The Ship Model Forum

For what it is worth, I edited the CAD model of the AN/SPG-49 radar antenna and removed all of the nuts, bolts, screws and washers.

Original file size = 46.4 Mbytes.

Strip down file = 33.3 Mbytes.

28 percent of the file was fasteners!

Replacing the original directors with the stripdown version will reduce the final file size by 26 Mbytes. Maybe after I strip the fasteners out of the entire whole ship model it will be small enough that I can work with it in reasonable amounts of time.

Someday, with 100 GHz processors and unlimited RAM maybe I can work with the detailed files again. Guess what I want for Christmas?

Phil

Don't worry Phil, The Nice People at Intel and AMD are bringing 6 and eventually 10 core processors to the desktop.

Until that time, you could consider a Power Mac.

Before someone... me, just think...

They runs Windows flawlessly, have up to 8 real and 12 virtual cores and can stash up to 64 Gb of memory.

The only downers are the price (ouch!) and that it comes with an ATI video card.

Owen

Owen,

I am very skeptical about the value of multicore processors in PCs. My company has written software for specially designed 16 processor machines, and the program has to be crafted carefully to make simultaneous use of a specific number of processors (it won't run on single processors). It is an extraordinarily complex task, and the average programmer doesn't have a clue about how to write software to run simultaneously on multiple processors.

I know of no off-the-shelf applications currently written for multiple processors. Such programs wouldn't run on machines with fewer than the number of cores the program was designed for, and there are a lot of single core machines out there. What company would go to the expense of writing multi-core code that they couldn't sell?

The fundamental limitation is the operating system - it has to be designed to supervise the use of multiple cores. In theory, operating systems could parse out individual single core programs to single cores, so that multiple programs could operate simultaneously. I have observed multiple core use by Windows XP and Vista and only one core is being used. The rest are idle most, if not all of the time.

For example, my CAD programs, Microsoft Office programs, Adobe software and everything else I have looked at all run on the same single processor in Windows XP and Windows Vista. It even appears that Windows also uses the same core that the applications are running in. The other cores are idle. I have never seen two cores operating simultaneously as would be the case with true multicore software. I haven't examined Windows 7 or the latest Apple OS, so it is possible that they have some multiple processor hooks, but I would be surprised. I would be more surprised to see application software that tried to use multiple cores when the operating system didn't support it.

Don't get me wrong - no one will write multicore software until multicore processors are common (unless you want to shell out a few million dollars for your special application). So the proliferation of multicore machines is a good start. However, as long as companies can get by selling their old single core software the suits in charge won't shell out the really big bucks necessary to rewrite code for multiple processors - I am talking about total rewrite, starting at the fundamental libraries, and then extensive debugging. Years of development expenses. The programs that are rewritten are going to be very expensive - not hobby use software.

Right now I suspect all the multicore hoopla is just advertising lies put out by sleezeball marketing types to separate the suckers from their money.

Phil

I know that some soft is 'optimized' for mullticore processors. There are benchmarks for processors and if they add more cores, the performance climbs. It is a minefield though, with some soft (yes, Adobe products ) seemingly impervious to adding more processing oomph. They only respond to a higer clock speed rather than more cores to operate with. I have no idea how much 'optimization' takes place though to get it to do that. I know that large parts of Blender are not multi-core optimized, though for modelling and rendering sections, you can actually tell it how many cores to use if you want to do some surfing while waiting for a render to percolate.

I also have no idea what percentage of the sharing is done by the OS parceling up tasks rather than the application knowing what to do.

The other problem is the new Intel chips, which, from what I can deterimine, use Hyperthreading on top of multiple actual cores to give the effect of a 6 die chip on 4 die hardware.

Owen

Owen,

You are correct! I talked with our head programmer today and he said the current software tools almost force programs to be broken up into multiple threads. It helps with efficient programming for Internet applications and some operating systems can manage multiple threads. Windows 7 comes in seven versions - the cheapest don't support multiple threads, and the more you pay the more threads each version can handle. Apparently Vista can handle two threads.

Shows how much I have been paying attention!

A "thread" is a section of code that can operate independently from the rest of the program. If the operating system can manage multiple threads it can parse out each thread to separate cores, allowing true multi-core operation. Each core can also run multiple threads.

My friend has observed four cores in use, with four threads each, on a Linux system he is programming.

So true multi-core operation can be had for a price. The times they are a changin' faster than I thought.

Phil

Yeah, things are changing at a ridiculous rate. I only keep up as I used to work in tech support. I was surprised recently at how much soft ISN'T aware of multiple cores, or are only aware of two cores. In ye olde days people used to use games to test a computer, but not so much now.

Owen

Hey,

Things are slowing down a bit on the virtual modelling forum so I thought I would stir up some action.

Building the hull can be challenging, especially if you include plating details. The basic superstructure usually isn't too bad - just a bunch of boxes, although they can be some pretty strange shaped boxes. Then adding life rails, doors, etc. adds a lot of repetitive detail that is necessary but not too challenging.

It is the enormous number of little fiddly bits that really drags out the model construction, and the soft, marshmallow and wet spaghetti like things are the worst! I hate making knots in ropes!

The attached picture shows another necessary detail that requires a huge amount of work for one small detail - fire hoses. The Oklahoma City had 1 1/2" and 2 1/2" fire hoses arranged on racks all throughout the ship. They were mostly 2 1/2" topside with the smaller diameter hoses in the interior (ever try to wrap a charged 2 1/2" hose through doors and down ladders?). Some of the racks were single layer short hoses, but the really long hoses were laid up two layers deep. And they were usually close to fire mains with strainers. Spud tip (low velocity water fog) extensions - the long pipes with curved ends and the brass spray heads (spud tips) at the ends - were often mounted on bulkheads hearby.

Drawing these folded hoses is a test of my patience! It takes me two to three days to finish one of these things (the first time). Of course, there really are only two types here - large and small. I mirror image and flip each basic type to get some apparent variety. Besides, they are scattered fairly far apart topside so no two are visible at the same time for a comparison that would reveal that each instance isn't unique, as it would be in the real world.

The kinked curve - just a bit pointed - at the ends of the loops of the 2 1/2" hoses was a bit tricky. I made smooth curves in my first attempt, but when I compared them to photos they just didn't look real. The hoses are actually thick rubber hoses on the inside with a canvas wrapping. The rubber kinks when it is folded and produced the somewhat pointed look. The smaller 1 1/2 inch hoses bend more easily.

These are just a few of the bazillions of details on this ship that had conversions on top of conversions and numerous additions and changes over its 35 year career.

Phil

Phil how is your actual Okie city model progressing or are you concentrating on the 3d version??

Cheers Bruce

PS can all your CAD work be used for Laser cutting superstructure pieces?? I wouldn't mind doing a USS Canberra one day for my RC Cruiser squadron....

Bruce,

Work on the physical 1:96 model stopped a couple of years ago. I kept running into contradicting blueprints and other puzzles so I decided to stop until the entire CAD drawing is finished. That way if it is wrong at least it will be consistently wrong!

I tried to work on the smoke pipes but had problems with very fine detail photoetch for the catwalks - and that turned out to be a good thing. Recently I discovered some black and white negatives I made while on the ship 40+ years ago and had forgotten about completely. After I digitized them I found a very clear photograph that showed that the catwalk on the OK City wasn't built like the blueprints. So, if I had succeeded with the photo etch I would have to reconstruct the after funnel.

Now I want to go back and examine photos of the original Cleveland class to see if other ships had this altered catwalk - my guess is that the five ships built by Cramp Shipyards probably all had modified catwalks, since this type of change seems to have been shipyard specific. But it could have been a change for the square bridge version. To be continued ...

In addition to my long-lost photos I have received a large number of pictures from fellows who have contacted me through my web site. Many of these have cleared up little mysteries that I have run into over the years. Right now I have photos of almost every square inch of the ship topside, so I am guessing about only a very few small details.

I am converting my 3D models into 2D drawings that show the dimensions and details of the parts. The superstructure drawings could be used to drive a laser cutter. The same is true for the decks, platforms, etc. and any flat pieces. It would be pretty cool to lay out all the parts and have a laser cut them out of a sheet of 0.010" brass! It would make an awesome kit!

The drawings can be scaled to any scale by changing a single controlling dimension.

Phil

It has been a while since I posted on this project. Business, life and other complications really cut into my hobby time since last November, but I have been slowly plugging away when I get the time.

I am working on the midships deck house, and it is far and away the most complex part of the project. However, I discovered a bunch of old B&W negatives I had completely forgotten about that I made while on the ship and they are a treasure trove. I actually took photos with the intent to model the ship! In addition, one of the fellows I served with sent dozens of his photos. When I get time I will post them all on my web page:

http://www.okieboat.com

Here is a picture of part of the deck house - bazillions of tiny details. I still need to rig the kingpost booms for handling the boats. I also need to create a 28 foot personnel boat (three on the ship) and the very complex boat winches. I have a lot of wiring and light switches, etc., and the cables going up the midships tower that sits atop the small house in the center of the picture. Maybe I can finish this part by new years. Then all that remains is the forward deck house, and it is trivial compared to this one!

Phil

Looks great Phil, awesome detail! I give you guys that do these mega detailed models a lot of credit, I just don't have a long enough attention span to take on a project like that.

Dean

Your thread on the winch gave some idea of the amazing depth of zoomable detail here. A thing I also noticed is the sheer artistry of a detail like the wiring in the foreground. It would be easy, because CAD mechanizes the drawing process -- to have this wiring turn out looking too perfect, like a schematic -- but it looks hand wired. Expertly hand wired, but clearly done by hand nevertheless. I don't know how you do it, Phil.

As they say in England: "Gob-smacked".

The wiring is a real pain in the posterior, in several ways. I create it by drawing a curved line along the desired path and then extruding a polygon along the curve.

The nasty part is where you have a bunch of cables in a run and you want to break out one of them. The first atttempt invariably has one cable running through others. It takes some fiddling to get the extrusion curve just right.

The really big problem is the file size increase from these cables. They make huge files that take forever to render. I "cheat" by using an octagonal cross section. A 36 sided polygon would render much smoother - in fact, the octagon will not render as a smooth cylinder. But the octogon produces only about 1/5 as many facets in the extrusion, and a much smaller file size. I also minimize the intermediate steps along the length of the cable.

The result would be a really crummy looking "pipe" but actually makes a better looking cable, since cables don't have smooth surfaces. Just don't look too closely!

It would be a lot easier to leave out the cables, but they add an important bit of realism to the picture. And when I get back to the 1:96 scale model I will want to add the wiring, and having already put it on the CAD model I won't have to scratch my head as much figuring out where to run it.

Phil

Is there any plan to print this with a high res 3D printer?
Dave

Dave,

I have had very poor results with 3D printing over the years (grainy material, low resolution, structurally weak), but it is improving. However, at 1:96 some parts would very large and phenominally expensive, if I could find a printer that could create parts that big.

3D printing could be used for individual small parts like winches and boats. It would be even better for the radar towers - there are places where more than a dozen pieces come together and I don't know how I will make those assemblies.

Phil

I have made a bit of progress. The ship had two kingposts with heavy lift booms, used mostly to move the ships boats. One of the details I have dreaded is adding all of the rigging. If you run ropes through blocks in the real world, when you pull them tight everything assumes the "right" positions.

In the CAD world you have to position everything one piece at a time. Getting the blocks at the correct angles is tedious, basicllly a graphic solution to vector geometry. And when you reposition one block it changes the angles for all the rest.

The diamond blocks in the picture were pretty easy - the end positions were fixed (no swivels) and the heavy tension pulled everything straight.

The circular blocks were a different story. They were mounted on swivels with three degrees of freedom, and that meant they could point any direction. Anyway, at least the topping lift is done. The boat lift snakes through several of the round blocks (not shown) with three degrees of freedom, and I haven't finished the boat winch yet.

Phil

I see it has been a year since I posted results of my build! Time flies when you're having fun! It has been a busy year for things other than 3D CAD modeling.

I have been adding details to the midships superstructure, far and away the most complex part of the ship. I now have all the doors, air ports, railings, ladders, lights, wiring, switches, speakers and other details incorporated into the file. I have added most of the rigging for the kingposts and boat booms, although I still have to decide how I will rig the boom vangs. No two photos show the same rigging, and it changed a lot over the 19 years the CLG/CG version of the ship was in service.

The pictures show the almost completed deckhouses (less the midships tower). I still have to create the boat winches (port and starboard) and the small personnel boat on the port side.

Phil

Very nicely done Phil!

I always look forward to seeing your work. Very meticulous.

As always, totally amazing. I particularly admire the ship's boats and the fine detail of their stowing and launching equipment. Michael