken up in pieces internally, and removed bit by bit - as was the case with the clay cannon patterns.

    Middlebrook says that at Salisbury the molds were two-part for removal of the pattern. His putative "two-part" mold could only have a parting plane lengthwise of the barrel (since a barrel pattern with numerous undercuts, encircling raised rings, protrusions, or muzzle "swell", could not "draw" endwise from a mold). What may have confirmed him and other secondary authors and illustrators in this erroneous idea, I will shortly indicate. Referring to the metal straps around the exterior of the mold (visible in the Diderot drawings from contemporary French sources), he says they were to keep the mold halves together during pouring.

    Let us now consider a documented (Jackson, 1973) method by which the English made cannon patterns at this time. First, they made three separate patterns: one for the barrel, one for the breech end (which included the cascabel), and one for the "head" which included the muzzle, sometimes down to the swell. That the French, too, used separate breech patterns, seems confirmed by the depiction of a breech mold, clearly shown in the lower right corner of (Fig. 7-B: Though a critical illustration to my thesis here, this original not returned by publication). It is my hunch that the vertical cutaway of an assembled cannon mold shown at the left in this same illustration, is the source of Middlebrook's confusions. I believe, that despite the fact he was trained engineer, he interpreted this cutaway as an actual half of a two-piece, vertical split mold! Harte and later authors have perpetuated his error.

    Assuredly, the French draftsman meant only to show relation of various components of the mold in his hypothesized sectional view.... Note the "head" portion in this "mold" extending beyond the muzzle (which ends just above the letter "M"). Interestingly, no separate "head" mold is shown (as is the breech mold noted previously), and the "head" shape here closely duplicates that of the pattern previously analyzed in Fig. 5. I suggest the lettered details shown in the lower half of Fig. 7 are actually the very pattern shown in Fig. 5... Perhaps integral rather than separate "head" sections were more common among Continental founders.

    Perpetuation of the "split-mold fallacy" for cannon founding at Salisbury, explains a detail in the modern recreation Fig. 1-B, where a group of pristine "mold halves" is shown lying adjacent the casting floor or tubb. That such longitudinally split mold halves never existed at this or any other contemporary cannon foundry, I hope to have shown.

    Patterns for swivel guns cast at Salisbury would have been made in a fashion similar to that for the larger bores, save that the smaller swivel guns or "murderers" as they were sometimes called, would permit a bench operation. As shown in the accompanying illustration (Fig. 8: This original, too, not returned by publication, but I have another illustration here that gives the idea pretty well.. "Murderers" This involved a mandrel turned by an apprentice. The patternmaker built his clay up on the rotating mandrel, and then used wood or metal templates ("strickle boards") which he pressed against the form as it rotated. This contoured the soft clay and additional smoothing and forming of the exterior barrel form could be done by hand. Molds for casting swivels would have been made similar to molds for the large barrels, next to be described.

  Mold Making

    When the pattern was completed and thoroughly dry it could be handled and moved with care. By means of the overhanging mandrel ends, it could be supported on cross timbers or cradles. The next step was to make the shell or mold. Graphite and carbon mixes were painted on the dried pattern surface to seal it before the clay mold was started upon it. Steps in making a cannon mold are shown in Fig. 6. "Turning bars" on the ends of the overhanging mandrel facilitated rotary sculpting of the damp clay with contour boards and other tools. . English moldmakers sometimes wound strips of coarse, hempen cloth into the mold clay to strengthen it. The outside of the tubular, open-ended mold wasfurther strengthened against the rigors of future rough handling by the casters, with a lattice-work of iron bands which completely encircled it. This iron "cage" is clearly visible in Fig. 7.

This mold, too, was carefully dried over "ember pits" as illustrated in Diderot (Fig. 7), or in the English foundries, over coals retained within low brick-walled fire boxes - often with ancillary supports permitting slow turning of the drying molds, somewhat like roasts upon a cooking spit... When judged cured by the master molder, the pattern within was removed in the following fashion. (Remember, the mold, like the pattern inside, is at this stage open on both ends.

    First, the model-and-mold assembly was removed from over the drying fires, and the tapering, thin, protruding end of the mandrel was given a sharp rap with a sledge (Fig. 9). Once started, the tapered mandrel was easily withdrawn.  Temporary skewers which had held the wax embellishments in place while the clay shell mold was being built up, had been previously removed, so the mandrel just slid out from under them - leaving them still invested in the thick clay shell.





























   Next, the ends of the straw rope (really a sort of core to both pattern and model) were severed, and it was unwound, spiraling out endways from within the mold. (Fig. 10: Illustration no longer at hand). One authority (Crossley, 1975) says at this juncture coals were sometimes introduced into this space, igniting any remaining straw rope, and heating the pattern. This heat would help shrink the pattern clay, causing it to pull away from the mold walls. If grease had been used as a separator, it would begin to melt. The same was true for any invested wax fillets or wax pattern components. Care was taken not to heat the surrounding mold. Iron or wooden rods were used to break up pieces of the pattern and push it out the ends of the mold. (This is why rigid wood components would be unsatisfactory as pattern components). An illustration from Woolwich (Fig. 11: no longer extant) shows a barrel mold propped with one end on a saw horse, and the other on the foundry floor, where a small fire has been kindled to waft additional heat up the mold tube and melt out wax and grease.

    When the pattern had been completely removed from the mold, the interior mold walls were painted by means of a long-handled brush with compounds of lead, soot, and other ingredients to seal them, and to facilitate separation from the later casting. In similar fashion, the breech end and "head" molds were made from their respective patterns. At this juncture, the trunnion cavity was integral with the barrel mold, but the outer trunnion ends were still open. They were later closed by specially ground, perfectly flat tiles at the time the mold was set up in the casting pit, next to be described.

Casting

    FOLLOWING PARA. SUBMITTED AFTER INITIAL SUBMISSION... ( The finished, fire-hardened molds were next conveyed to the cast-house floor (at Salisbury, likely under an ephemeral shed roof). Here, at no great distance in front of the furnace was the casting-pit, sometimes tubb, and the molds were set up vertically in this empty pit. In French founderies, barrel molds were sometimes baked or fired separately after being lowered into the casting pit, and prior to emplacement of the sand fill. English practice seems to have stressed firing more as an adjunct to mold-room operations.)

    In roughly contemporary English furnaces, (Crossley, 1975, 1979), where careful archeological procedures have documented the finds, these casting pits casting pits were about 10 to 12 feet in diameter, and perhaps 8 to 10 feet deep. Often they had oak-plank floors, and were walled with oaken staves, close fit along edges to keep moisture out. One pit was double-floored; in another were remains of a mold table, with sets of different legs to permit height adjustment to gun molds of different lengths. Pipe fragments in the pits were interpreted as part of a water removal system.

    Since the gun-casting pits in these furnaces were often not far from the water-filled wheel races, the operators were always concerned to keep them dry. At the Royal Arsenal, a compound casting pit was in use, nicely faced with brick and stone. Stone kerbs or rims protected the upper edges of the pits where molten metal crossed over coming from the furnace. Adjacent to the pits were often planked floors at various heights which were beneficial when removing the heavy castings by block-and-falls.

    To set up the molds, pitmen would first carefully tamp down a layer of prepared earth in the bottom of the pit. Upon this they would place a breech mold, open-end up, well-bedded in the earth. Next, other foundrymen would swing the barrel mold, slung in block-and-falls, over the pit, breech-end down. In the pit, the pitmen would carefully stretch a white cloth taut over the breech mold opening and fasten it in place temporarily against the inside walls of the mold. Upon the center of this cloth they placed a lighted candle, whose function will soon be seen (Fig. 12: this extremely interesting illustration no longer at hand).

    A length of string was dropped vertically down the inside of the barrel mold by the crew at the top of the pit and the pitmen made the string fast to the white cloth. Then the barrel mold was carefully lowered into place. By virtue of the lighted candle in the exact center of the breech mold and reflected light from the cloth, the pitmen could clearly see to guide the barrel pattern into place and mate it to the breech mold! This clever expedient might aptly be termed an 18th century "optical centering device"! It is illustrated also in French manuals of the Period (Fig. 13: not on hand).

    THIS PARAGRAPH IS A POST-PUBLICATION INSERTION TO HELP MAKE SENSE OF THE DESCRIPTION, AS WITHOUT THE MISSING ILLUSTRATIONS, IT IS HARDER TO VISUALIZE HOW THIS WORKED...( A candle snuffer mounted on a long pole was then inserted down the barrel mold to extinguish the candle still burning inside the closed mold. After the mold had been placed in the casting pit, and the sand fill had been packed around it up even to its open muzzle, a long pole was thrust down the length of the barrel mold and pushed against the still-taut linen kerchief above the cascabel section. By means of the still-attached string, the kerchief was collapsed "parachute-fashion" around the tip of the pole and withdrawn up the mold, bringing with it the candle stub, and any dirt or other material which might have accidentally fallen down the barrel mold during the tamping and filling operations.)

    The entire casting crew next mustered to tamp down the rest of the casting-pit fill. Since the fill was damp to facilitate packing, and the molds were bone-dry, it was necessary to complete this operation as soon as possible so moisture did not seep into the molds. As mentioned, when the fill level rose to the height of the trunnions, special tiles were set in their openings and filling continued. The "head" mold, if there was a separate one, was attached last and the pit filled level to ground level or level of the mold sprues.

 

(Continued)