lofting basics help

Good morning,
Thank you for the inputs; I am seeking a simple machine that is easy to set up to take the section lines from the half model in my effort to fair the opposing side.
Upon taking lines I will import the data into my CAD program for a bit of refinement.

I know of several schemes for taking off lines on full size craft and half models. The schemes for full size crafts are of little to no value as the scale lends itself to large margins of error.
Attached are a few examples of a half model duplicator; I like the simplistic design of the one on the right and I may fabricate one but it seems to lack the ability of taking the off the lines below the chine when the traced section has a bit of concave shape that bends back behind the chine line.

It is refreshing to see that some of you fellows understand that a measure of error at half model scale is amplified greatly when you scale the section data to a greater scale. I made a noble effort of conveying that thought some time back but I did a poor job authoring it

The mechanical tracing gadget is a viable solution but it requires way to much to set up and I am seeking simplicity, it also would require some additional hardware to make it work.



For what it’s worth…
In my humble opinion, I believe it to be quicker to fabricate the half model than it is to sit at the computer and CAD the subject.

A few of the half models that I need to fair

http://www.rcgroups.com/forums/showthread.php?t=1383457#post17444575

http://www.rcgroups.com/forums/showthread.php?t=1392032


I welcome any inputs

Thanks again,
Dave

 

In the distant past I used a photographic method for scaling up. It involved taking a photo using transparency film and enlarging the image using a projector. All that way back I had to trace it out with a steady hand - I had one back then but that was about 20,000 beers ago ...

The stations would have been lined out with tape back then, if I didn't want to slice up the half model. And if the scale was a ratio rather than a known final size a small rule would have been included in the scene, together with an accurate circle (coin?) for checking the aspect ratio.

These days I would use a laser level to create the station lines and a digital camera no need for slicing. If the final image size exceeded the size of my printer there is a local print outfit who will print out the file, or they can enlarge a printout on a roll-fed xerox that will output 24" wide - any length. That's a method I plan to try for plank developments on a design I am working on. My printer has a 0.5% aspect ratio error which is not adjustable so I will have to compensate for that in software, but it pays to check ...

 

Technically FreeShip and DelftShip use a subdivision algorithm to create curves and surfaces, not NURBS. But the results behave similarly to NURBS and the subdivision curves and surfaces approach NURBS curves and surfaces as the mesh is refined.

This may be a topic for a different thread.

 

I'm not sure I've encountered the idea of lofting for model aircraft before!

I think that we need to remember that a good part of the reason for drawing lines at full-scale was to derive frame shapes. In the modern era, you should have sufficient command of your CAD tools to be able to produce a smooth and fair shape.

Most CAD tools now have render capability (which is very effective for seeing discontinuities), and curvature analysis. 20 years ago, computer graphics weren't good enough to ensure fairness on a hull. With good CAD work, I don't see the need now for doing extra lofting work by hand. You should be able to use the CAD to produce a sufficiently accurate frame at any location along the hull (or fuselage). This being the case, then when the skin is wrapped over it, the skin should be fair.

This is somewhat dependant on having a good understanding of your CAD tools, of course,

Tim B.

 

Good morning designers,
Thank you for all the wonderful responses, it would seem the some of the comments are a bit off the later topic but somewhat valid points are being presented.
It should be no secret that a Master of his craft can outperform/kill two birds with one stone; loft & fabricate in one step. All that is required then is to simply take off the lines and tidy them up a bit.

I have designed a simple machine to perform the task, and yes it was done on the computer.

Cheers,
Dave

 

Hi , I have a lofting question as well so I bumped the thread nistead of creating a new one.

My boat will be in the 35 to 47 foot range so I would require a say a 30 x 70 foot lofting floor. Well , There isn,t a shop that big around here and ,well it constantly rains just about every day! I was planning on a temporary shelter for constuction. Is building a huge shop to protect my drawing nessisary?

I am a registered user and owner of autocad 2002 so I already have a great cad program. Given offset tables and autocad has infinite space, I should be able to computer loft and fair to 2 or 3 decimal places by keyboard by inputing my X,Y coordinates. Should I not? in layout if you set your viewport and final scale to 1:1 you would have full size lofting. Could you not derive full size patterns from finding a plotting service? If so do you know of a plotting service in the Pacific Northwest? Preferably Canada? Hey any help you can throw my way would surely help!
Thanks

 

yes - you are absolutely correct - you can loft 2d shapes on the pc from offsets without a problem. I guess the question that springs to my mind is why ?

If you are not using a proven design, which would provide all the offsets and measurements you need to transfer straight onto panels of construction material, then creating a 50ft boat, when you are just in the 'what can I do in Autocad 2' stage may be a bit over your experience level perhaps. I think Autocad 2 was not even 3d - but I am not totally sure.

You might be better off spending $150, getting Delftship, or some other specialist boat design package, so you could calculate displacement levels, assign weight tables, calculate righting moments etc etc - which are the very essential starting point of big hull design.

a 50' boat is not something many people can design 'by eye' on a 2 dimensional cad program.

or, you might be better off building a 1/5 scale model, and floating that, than starting into 2d cad for the first step.

Perhaps a few more details of you project and the stage you are at could help people suggest the most effective approach.

 

New to Lofting

The task of lofting is one of my favorite challenges and in 26 years of doing so the task is always revealing a new twist on what you might expect as sections being totally wrong to get a desired form. The beauty of CAD is that you have an unlimited scale, weightless splines and no cost of large sheltered loft space. If you want to learn the art for design, and it is an art since the swoop and curve of designing your own ravishing form is subjective. Keep in mind the resolution and acceptable minimum fairness for builder sections is less so. The fidelity of CAD is great and it cannot be forced to spit out a section that will not result in the developed shape. That is what the origins of the loft was used for, as a "refined, full scale geometry provider" between the designer scaled-lines drawing and the ship builder. I have yet to find a perfect scaled-lines set of offsets that survived without revision to the real build.

Especially with older projects like the plane you note. many times the drafting used "near sections" to show the construction of say a bulkhead, and more typically a walking flat too wide not allowing the clipping of the plane only revealed in the loft.

I think ACAD "inventor" has "bezier curve" spline capability, you need this for weighted stiffness of those defining lines you will develop first when inputting the shape. I suggest the following from years of practice:

Overshoot the major lines like chines, sheers and keels, or use a few more sections past the ends of the form. its easer to cut off "free" material in CAD to the actual lengths.

You can get a cleaner "steps and transoms" cut from a projected surface, its much faster since your not trying to guess the unknown, eventual "lofted edge" of surfaces at intersections. This factor is why its harder to input a existing shape into 3/D as opposed to a new one. The "stiffness" of the battens used for any existing shape is never known nor consistent plus a number of other similar reasons.

Its been my experience to use as few defining sections as possible for the desired shape.

The most challenging hulls I have ever fared are Buddy Davis and Rybo SF hulls since the concave bow flair turns into convex sides going aft! AND the sheer is usually split and butts into the cockpit coaming. This can only be done by projecting fair surfaces and trimming them at their computed intersections.

 

CAD lofting must!

I forgot to add, it is most important that you "preview" any 3/D CAD model with the proper perspective turned on when viewing!

It's not funny to watch and it happens every day that a designer will model a CAD form while only checking the shape viewing it in isometric view. That is only sufficient for a table top size shape. Your full size boat will always be viewed from a human being (the client?) who is somewhat smaller than the actual form. These boats are photographed usually at a full scale friendly lens length of 55mm. The images you see in magazines are usually shot with a "sexier" 30mm lens! ACAD and Rhino has a great utility to adjust these "depth of field" settings and gained me a substantial edge over other designers since I knew to preview and show the evolving shape in its proper perspective. Keep that in mind as you develop large shapes on a 24" screen.

That remains the major advantage of CAD lofting over old loft floors. Originally, no one knew what the actual full size ship, plane or boat would look like in real life until it was mostly framed. It's a very big deal if you do lofting for product design of large vehicles. That appearance distortion is also an unfortunate outcome with scaled physical models, ever notice how you can tell if your seeing a scale model train setting over a real landscape?

 

Thanks guys Hopefully I will recieve good plans that come with full sized patterns. In the event they don't I just want the ability to input the given offset tables into autocad 2002 and loft /print the patterns if nessisary. I am not designing from scratch. I noticed only a few designers offer full sized patterns so lacking a facility and a ton of plywood I thought this might better suit my needs.

So I think if I set the precision to xx.xxx I should be fine for fairness. I am good with a torch but I doubt I would get within a thou ! Atually I am a registered owner/user of the invetor 5.3 as well so I do most of the 3d work I need in there. It feels more natural to me and less clunky than doing 3d in autocad. I just don't like the layout in paperspace as much in 5.3 though.
With the accuracy aspect, I can computer loft/print simple offsets very acurate but doing a 30 x 70 grid would require a transit and plunging/doubling your ninties to get a really acurate grid. I

Thanbks again guys , very much appreciated. This will open a few more designs for me without the intimidation or lack of shop space

 

Two and a half different approaches to lofting/fairing from a table of offsets and lines drawing on a computer:

Virtual Lines Drawing 2D: Direct electronic equivalent of the loft floor. Work in three 2D views. Create the relevant curves in each view (sections, profile, half-breadths, waterlines, etc) from the offsets and/or a scan of the lines drawings. Modify the curves in each view to reconcile the intersections of the curves (sections, profile, half-breadths, waterlines, etc) the same as you would on a loft floor. Check smoothness of the curves in each view using available curvature plots, etc. Only information which is available are the shapes of the curves and their intersections. How closely the curves intersect is dependent on the operator.

Virtual Lines Drawing 3D: Work with 3D curves in 3D electronic space. Create the relevant curves in 3D from the offsets and/or a scan of the lines drawings. Modify the curves as needed so that they intersect. Easier and simplier to see if the curves intersect than when working in 2D. Check smoothness of the curves using available curvature plots, etc. Viewpoint can be moved around to check for any reverse curvatures, etc. Only information which is available are the shapes of the curves and their intersections. How closely the curves intersect is dependent on the operator.

Virtual Half Model 3D: Create a 3D surface model of the hull shape analagous to a half model based on offsets. This can take some practice. I start with creating set of lines the same as I would to start in the Traditional Lines Drawing 3D method. 3D curves modified to eliminate any significant gaps and correct errors. Surface is created using the curves. Surface is then modified for fairness. Again, practice helps. Fairness is checked with a variety of tools including curvature color plots, "zebra" plots, and reflected views of "fluorescent tubes", etc. Once the 3D surface is fair, offset surfaces to represent the inside of the hull and developed surfaces (if appropriate) etc can be generated. Then the needed numerical data is extracted, including data for CNC cutting of panels, molds, etc.

 

Fairness is different than accuracy. Fairness is whether the curves and surfaces are sufficiently smooth. For a 3D surface the cross section curves need to be individually fair, but that isn't sufficient. They cross sections needed to be fair relative to each other so the resulting surface doesn't have bumps and hollows in the transistion from one curve to another.

 

Fair?

I agree about accuracy not being related to fairness. General shape comes first, fairness next (design spiral) and accuracy onto the full size frame is the way.

 

Plotting services should available in any city and most towns. You should check the kinds of file they can use, also size limitations since for a large boat you may need to break down some plots into sections. You may need the drivers for their machines but it is unlikely these days, they probably use a standard file type. I can't give you specifics as it is a very long time since I needed to do this, and I am a couple of decades behind the times.

 

Thanks again, I really do think when it comes time for me to choose the final design/designer I expect that they knew what they were doing and faired the hull already. I expect to pay $ 1000.00 to $ 4500.00 for stock plans so one would hope that the given offset tables are acurate. I hope I can find a designer who would have full size patterns,if they are not included then thats all I need the lofting for. I just need the frames and plating if they didn't full size them.

I mentioned presision settings for acuracy in the grid and offset coordinates. If the coordinates are not precise, then splining no mater how much care will not line up with the fixed reference offset points. Maybe I am thinking backwards, but why would you have offsets if not for control points?

I will not be G -coding (NC) If I did, I would probably export to mastercam 9 for my pathing. It is alot easier to proof your pathing in a simulation mode. I have no access to a CNC table , however..... I do have access to an optical oxy/acetylene table and therefore paper would suit me better. I use the table at least once a week and it has two torches so making twins for port/starboard off of one pattern path is possible.

Whew so much to think of prior to actually getting to the build.