More stupid problems from a noob

hi all

So, I make boat covers and and am converting to digital, I have a flybridge cover I have been struggling with and the main part of the cover is ok as I had it cut in cheap plastic and test fitted it on the boat and thats ok. except for one part of the cover.
No matter what I try I can’t get a curved panel to flatten without 16% or more deviation,
I have tried to loft and Devloft and used squish, smash and Unroll and flatten, I have also tried cutting the panel vertically down the center making a less complicated shape for Rhino to work with, the product is fabric so I can add seams almost anywhere I want.
There has to be a way to get a reliable flat pattern from this shape but I’m lost now … ?

destroyer screen only.3dm (66.9 KB)

Hello - are both ends meant to meet? You have one edge pinching to a singularity and the other open.



Can you post the original curves you’re working from?


Hi Pascal
No the panel sews to 3 other panels I will post the whole thing . I need to learn how to flatten this type of surface. I can make that surface easily from plastic and it will lay totally flat , I did get a flat surface in Rhino but it was 100mm too narrow for some reason ?
fly 02 again working.3dm (525.0 KB)

Hmmm - is there any reason not to simplify your curves so they’re smoother and match up more cleanly?


hopefully as I get better yes I should be making the curves as clean as possible, it’s just that I haven’t learnt enough yet to spot a good curve from a bad one :slight_smile:
When making the panels in fabric and sewing them together I can smooth out minor errors with the sewing machine for now.

OK - the thing is, if your curves are clean and simple and within whatever tolerance you need from some ‘standard’ - possibly ugly but more or less accurate digitized curves(?), then the process of making the patterns should generally go much more smoothly. Is this particular piece symmetrical, in theory? I ask 'cause you’ve built each end differently but the neighboring surfaces seem about the same on each side.

For example, are these close enough?

fly 02 again working_PG.3dm (108.3 KB)


it will look more or less symmetrical in real life but boats are never that symmetrical, sometimes they can be 3 or 4 inches different from one side to the other. often as long as the cover looks right and fits from the side you are viewing thats ok because being a 40+ foot boat you will never see both sides at the same time. when I trained in the States 10 years ago the training guy told us the only straight thing on a boat is the centerline :slight_smile: and he was right.

some covers I make are a tight tolerance maybe 2 or 3 millimeters but this type can be +20 or 30 mm and still be ok

The digitizing tolerance of the boat should be +/- 0.5mm

i was hoping there is a relatively simple way to flatten this surface ?
otherwise I have no use for Rhino if it can’t help me make canvas cover, which seems odd because canvas guys do use it ?

Hello - well, as I say, if you make clean curves the whole process is likely to be a lot cleaner - you did not say if my example above is too far off to be useful…?


Sorry Pascal I did not notice you had attached an example, It does look very much cleaner than mine. I think maybe for this cover It would need to be a little closer to the measured lines , i only know this because I made a cheap plastic cover from the Rhino panels and when I fitted it to the boat it was quite tight. but I could play with that. I would be interested to know what tools you used to clean up the curves… when I tried to simplify a surface I ended up with bulges and rounded corners instead of sharp ones ?

Without cleaning up my effort I managed to get a flat surface with less distortion by converting the panel to a mesh From Nurbs object and then converted back to a surface with meshtoNURB which allowed me to Unrollsrf … is that acceptable ?

I thank you for your expert help, I am learning a lot and also learning how to find online guides for my problems… it’s only my lack of skill with Rhino that is holding me back from expanding my business.

Don’t give up yet @fragged8 on the Port side of the offending surface, if you make control points, at the point, coincident then the unrolled area is out by 0.22% instead of 16.7% - and that is without rebuilding the input.


I appreciate you guys spending your precious time helping out a noob but I don’t really understand what you mean by coincident points Brian ? i’m obviously missing huge amounts of knowledge to be able to do what I need. My problem is now that I can’t trust what I am designing to be worth wasting the canvas material at £20 a meter as I am one man band :frowning:
i seem to be getting somewhere with converting from NURBS to object and then meshtoNURB, then unrollsrf I don’t really understand what I am doing which bothers me and I don’t really know when to use this method and when not to, but the flattened surface looks very different to my failed attempt that didn’t fit on the boat so keeping my fingers crossed this one might.

I’ve been hacking this cover around for a few weeks now in Rhino when I could have patterned, made it and fitted it in 2 days using plastic and a sharpie …

This image should illustrate what I was trying to say… The area of the unrolled surface in your example is quite far out, this fix will give a closer result. I don’t think converting to mesh the way you describe is the way to go here as unroll should be accurate enough for fabric.

It’s probably worth unpicking the Fwd panel on the plastic trial and re- unroll the fixed panel.

Attached is a sample with the original curves rebuilt just a touch (3mm deviation from original)

Canopy Trial.3dm (141.5 KB)

goodness that was tricky, I found what you were referring to thank you. I think my zoom on my mouse wheel is fooked, I zoom in ok up to a point and then the zoom goes way over the top and the surface disappears. I did have to zoom way in to find the double points though, I would never have seen that without your expert help.
Because I know the rest of the cover fits I will package this up and send it off for cutting and see what happens, it will be much easier when I can afford a cutter of my own but thats at least another £20k to £30k I will post the results.

How did you spot the problem area in the first place ? I wouldn’t have been able to spot where to look to find the problem

I have a long way to go yet but your and Pascal’s help is worth it’s weight in gold, thank you.

1 Like

‘_Zoom _Target’ or ‘_Zoom _Window’ will allow you to zoom in closer when the mouse wheel reaches it’s limit.

This kind of situation often results from problems in the curves used to create the surface, for example, too many control points, or control points which could be faired better with their neighbours, or (this case) the curves may not be joined (coincident) at their end points.

The construction curves in your example can be cleaned up considerably (see image) before making surfaces. Rebuild the curves into a degree 3, in this case and reduce the number of control points, depending on the amount of deviation you can live with.

Then further manipulate the points with ‘_MoveUVN’ until they flow nicely - you can also delete them individually. Then another rebuild may help. This will go a long way to achieving clean curves and resulting clean surfaces.

A good few cover jobs will be needed to pay for a cutter, yet the convenience and control it would give you, as well as the cost saving on cutter services would probably be worth a cost exercise if you have ongoing work…

Move UVN is another new tool to me thx, I see the sense in cleaning up the digitised curves before making the surfaces and am starting to get better at those jobs, I’m just worried that by cleaning up the curves I am actually adjusting the geometry slightly which then might adversely affect the cover fit. I have had a few issues with the proliner itself or more accurately my operation of the measuring device, because it is so accurate I have to be more aware of the way I measure the boats to reduce some of the deviation at source.

The one day course I did using Rhino was far too simplistic and hasn’t given me enough knowledge so I will look into more specific training if I can find some in my area but that’ll have to wait until the winter.

I’m saving all the great tips, thank you for your help I really appreciate it…

Cleaning up the input curves will more likely improve the cover fit, rather than starting off with the Proliner’s straight segments between measure points.

Ideally you want to end up with as few control points as possible on the new curve, within an acceptable deviation from the original. ‘_CrvDeviation’ will show you the difference between the old and new.

Here are a couple of adjustments to you working model that would improve the resulting surfaces.

Another way of starting with good curves is to trace over the digitised input with ‘_Curve’, selecting carefully considered pick points, then visually inspect the result compared to the input.



Just keep at it… and tune in here for tips

1 Like

Your help is superb thank you, honesty I was told by a canvas guy who does this kind of work it would take 6 months or more to get up to speed to be able to make covers this way. I’m around a month into it now and with your and Pascal’s help I’m well on my way, if I can remember all this golden information :slight_smile:
Just about every penny I owned went on the proliner and mostly train at night sometimes till 3am while keeping my business going 9 to 5, 7 days a week so there really isn’t an option, I have to make it pay hehe

This is my next nightmare once the flybridge is done, but i may need to re-measure the boat because of errors hehe
tonneau05 new orient box05.dxf (189.9 KB)

What do you do with Rhino ? it has to be your main job ?

Have you checked out

1 Like

thank you stratosfear, yes I have seen it and been to the website. I figured that I should get better with Rhino before I start adding in other add ons I don’t understand :slight_smile: Once I am confident with making boat covers this way D.Loft could be an option. Exactflat would be the ideal for me but the cost is horrific.