Jump to content

JSGerson

NRG Member
  • Posts

    2,482
  • Joined

  • Last visited

Posts posted by JSGerson

  1. Grating and Seats of Ease

    The kit plans called for pairs of 1/32” x 1/16” rail support beams on either side of the rail supports. Accordioning to the plans both the kit and the US Navy, these cross beams are flush with the top of the rail supports. They support the gratings upon which the seat of ease rest. However, either I or Mr. Hunt misinterpreted the plans because at the end of Chapter 8.1 he states: “I chose not to model these as it's nearly impossible to see them.” See them or not, they are important as I interpret the plans. As result, his construct of the gratings and seats of ease, are lower than the cross beams had he installed them, while mine are resting top of cross beams. Be that as it may, here is what I did.

     

    Following the practicum, I cut the seat of ease side pieces from 1/16” basswood stock and assembled them using the kit’s plan as my template. The kit actually called for 1/64” thick pieces to be used, but since their thickness would not be seen by the viewer, the thicker pieces were more stable and easier to handle. The kit provided three laser-cut pieces of the grating for each side, which I temporarily taped together for dry fit purposes.

    IMG_2287.JPG

    IMG_2288.JPG

    IMG_2289.JPG

  2. I assume the boomkins on the actual ship had some sort of fasteners to hold the beams in place to the hull, but I could not find any information on that. Therefore, they were just glued into place. The boomkins did have two metal rod hull braces each. The kit called for 0.023” wire for these braces. I needed something stiff, so brass and copper wire were not suitable, but I did have “0.020” and 0.026” music wire. I chose to use 0.026” wire in lieu of 0.023”, which I did not have. Holes were drilled into the hull and boomkins in the appropriate places following the kit plan with a #72 drill bit. After a little trial and error, the wire s were sized for length and the wire ends bent for the appropriate holes.

    IMG_2302.JPG

    IMG_2303.JPG

  3. Looking at my efforts, I realized I needed to do this all over again but with a harder wood because the square edges of the boomkins were already beginning to wear down during the fabrication process. Using 1/8” x 1/8” stock pear wood that I had left over from when I constructed my Rattlesnake, the results were much better. They were then painted black. The brass bands held three eyebolts. Two eyebolts, one on the aft side and the other on the bottom of the boomkin, were aligned lengthwise following the line of the boomkin. The third eyebolt located on the top of the boomkin was installed perpendicular to the boomkin as was shown in the various photographs of the actual ship. To do this, a pin was used to make the initial dent in the brass for the #77 drill bit. The holes were drilled through the brass into the wood. After each hole was drilled, a full-size eyebolt was inserted so the brass band would not move while the next hole was being drilled. Then the eyebolts were removed, cut to size, and permanently installed with CA glue. This secured the brass bands as well.

    Boomkin – PearWood.png

  4. Bow Boomkins

    I was about to install the grating and fabricate the seats of ease in the bow when I realized that the pair of bow boomkins needed to be fabricated and installed first because they were physically below the grating. Surprisingly, the practicum did not address these components.

     

    Initially, utilizing the kit’s 1/8” x 1/8” stock basswood, the boomkin were fabricated by tapering the pieces of wood to 3/32” x 3/32” at one end. Then, with a square cross-section needle file, the triangular indents on the tapered ends were formed. Using 0.005” brass metal sheet, two strips of the metal were cut to size, wrapped around the boomkins to form the brass bands, and then blackened.

    Boomkin - Boxwood.png

  5. As I mentioned in an earlier posting, the MS plan drawings of the rail supports do not fit properly on my model. Whether I’m at fault or the plans are, is immaterial. These have to be custom made now., By trial and error measuring using card stock, a first try template was drawn and cut out of the cardstock. Once I got the card stock to fit, it was traced onto 1/8” thick stock wood. Once I got the wooden piece to fit, I realized that it had to be refined a bit more. A second wood rail support was fabricated. This was able to fit properly onto the rails. The support was removed, painted and pin striped. Finally, it was installed. This was repeated the same manner (with or without an extra preliminary wood support as needed) for the remaining nine supports.

    IMG_2281.JPG

    IMG_2282.JPG

    IMG_2283.JPG

    IMG_2284.JPG

    IMG_2285.JPG

    IMG_2286.JPG

  6. This is when I discover a new problem. According to the MS plans, the precut gammoning holes in the stem are supposed to be between supports Nos. 2 and 3. Per the marks I just made with the gratings, the gammoning holes straddle rail support No. 3. The gammoning holes are where the gammoning chains pass through. Looking at the US Navy 1927 -31 Restoration plan No. 24779, the gammoning holes have been replaced with eye bolts embedded into the stem. The gammoning chains are then connected to the eye bolts. The chains are fastened to the bowsprit in the same manner as shown in the MS plans. These stem bolts are positioned half the distances apart as the pre-cut holes are and appear to be straddling rail support No. 3 from the numerous photos that I’ve looked at. Based on the all the ambiguities, just about nobody will even know about or bother to look inside the rails supports to see if there is even any gammoning there, let alone whether it is correct or not. Therefore, I am going to use the existing gammoning holes as they exist on my model in lieu of the embedded bolts shown on the US Navy plans, just because as builder of the model, I like the way they look.

    24779001.jpg

    IMG_2277.JPG

  7. Now it’s time to install the rail supports. These are positioned such that the joints of the floor grating in the seats of ease area, rest on the horizontal structural beams attached to the rail supports. In other words. Where the gratings meet, that is where the rail support is supposed to be. Using the grating parts supplied in the MS kit, I taped the three pieces together that make up most of the flooring, to mark off where the joints are with the architect tape on the stem.

    IMG_2275.JPG

    IMG_2276.JPG

  8. I don't know how much "easier" building a cross section is than the full ship. The cross section has a lot more detail below decks. You might want to look at Tom Culp's build of the Model Shipway' Cross Section of the Constitution.  In the mean time, I've attached a number of US Navy plans showing the cross section. You can find more plans with other details at the US Constitution Museum website.

    24472001 - Midship Section.jpg

    Structural Views.jpg

    130241 - Midship Section sh1.pdf 34535001.pdf

  9. BTY, I’m not knocking Mr. Hunt’s practicum. I could not have gotten this far if it weren’t for it. But I have learned that he’s human and he's the first to admit that his way may not be the best which is why in addition to using his practicum as a guide (not my bible), I also check how other builders solve these problems.

     

    My second fabrication attempt at the 2nd rails went relatively smoothly; and they were glued into placed. In the first image below, I attempted to show the slight curve in the rails.

    IMG_2269.JPG

    IMG_2270.JPG

    IMG_2271.JPG

    IMG_2272.JPG

    IMG_2273.JPG

    IMG_2274.JPG

  10. The next thing the practicum stated about the rail was, “It is not curved from side to side.” Wrong again. According to the MS plan, the second rail has a slight horizontal bow inward towards the stem, a much simpler curve than that of the 3rd rail. But I didn’t realize that until I re-read xKen’s build log (starting at post #294) of how he made his rails where he accounted for the inward bow. To double check, I looked again at the US Navy plan and saw that the bow was actually a slight S-curve. To be fair, the MS plan didn’t show the full length of the 2nd rail so you couldn’t see that it swung back slightly to form the S-curve. Unfortunately, I had already made the pair of 2nd rails. I reluctantly abandoned my first attempt and started over again this time using 1/8” layered stock (1/4” total thickness) to account for the S-curve.

    2nd Rail From Top - Plan View.png

  11. However, using the 2nd rails to dry-fit the rail supports, I soon found out that the rail supports did not fit properly between the rail and the stem. Then, reading section 8.1.4 in the practicum about installing the “2nd rail from the top” I discovered that Mr. Hunt also ran into this problem. He believed through his investigation, that the MS plans were in error for the shape of the rail supports. I then checked the US Navy plans and there were only very slight differences that I could discern, so both Mr. Hunt and I don’t know where or what error(s) occurred. His solution was to fabricate and install the 2nd rail first, then custom fabricate and fit the rail supports then, work backwards to install the 3rd rails last. Because I only glued the 3rd rail to the starboard side at just one point, it was relatively easily to pry the rail off to continue following the practicum’s solution.

     

    Going back over the practicum instructions, I noticed the practicum’s instruction for fabricating and installing the 2nd rail weren’t correct. The first error was to use the MS plan profile (elevation) view of the rail directly from the MS plan. Unfortunately, this is a foreshortened view because the rails are angled from the hull inwards towards the tip of the bow stem as indicated in the plan view. (I had this same foreshortening problem when I constructed the transom) To get the true view, that is the true length of the rails, I scanned the elevation view of the rail into my computer, using PowerPoint, rotated the image to match the angle of the rail in the plan view, then stretched the image till it matched the length shown in the plan view. Only then could I print the image with the proper rail length and make a template.

    2nd Rail From Top - True View.png

  12. The practicum stated:

    Quote

    The second rail from the top is made from 1/8" stock. This rail is fairly simple and straight forward. It is not curved from side to side. So, the only template we need is one of the profile of the rail, which we can take directly from the profile drawing.

    Great, easy straight forward pieces to make…or so I thought. So, I proceeded and made the 2nd rails as instructed, complete with paint and pin stripes,

    IMG_2268.JPG

  13. The next logical step seemed to be installing “3rd rail from the top. First, I used PVA glue to fasten the starboard side rail just to the underside of the cathead only. I left the stem tip end unglued to allow me some wiggle room when adding the rail supports. I anticipated that the rail supports had to be at least fitted with the rail installation. Therefore, the rail supports (port and starboard) were fabricated based on MS plan detail 4A at the same time as the 3rd rail.

    IMG_2266.JPG

  14. What the first diagram is try to convey is, as you apply the copper plates to the hull, you start at the stern and work your way forward and upwards overlapping the edges of of the previous aft plate as well as the plate below's top edge.

     

    The second instruction is trying to state that because the plates are applied to the hull with no tapering unlike the wooden planks which are tapered. Therefore, sharp curves are going to be created at either or both the stern and bow. In order to minimize those curves a steeler, an extra row for a short length is added. The BlueJacket instruction is telling where a steeler should be placed. At each band, you in effect start a new horizontal uncurved row. The higher band cuts off the curving from the lower band. The diagram below illustrates a coppering method. I hope this helps and not confuse you more.

     

    Jon

    Coppering.jpg

  15. If you use the glue blob method, you can wipe off the glue if you notice the error early enough or use an Xacto knife and just pop off the dried glue. I used (0.6mm punch) with brass punch-out's. I used Wipe-on-Poly as the adhesive.  While still wet, I  had time to maneuvered the punch-outs into position.  It dried fairly quickly and adhered the punch-out to the surface. Drilling is permanent. I wouldn't worry about out of position bolt heads. Once the hull is painted, you will be hard pressed to see them, let alone be able to identify which are out of position.

     

    I deliberately used this phenomenon when I copper plated the hull. Most of the builders who decided to emboss the copper plates to simulated nail dents either went around the plate edge with a ponce wheel or made a pattern embossing stamp. This resulted in what I felt were out of scale dents on the plates for these model scales. I embossed the plates with random indentations made by pressing the plates with coarse sandpaper for the face of the plate and adding edge markings from the straight line teeth of a fine miter saw. From three inches back you could not see if my "nail" indentations were random or patterned. What it did do was remove the flat shiny surface. The effect worked great.

     

    Jon

  16. You know, I never noticed there was a subtle difference in the pattern. Oh well, it's too late for me but after a coat of black paint, you can barely see them at all let alone discern the difference in the two decks. That, and 99.9% of the people who actual take the time to look at these models would not have the intimate knowledge to recognize that detail. I believe it's because the spar deck has cannonades and the gun deck has, well, guns. Still, if I was a aware of it earlier, I would have done it more accurately. Good eye!

     

    Jon

×
×
  • Create New...