This could verywell be the case. But be careful with flattening curved surfaces.
SketchUp is accurate enough to be used for straight objects like most architectural parts of a building are. But like here with a hull that needs to fit over an also applied frame of fixed cross sections you’ll need to think twice. You haven’t convinced me yet. But that’s my problem.
I hope you’re right in having handed correct drawing plans.
Again, is this hull the first built?
Andrew, do you build the hulls you draw? You should at least build a scale model from the plans you create.
Did you check the plans to make sure they are clear and that the dimensions in print make sense? I think you are probably right that it is a setting out error on his part but in all fairness to him, I think you should spend some time going over the package you sent to him to make sure the numbers are correct and that you’ve clearly communicated the shapes of the panels.
Have you considered giving the customer that patterns for the frames and bulkheads and instructions for setting up the strong back? Let them take the curves and the lengths from the that. It would be a more traditional approach to building the hull but at least they’ll be cutting the panels to what they’ve already set up.
Hi Wo3dan
I’ve sold several sets of boat plans, this is the first report with accuracy issues.
A few years ago someone took the CAD files for another design, CNC cut the panels and built a model about 1m long - No error report.
I received photo’s of a completed build of yet another design a few months ago - the builder was happy with the plans, the boat and it’s performance - no error report.
This guy was so adamant that he does not make setting out mistakes, and is of the opinion that CAD drawings are inherently inaccurate, that I felt I just had to post this topic.
Right now, I’m almost certain that my drawings are correct. At some stage I’ll make the hull myself.
I realise that SketchUp divides curves into segments, but the test I did in my previous post has demonstrated to me that it won’t make any difference if each segment is 5mm or 30mm long.
Hi DaveR
As I said in my original post, when I received the error report, I checked everything - the Layout dimensions, the dimensions within SketchUP to make sure they correlated, the edge lengths on the curved hull panels and the flattened hull panels, I redrew to check for drawing /flattening error. I’m convinced that my drawings are right.
This guy is of the opinion that CAD drawings are inherently inaccurate … I personally consider that assumption to be wrong… But I decided to post this topic in order to find out if anyone else had had inaccuracy issues with SketchUP …so far no-one has.
I still fail to see how two flattened panels with the same edge length on both the 3D model and on the flattened panels can throw up a 1-1/2" discrepancy when stitched together.
Yes, I built a scale model out of card and everything fitted as it should. His response to this was that I should build full size!
Any good CAD program could be considered accurate. So is SketchuP. Cad drawings are what people make of them… either accurate or not.
A small scale model may reveal errors. It’s a first step. But if not, if seemingly no errors are found, it does not mean that at full scale there can’t be a mistake of 1 -1/2" over a long stretch like a panel of 4.5 meters. You’re talking about less than 1%. Hardly noticed in a model 1/100 the actual size.
Anyway if others have already made hullls to their satisfaction using your plans, you are probably right.
I would guess that if the error here is not entirely on the part of the builder, then perhaps on the process of “flattening” the geometry. As a seasoned veteran of Autocad myself, I’ll compare the “flatshot” command, which effectively takes an orthographic picture of your 3D geometry and creates a 2D copy… not OF the 3D geometry, but WHAT IT LOOKS LIKE from a 2D perspective. Perhaps you’ve already considered this, but sometimes it is the easy things that we trip over. For example, if you cut something simple… say the cardboard tube in the center of a roll of toilet paper… in half from end to end so you have two, 180-degree curved lengths of tube. If you look at the side view, you effectively see a rectangle. While it is true that the shape, when flattened still makes a rectangle… it actually makes a larger rectangle that what you see orthagonally when viewed from the side due to the ‘unrolling’ of the curve when flattened. If the curved portions of your hull are subtle and not 180-degrees like the example here, the sum of the differences might add up to the small amounts the builder is experiencing.
I’m curious about the flattening process your 3rd party uses, and just how curved your pieces are. My perception of a boat hull is that the curve is not just a linear curve (like my cardboard roll example)? But rather a more complicated curve that bends in more than one direction? Obviously people have been building boat hulls for centuries without CAD software, but translating a 3D curved shape to something that is flat so it can be traced and cut and then reassembled into something with a complex curve, seems problematic. Again, just curious about the process.
The problem of getting flat material to lie fairly on the compound curves of a boat hull is as old as boat building and is the source of endless issues that experienced builders learn to deal with. The wider the planks, the greater the problem because a flat board will only bend into a “developable surface”, that is, one with zero Gaussian curvature (too much math for most, but basically it means the surface is locally flat in at least one direction). On carvel-planked boats they sand out the flats. On lap-strake boats they use narrow planks to localize the issue, but the planks have to be carefully spiled to avoid gaps. And that’s why many plywood boats with wide planks have such a slab-sided appearance. A large developable surface looks flat-ish to the eye.
Laying out errors can happen to anyone. I took a course at Wooden Boat School taught by the designer of the hull we were building. Several instances had been built before. But partway through we noticed that the planks weren’t landing quite right, and he had us use long battens to find new places for the remaining ones to land. All through the exercise he was mumbling under his breath “what idiot layed out these lines, anyway?” (referring to his own plans). We also found that one plank needed to be about a foot longer than possible from the plywood we were using. He grumbled something about metric vs imperial plywood, but after looking up the standard sizes I could never believe that was the explanation.
So my conclusion is that the person complaining about CAD simply doesn’t have much practical experience building boats! Errors in the plans and the laying out are facts of life.
1 Like
That’s why experienced boat builders loft a boat before they build it and why I suggested to the OP that he supply plans for building the frames for the strong back and let the builder take the plank shapes from the built frames.
I agree. The problem is that too many people want to build a boat, but don’t want to learn the art of taking the plank shapes off the frames. So, pre-cut kits and full-size patterns are the trend.
That’s true. I’d build a boat from a pre-cut kit from a known supplier but for a boat I’d build from scratch, I wouldn’t think if cutting the planks without taking the shapes off the building frame.
Even then getting “true” lengths can be issue unless is very careful:
- The bulk heads are just basically a sample data system and many times there are not enough to establish the true curve;
2 True lengths can only be obtained if one has a normal view. The OP pics seem to indicate some are not; - Many shapes are based on true spline curves which Su does not have although OP seems to indicate all fat panels;
- Use of fractions are ok to give info to trades but unless you have snap to length off or input values in the vcb the true lengths can be off.
5 IMHO any one presenting plans / drawing of hull design should have info on buoyancy and turning capability.
I think the point Dave and I are making is that the client of the OP is being OCD in a way that suggests lack of experience building boats. The reality is that as you build, errors creep in and you have to adjust the rest to accommodate them - even if the parts were laser cut from a CAD model. You learn how to cope with it rather than expecting perfection from the plans.
3 Likes
This is the key issue here and SketchUp can deal with such designs.
But to get the shapes of the panels from accurate plans requires quite some experience.
With less or no experience one needs to build the assembly of frames and work from them. … knowing that the panels are straight in one direction.
Not to say that you can get quite complicated hull designes with chine shaped framing and plywood or sheets of steel or aluminum.
My thanks to everyone for responding to my post.
My original post post advised that I’d checked my drawings and model and I couldn’t find an error. As a consequence of this and of the builder’s insistence that his setting out was right and that ALL of my drawings contained errors, I wondered if there may be accuracy issues with SketchUp. I am now satisfied that, for the purposes that I use it, SketchUp is very accurate.
I say this because, through this post, it’s been pointed out that SketchUp’s way of breaking curves into segments may introduce errors and that the length of the inner edge of the hull panel may be shorter than that of the outer edge, due to its thickness. I checked these by increasing segments and by adding thickness to the hull and checking inner and outer lengths… Not an issue.
I’ve made models of this and other designs and I see no errors. I’m also happy that the developed hull panel drawings are correct … This is not a difficult geometric exercise, the hull curves are created by a series of triangles (see image on earlier post above), all one has to do is twist every triangle to be on the same plane as the adjacent one. It can be done manually, I use plugins to create the surface triangles and 3rd party software to flatten them. I check by comparing edge lengths and lengths of frame edges and the corresponding joining lines on the hull panels.
So thank you all for your feedback. I’m now happy to close this topic.
Best wishes, Andrew
I design sets for a local drama group. We have standard sizes of canvas flats which I have set up in my componants file and just move them out onto my stage plan where nesessary But I have found on occasion that when I come to label up the sizes of the flats for the construction team that some of the 1000mm flats for intance have becom 998 mm although I always use the same component template
I have also hit trouble in designing an asametric joint. When tansfering a copy of one part of the joint from one drawing to another to find it dosen’t’ fit bacause I had not noticed that the half I copied had flipped in transfer
For flattening, have you tried the “unwrap and flatten faces” plugin by alexschreyer ? I have installed this, and it works fairly well, but I haven’t employed it for final product yet. This may, or may not, assist you in accuracy, as it doesn’t require export of your SU drawing to another software program to flatten.
Additionally, a question, I cannot find the “joint push pull” plugin that you mentioned. Can you suggest a search term, or is it part of a suite of plugins?
Joint Push/Pull is available from Sketchucation’s Plugin Store. Make sure you also install LibFredo6 which is available from the same place.
Thanks. Please see my 26 Feb email. I’m very happy with the accuracy of Sketchup. The cause of my concern arose from accusations by a third party who didn’t trust the accuracy of computer generated drawings. Needless to say… I am no longer working with him.