Tooling Development
Having arrived at the decision that the hull would be of sandwich construction and would be fabricated in an upright female mold, attention was directed at the determination of the best mold fabrication method consistent with reasonable cost, surface quality, and the limited number of hulls to be molded.
Early thinking involved a steel female mold, externally reinforced and with provisions for heating the mold by means of steam or hot water. Heating assemblies would also have to be provided inside the mold in order to cure the inside of the layup because of the insulating effect of the hull core material.
The apparent difficulty of fabricating such a mold to the desired accuracy and at reasonable cost led to the investigation of other possible construction methods.
The procedure eventually decided upon was to fabricate a wood and plaster male plug upon which a reinforced-plastic mold shell with a sprayed-zinc surface could be laid up. A suitable external steel structure would be secured to the mold shell to provide rigidity and permanence of shape to the finished mold. Built-in heating provisions would be no longer required since exclusive use of room temperature curing resins was now contemplated.
Before beginning the mold itself, it was thought best to try a smaller section of the hull in order to determine whether any unforeseen difficulties were likely to be encountered. A half section of the bow having an area of about 50 square feet was selected as a suitable trial unit and a wood male plug structure was lofted, assembled, and surfaced with metal lath and plaster. The surface was worked fair and smooth and coated with a proprietary sealer, Renite RP-802, followed by several coats of paste wax and a coat of parting film (Marco ML-4). A film of zinc was flame-sprayed on the plug surface to a thickness of about .010 inch and a laminate consisting of 4 plies of U.M. style #4044 cloth and 5 plies of 1½ oz. mat laid up directly on the sprayed metal. After the laminate had cured, the female mold shell was easily removed from the plug and the metal surface was found to be securely bonded to the laminate. The resulting mold was deemed satisfactory and its fabrication gave no indication that any serious difficulties were likely to be encountered in the fabrication of the full-sized male plug and female mold. It was therefore decided to proceed with their construction.
Mold plug frames were lofted 1" inside the lines of the hull and were constructed of 2" dimension lumber. These frames were erected on 24" centers and braced to a suitable foundation.
A keel strip, provided with an aluminum divider plate was secured along the center line and the plug frames were further tied together by 3/4" × 3" longitudinal stringers let in and secured to the frames on 12" centers.
Metal lath was then nailed to the frames and stringers and 3/4" grounds were screwed along stringers over the metal lath. A base coat of wood-pulp-filled plaster was applied to the metal lath and screeded off fair with the grounds. The grounds were later removed and the resulting voids filled with plaster. After the surface of the base coat had been roughened, a 1/8" coat of unfilled plaster was applied as a finish coat. The resulting surface was sanded and patched as necessary to fair and was checked with loft templates. When the plug's plaster surface was true, fair, and smooth, it was sealed by application of three coats of a clear lacquer and given three coats of Johnson's Paste Wax and two coats of Marco's ML-4 parting film.
The female mold was designed in three sections - port and starboard sides and transom - to facilitate parting of the mold and plug as well as the subsequent removal of the hull from the mold.
The first operation in the fabrication of the female mold proper was the application of a film of zinc over the parting film on the plug. The zinc was sprayed in a continuous operation and to a thickness of about .013 inch. A total of 612 pounds was applied.
One complication which arose during the metal spraying was a tendency for the blast from the metallizing gun to remove the parting film on the plug, particularly at inside corners where the blast would be confined. It was necessary to repair the parting film at several points, but as far as the overall picture was concerned, the delay was slight.
The mold shell laminate was laid up directly on the completed zinc film. This consisted of a shingled layup of 6 plies of 1044 cloth followed by 3 plies of Bigelow-Sanford 2 oz mat, butted, and a finish ply of 1044 cloth. These layups were done in a three-shift operation, one side of the mold at a time. The mold transom had been done first as a further check on procedure suitability.
The external steel reinforcing and supporting structure for the mold shell was fabricated as a unit and when mold shell layup operations were complete and cured, the steel structure was placed in position and plastic bounding angles were laid up, connecting the steel webs to the plastic mold shell. A number of plate longitudinals had been incorporated in the structure, and these were also tied to the mold shell by means of plastic bounding angles. Bolts were used to further secure the bounding angles to the steelwork. The resulting mold sections were then removed from the male plug and inspected.
A number of small imperfections in the mold surface were found to have resulted from air bubbles trapped under the mold shell laminate. The zinc would at first hide these voids, but polishing the mold surface would cause the relatively thin zinc surface to collapse into the air bubble. These voids were filled with an aluminum-filled auto body cold solder and sanded.
In the sanding of the patching material, the zinc was removed in some small areas so that the laminate itself became the mold surface. In all cases the surface in these areas was as smooth as that of the zinc. If a great deal of polishing of such a zinc surface is contemplated, it appears that a thicker film should be applied.
There were also a number of very shallow transverse wrinkle marks in the mold surface, probably caused by the longitudinal shrinkage of the mold shell laminate. These imperfections were also filled with cold solder, although their effect on the hull surface would have been negligible, affecting only the gel coat. It appears that these flaws could be eliminated either by applying a thicker coat of spray metal or by applying the shell laminate in smaller increments and thereby eliminating the simultaneous curing and shrinkage of the laminate as a whole.
As soon as the mold assembly had been erected and patching and polishing completed, a number of small test layups were made to determine parting qualities. It was found that the lacquer-based cold solders used in patching had an inhibiting effect on the resin cure due possibly to the effect of residual solvents which were able to act through the parting film. An aluminum-filled epoxy as a patching compound would probably have given better results.
After the mold had been erected and leveled in the fabrication area and prior to the beginning of the layup of the hull proper, the mold was provided with staging around the main and raised deck sheer lines. Steel rails were erected along each side of the mold at main and raised deck levels and cars were built which could be moved fore and aft along the rails. These cars in turn were provided with transverse rails upon which a second smaller car was placed. This allowed movement either transversely or longitudinally. The smaller car was equipped with attached staging which could be raised or lowered as required by the nature of the layup. One such car assembly was provided for the main body of the hull and one for the raised deck portion. Adjustment of the staging was accomplished by use of threaded hand wheels for vertical adjustment and by pushing the cars for longitudinal and transverse movement. See photograph 45. In a program involving a number of hulls from a female mold it would be well worth the cost to have the staging motorized, particularly for vertical adjustment.
The matter of locating reference lines within the mold was given considerable attention and the fact that any layup in the mold would obscure any marks made on the mold surface, contributed to the difficulty of checking locations as work progressed.
Frame locations and other transverse reference locations could be marked on the mold sheer flanges and transferred inside the mold with taut wires and plumb bobs or by other standard procedures.
A relatively simple method for checking the C/L plane regardless of the amount of layup in the mold was devised. It consisted of stringing two taut wires above the mold, one above the other and separated as much as the available space would allow. These wires then defined the plane of the mold or hull C/L. To find the C/L at the mold or layup surface, a small mirror was placed in the area of the C/L and the wires sighted through the mirror and the line of sight adjusted until both wires appeared as one line. The position of the image of the registered wires on the mirror would then coincide with the actual location of the C/L plane at the mold surface. An index mark on the mirror made it possible to hold the C/L once located, and a mark could be made on the layup. The accuracy of this method was, of course, dependent on the maintenance of a level mold assembly and a rigid structure supporting the wires. In practice, the idea worked out satisfactorily and could be used over the entire length of the mold, requiring only moving of the mirror.
Once the C/L was established, other longitudinal reference lines could be located by means of scales, straight edges and spirit levels.
Bulkheads and flats were to be fabricated on plywood molding tables provided for the purpose.
For fabrication of girder sections, tables equipped with mold strips for forming the girder flanges were provided. Since it had been noted during certain practice layups that right angle flanges tended to close up slightly upon curing, some sample girder flanges were laid up to determine the extent of the distortion.
Results were conclusive and sufficiently uniform that a wedge shaped strip was provided for use with the flange molds which would correct for this tendency. As a result of this precaution, all finished girder flange angles were very close to the desired 90°.
The initial concept of deck fabrication called for the molding of 5 separate deck sections and fastening them together and to the hull. Variations in sheer throughout the length of the hull would require that each deck section be fabricated on a separate mold. Fitting these separate sections together would doubtless be complicated by warpages during curing and while possible, would certainly be expensive. Efforts were directed toward finding an easier way.
The process as developed called for fabricating the main deck in one piece on a plywood mold platen fixed to the deck beams and to the hull. The process is detailed further in the section on production of ExMSB-23.