Dr PR

This all started on Christmas of 2004 when my oldest son presented me with a 8"x8"x72" present. "They got me new skis?" I wondered. No, it was a 1:96 scale Cleveland class fiberglass hull, and one deck plan drawing. I began work on it immediately and soon realized I needed more information. I had 281 high resolution photos I took while I was on board. I found a few hundred more on line, but these were all very low resolution. Over the years I have received a few hundred more high resolution pictures from people who contacted me through my web site, plus a dozen or more from Navy archives. I also studied pictures in the ship's cruise book, and eventually found good photos of almost all of the ship's exterior. I wanted to make an accurate model worthy of a place in a museum. I have seen pictures of several so called "Oklahoma City" models and they were all highly inaccurate. For an accurate model I needed dimensioned blueprints. I found these in the US National Archives, and that was an adventure all in itself! I live in Oregon, and it would be very inconvenient tripping off to College Park, Maryland, every time I needed another drawing - especially with all the TSA hassles and the expense of flying across the continent. So I relied upon email contacts with the Archives. When you email the archives you should expect a 3-6 week delay before you receive an answer - there are a lot of people contacting the archives and only a few Archivist to answer questions. These people cannot read your minds, so you have to be very specific about what you want to find. Ask about only one ship at a time, and give it's name and hull number, and describe what type of drawings you are looking for. Some Archivists really know little about ships, and can give only general answers. I lucked out on several occasions and worked with people who had good knowledge of Record Group 19 where ships' blueprints are kept. I got a lot of good information from these people. With a little patience (and about $1500) I eventually obtained 43 reels of 35mm microfilm with 7700 blueprints for the Cleveland class and another 2200 for the CLG Talos conversions. Everything I wanted to know (almost) and a lot more! Note: Back in 2004 my only choice was microfilm reels. Today you can order scanned images on DVDs. Don't even think of trying to transfer bazillions of bytes of images via the Internet. When you look for microfilm at the Archives you will find that each drawing set, such as the original USS Cleveland CL-55 drawings or the CLG Talos drawings, is listed as being on a "reel." But these "reels" are actually sets of several individual physical reels of microfilm. The Cleveland drawings are "reel" 5537, and there are 19 physical reels of 35mm microfilm. Each reel has an index of the contents of the reel. Later the Navy changed the way microfilm was archived and provided a single Index reel for each set of drawings. It is a lot cheaper to get just the Index reel (if it exists) to see if there are drawings you might want, rather than buying the entire set and then discovering it is only wiring diagrams! The next step was to compile a complete printed index for the reels of microfilm. I scanned the index reels into Photoshop at our local library, and printed them. They make a stack about 6" high, single sided print, with about 25 single line entries for individual blueprints on each page. The problem was that most of the drawings were for internal wiring, plumbing, ventilation, and many frames listed furniture, doors, keys, and even the brass plates attached over each door. Not very useful for ship modelling! But I eventually did scan about 400 blueprints into Photoshop. The detail on most of them is amazing. They used very fine grain film, and I can see the tiny pencil dots the draftsmen used to align rows of text and the centers of circles! But getting these images was a chore! The microfilm scanners at our library are fine for scanning images of pages of newsprint or typed documents. But a maximum resolution of 800 dpi was not good enough to scan a full film frame and be able to read the fine print. So I zoomed in and scanned each frame on the film with six overlapping images. For the largest drawings (36" high and 144" long) I had 36 separate images to paste together in Photoshop. Each frame on the microfilm had a bit of spherical (pincushion) distortion, and each scan from the microfilm scanner added even more distortion. I had to rotate, scale and "warp" each image to straighten things out and get them all to align. It could take an entire day to paste together one large drawing. But the resulting drawings show the tiniest details and the smallest print is very legible. Are you beginning to understand why this has taken 14 years (so far)? But a few key drawings were poorly exposed and almost illegible. Even the best scanner tricks and Photoshop magic couldn't resurrect some of them. Fortunately, some of the important drawings were repeated on two or more drawing sets on different microfilm reels, and I was able to find good drawings of most essential parts. Another problem was missing drawings. Each blueprint has a reference list naming other relevant drawings. But a few key drawings were missing from the microfilm. These might possibly be in collections of paper drawing sheets at the National Archives, but I will probably never know. There is one other thing you should know about blueprints for ships - especially from the world War II period. The Cleveland class ships were built in four different shipyards. One yard was the initial builder, and this is where the original blueprints were hand drawn - pencil on paper. Each drawing was hand copied - traced with pencil on vellum - to produce another set for each of the other yards building the ships. But each yard had its own way of doing things, so they often made new copies with changes. And while the ships were being constructed the Navy promulgated additional changes for each yard based upon past experience building and operating the ships. There were some very noticeable differences between ships of the same class built in different yards. The result is several different versions of the same blueprint produced in chronological order. You need to know which yard your ship was built in and when. I started out thinking it would be a simple job building the model after I got the blueprints. I am much wiser now! Phil

I have been working on a 3D CAD model of the USS Oklahoma City CLG-5 as configured in 1971 since 2004. It is just about complete now. I will try to describe the steps I followed to make a very accurate 1:1 scale model, including the problems I encountered and the research (which took most of the time) necessary. I am a former Lieutenant in the US Naval Reserve, and I was the Nuclear/Special Weapons Officer on the Okie Boat from January 1970 through March of 1972, hence my interest in the ship. If you want to know more about the ship, it's history and a lot more, have a look at my web page at https://www.okieboat.com/index.html **** To summarize, the OK City was commissioned in 1944, the 20th of 27 Cleveland class light cruisers. It entered the war in mid 1945 and earned two battle stars, one at Okinawa and one for attacks on the Japanese homeland. After the war it was mothballed, but was chosen in the 1950s to be converted into one of the first guided missile cruisers. It was to carry the Talos surface to air missile. Talos design was changed several times between 1945 and 1955, from a range of 20 nmi, to 60 nmi, and eventually 130 nmi. Meanwhile a Talos test missile was developed into the Terrier missile, and Terrier was the first missile to enter the fleet. The changes to Talos delayed introduction to the fleet until 1958. The Oklahoma City was recommissioned as CLG-5 in 1960. It was a fleet flagship, and served as First and Seventh Fleet flagships for most of it's 19 year service. It was Flagship of the Seventh Fleet for most of the Vietnam War, where it earned 13 more battle stars, becoming the most decorated of all the Cleveland class ships. The OK City was the first ship in the US Navy to use a surface to surface missile in combat successfully, and the first to use an anti-radiation (ARM) missile against an enemy radar successfully. It was decommissioned in 1979 and expended as a target ship in 1999. It was one of the US Navy's historic ships. **** In future posts I will describe the steps I took to research the design of the ship and how I proceeded to make the CAD model. Bear with me though. I am fairly busy with other things and there is a lot to tell. Phil

Dan, Excellent project! Your method of plating the hull caught my interest. I will soon be working to plate a 1:96 scale hull for a Cleveland class cruiser. I am fortunate to have the blueprints (hull plate thickness) and 81 pages of mold loft offsets that give the shell sight edges (positions of the edges of each hull plate) to work from. But I have been wondering how I would make the individual plates and all the rivets - very similar to the Leviathan's hull construction, including the backing plates with lots of rivets. I think your choice of styrene is a good one, but I may make the upper strake with 0.003" and 0.005" brass because it rises above the main deck edge about 0.050" in places and I'm afraid very thin styrene would be too fragile. Did you rely on the adhesive backing on the copper strips (backing plates) or did you use another glue? I have used the copper strips before and I am not confident that the adhesive will remain sticky years down the line. I'll be following your build. Phil

Jim, Welcome from another newbie to the forum - also from Oregon (Corvallis). I have seen some beautiful models of small boats in large scale. You have the opportunity to build a lot of fine detail - like some model airplane builders put into their planes. If (when) you decide to start scratch building the original construction blueprints for US Navy small boats (and some larger auxiliaries) are available on line from the Barbour Boat Works Inc. records (#758) at the J. Y. Joyner Library at East Carolina University, Greenville, North Carolina, USA. https://digital.lib.ecu.edu/11208 I accidentally stumbled upon this source and not a lot of people know about it, so I mention it to everyone interested in small boats. Looking forward to posts of your builds. Phil

I have been working on a CAD model of the USS Oklahoma City CLG-5 for 14 years now, and it is just about finished. When I have time I will start a build log on MSW (unfortunately, there is no place in the gallery for CAD models). I have been posting on The Ship Model Forum and on my web site https://www.okieboat.com/. Doing the actual work to build a 3D CAD model is faster than creating physical parts - if you are an experienced CAD driver. For one, you don't have the problem of doing the work to make a part only to learn later that it is not correct and you have to build another. It is very easy and fast to make changes to the existing model, and you don't have the expense of wasted materials (just wasted electricity). But if you factor in the learning time for the CAD program it is much longer. I have been using CAD programs since 1988 so doing the work is easy for me. It would be much more difficult for a novice. **** The best reason for making a 3D model is that you can be sure the parts fit together correctly when you get around to making a physical model. I started on a 1:96 scale real model of the ship and soon ran into a number of problems that I could not figure out from the original blueprints. In one case I finally realized that the original 2D blueprints had to be wrong! The parts of the light mast above the bridge could not fit together in three dimensions the way the designer had drawn them in 2D. The diagonal support members for the mast passed through supports for a foot rail on the mast (see attached picture) . I decided to make a 3D CAD model of the parts to see how the parts fit together and confirmed the problem. Fortunately I had many high resolution photos and I could see how the shipyard workers got around the problem. They just attached the diagonal support members at a lower position on the mast, with the foot rail supports above them. Using "photoguestimation" I calculated how far below the attachment point shown in the blueprints they had actually constructed the mast on the ship. A similar problem arose when I was trying to figure out how long the main vertical members of a radar tower were (attached picture). The blueprints show the side and front views, but because all parts joined at angles in three dimensions I would have to use three dimensional trigonometry to calculate the actual lengths of the parts from the 2D blueprints (did you ever learn 3D trig?). It was much easier, and faster, to just build the 3D model of the tower framing. And the 3D model showed not only the lengths of all the parts but the actual three axis angles that they fit together, allowing me to construct an accurate jig for putting together the pieces of the real model. I would have wasted a lot of time and materials, and experienced a lot of frustration, trying to piece all of it together one piece at a time with the real model. **** Most of the 14 years that I worked on the CAD model were spent gathering blueprints, plans, data sheets and photos. The OK City underwent many modifications during it's 19 year service as a CLG. Something was changed nearly every time we went into home port. This meant that to build an accurate model I had to pick a specific date to model, and then figure out what modifications had been made by that date to the configuration shown in the blueprints. I also had to rule out modifications shown in later photos that had not been added by the model date. I have 24 pages of notes describing changes over the years. I visited the last Cleveland class ship and CLG, the USS Little Rock CG-4 museum ship in Buffalo, New York. It was the sister ship to the Okie Boat. There I photographed and measured many of the parts of the ship, making dimensioned drawings. I also searched for data sheets and manuals for hundreds of parts on the ship. Research took far longer than the actual building of the model. But it allowed me to create a very accurate 1:1 CAD model of the ship, with details as small as 3/16 inch such as rivets, screws, bolts, nuts, etc. **** The CAD model phase is about finished. Now I can resume work on the 1:96 scale model with confidence as I create each part. And I can use the CAD model to create some parts with 3D printing - although my experience with 3D printed parts has not been encouraging. They are not an accepted material for museum quality models. But they can be used to create molds for lost wax castings. I can also create 2D plans and also drawings for photoetching. Phil

Outstanding CAD work! And it certainly has been a lot of work! Many people think of these CAD models as just a different type of model. However, if done accurately they are digital archaeology. By working at 1:1 scale the CAD model can be a much more accurate rendition of the real ship than a small scale physical model. And the files can serve to further investigations into ship construction techniques and traditions. Phil

Chris (and others), Thanks! I agree that the Clevelands were handsome ships - little brothers of the Baltimore class, and they are well proportioned. However, the CLGs were like the horse that was designed by a committee and came out looking like a camel. After being hacked apart and rebuilt hurriedly as early guided missile ships, with antennas hanging out everywhere, they certainly were a bit of a mess. To make matters worse, the already top heavy Cleveland hull had the massive missile house, flag superstructure and radar towers that made them even more unstable. In my opinion they may be the ugliest ships ever built! But I served on the Okie Boat for 28 months, and that is one reason I want to model it. Also, it was an historic ship, being one of the first guided missile cruisers in the US Navy, the flagship of the 7th Fleet throughout most of the Vietnam War, and the longest career and most decorated ship of all the Cleveland class. The Oklahoma City CLG/CG-5 underwent continuous modifications during the 19 years it was in service, with bits added and bits removed frequently. Some of the whip antennas were repositioned just about every time we went into Yokosuka. This made it difficult to find a specific configuration to model. Finally, after studying many hundreds of photos I chose the configuration of the summer of 1971. This was just before the FAST gear was removed, and about mid way through my time aboard. Someday I hope to create a CAD model of the 1945 configuration. That will be tricky because apparently all of the blueprints for the modified (square bridge) Clevelands have been lost - at least there is no microfilm for the superstructure, and there were major changes from the early round bridge Clevelands. To make it more difficult, no two shipyards built the Clevelands the same way. Each yard redrew the plans and added their own modifications. If you know what to look for you can often identify the yard where a ship was built by looking at the photos. When I find time I intend to start a thread describing the research I have done of the Clevelands in general, and the CLGs and Oklahoma City. Phil

I have been building ship models since I was a kid (about 65 years). I started with plastic models and then began building scratch built models out of balsa. In my 20s (when I could afford it) I graduated to wooden ship model kits. I was fascinated with ships and joined the US Navy, entering Officer Candidate School in 1968. I "retired" as a Lieutenant in 1972. I served on three ships, two minesweepers (USS Cape MSI-2 and USS Ruff MSCO-54) and a cruiser (USS Oklahoma City CLG-5), all having wooden decks. I wanted to model two of these ships, but there are no kits made for them, so scratch building is the only resort. I did nothing for years until Christmas of 2004 when my oldest son presented me with a 8" x 8" x 72" present. "They are giving me new skis?" I wondered. No, it was a 6 1/2 foot long fiberglass hull for a Cleveland class cruiser. The USS Oklahoma City CLG-5 was a converted Cleveland class ship. That started a 14 year odyssey (so far) to research, design and build a 1: 96 scale model. This evolved into a CAD model to create plans for the real model and an extensive web page for the ship, with the ship's history and a history of the modeling progress (https://www.okieboat.com/). I have been posting on The Ship Modeling Forum (http://www.shipmodels.info/mws_forum/index.php) in the Virtual Shop Modeling section (http://www.shipmodels.info/mws_forum/viewtopic.php?f=27&t=70810) for years. I have just discovered this forum. I am a former NRG member, and I am considering restarting my membership now that the Ships in Scale magazine has been acquired by NRG. Phil Hays

Wayne, Right now the model is in four files. I pasted renderings of the four files into one image in Photoshop. I have yet to combine all four - in fact I may not be able to in DesignCAD 3D Max. The largest file I have worked with was about 750 megabytes, but that was several versions back. I do not know if the new version can handle a gigabyte file, although it is 64 bit and should be able to handle it. I will take on that challenge this winter when I can't get out hiking. The four files add up to 1.071 gigabytes. There are a few duplicated reference parts in each file, but after they are removed the total file size will still be about a gigabyte. There are 2.86 million DesignCAD "entities" in the files, but I really don't know what that means in terms of objects drawn in the files. A single object like a cylinder may be composed of many "entities." But I do know that there are a LOT of objects in the files! There are 22 million points in the files. Hull: 220,258 KB Forward superstructure: 312,148 KB Midships superstructure: 234,964 KB Aft superstructure: 303,890 KB Initial render times for each of the four files varies from about 40 minutes to 95 minutes (all surface normals and shadows are calculated during the first rendering). After that new renderings take up to two minutes. Render times are not linear with file size, because the program has to check for shadows on each object that might be cast be all other objects. I expect the initial render time for a gigabyte file will be 4-5 hours. I'll start it before going to bed and it will be ready for subsequent renders the next morning. Note: After the initial render and when working in OpenGL display modes, rendered displays rotate quickly. However, if the file size is several hundred megabytes and all layers are enabled (everything visible), display rotations get a bit "jerky" - slow. **** My work station was built in 2013 and is five years old now. It was designed specifically for 3D CAD work: Windows 7 Professional 64 bit Intel i7-3930K 64 bit processor, 3.2 GHz, six cores, 12 processes (DesignCAD can use 11 processes, Windows typically uses 1) Intel DX79SR motherboard, 64 bit data bus Liquid cooler for the processor * Chassis has 10 fans to keep everything cool 32 Gbytes 64 bit DDR3-16 RAM - Corsair CMZ16GX3MAXx1600C ** Nvidia Quadro 2000 video card. 192 Graphic cores, 1 Gbyte DDR5 RAM. Drivers optimized for hardware support for OpenGL. *** Dual Display Port monitors, 27" and 24" Four hard drives: C 1TB, D 1 TB, E 1TB, F 3 TB. Drive D is used for the CAD model. **** I use a Kensington Expert Mouse trackball. * The Intel i5s and i7s have built in over temperature protection. This is provided by an extra CPU embedded in the chip that just monitors the hardware. If any core exceeds 65C the processor clock will be stopped until it cools below 65C, and then restarted. If this doesn't work the CPU voltage is reduced (this may cause instruction execution errors), and if that doesn't work the CPU shuts down. System performance plummets when the CPU clock is interrupted like this. If you want to run full speed under all circumstances (without processor clock interruption) you have to get rid of the heat in the processor. Air cooled heat sinks are only marginally effective. The liquid cooler has worked so far, and the machine runs without clock interruption even in the most demanding operations that run all six cores at 100% duty cycle for long periods. I can tell when the thing starts to heat up. The ten fans normally run with only a slight hum. But when I start working with large numbers of objects in DesignCAD, or rendering large files, or working with very large images in Photoshop, I notice the fans speeding up. If the room temperature is high (>85F) the fans can rev up so it sounds like the thing will lift off! Laptops do not have sufficient cooling, so i5s and i7s often operate with interrupted clocks. You may have a "3 GHz" CPU, but it may operate at a much lower effective clock speed if it is doing serious work. But at least it won't fry itself as some other processors were prone to do. The Intel motherboard came with a hardware monitor program. It allows monitoring temperatures, voltages, fan speeds, etc. My system normally runs with a CPU temperature of 32C with a room temperature around 25C (77F). I have never seen it go over about 42C. The i7 K series CPUs were designed for overclocking, and I have read of i7-3930s that are clocked at 4 GHz. I have not overclocked this CPU. It gets hot enough without overclocking! ** The fast RAM is the key. In 2013 this Corsair RAM was the only thing on the market that would actually run with a bus speed of 1.6 GHz at a bus voltage of 1.5 Volts (the i7 nominal bus voltage). Most other RAM makers claimed 1.6 GHz speed, but the RAM had to be run at greater than 1.5 Volts. Go over 1.6 volts and the CPU is fried, as has been discovered by many unfortunate people who did not do their homework. I enabled the XMP memory feature in the motherboard BIOS. This allows the CPU and RAM to work at the fastest possible speed (1.6 GHz) at 1.5 Volts. Without XMP enabled the RAM would have worked at 1.33 GHz. However, with other manufacturer's RAM and XMP enabled the memory bus voltage might be raised above 1.6 Volts and fry the CPU. You need to do your homework on this one! If in doubt, don't enable XMP! The RAM is quad interleaved on the Intel mother board to wring out the fastest performance. *** I work mostly in OpenGL in DesignCAD. The Nvidia drivers enable video card hardware execution of OpenGL operations. This is MUCH faster that total software execution of OpenGL code. Most other manufacturer's video drivers do not support hardware acceleration for OpenGL, it they support OpenGL at all. They are optimized for video games that don't use OpenGL. **** Hard drive seek operations are the slowest operations in the computer. Placing data files on a drive different from the drive where the operating system and program are located allows faster operation. Both Windows and DesignCAD use virtual memory paging to swap out code and data between RAM and the hard drives. With Windows and DesignCAD on Drive C and DesignCAD data files on Drive D, the code swapping seek operations on C can proceed while data seeks are in progress on Drive D. Phil

Hi, I'm a newbie to this forum, but not to CAD or CAD ship modeling. I have been using DesignCAD 3D MAX since 1988 in my work, and also for ship modeling. The program has three significant advantages over every other CAD program (about a dozen) that I have used. First, it is cheap - about $100 US. Second, it has the best and most versatile user interface of any program I have ever seen - five ways to execute commands, including a macro language, so you can work in the way most comfortable for you. And best of all, it has free technical support and a very active user forum (http://forum.designcadcommunity.com/). I have been working on a CAD model of the USS Oklahoma City CLG-5 in the summer 1971 configuration for fourteen years. It is just about complete. Most of the 14 years was spent researching blueprints, drawings, data sheets, etc., and collecting over 1000 photos of the ship. You can see images of this model here: https://www.okieboat.com/CAD model.html The entire model was done in DesignCAD, so if you want to see the capabilities of the program for ship modeling have a look at my web site. If you are new to CAD I cannot overstate the importance of an active users forum for the program. Many very experienced users from all over the planet follow the DesignCAD forum and are happy to answer questions from new users. DC tech support also watches the forum and answers questions. As others have said, learning CAD programs can be difficult for the first time user. You not only have to learn how to get the program to do what you want to do, but you also have to learn to think in a virtual reality. Forget any limits you have learned with 2D drawing on paper - the sheet is infinite. And if you work in 3D you have to learn to think in 3D - you are creating a new virtual world. It will take some time to become comfortable with this process. It is best to model in 1:1 scale - the actual dimensions of the ship. That way you can use the dimensions on blueprints directly without calculations and possible scaling errors. Later you can scale the drawing to any smaller scale. You can create 2D drawings from the 3D model, or you can produce 3D stereolith files (3D printing), or use your model to drive CNC machines (this is tricky). But the best thing about CAD modeling is that you can (and will) make mistakes, and later go back and fix them without wasting any materials. Phil