Here's my problem. I'm trying to make a shell of a very complex part for a 3D print. Because my part has a lot of fillet and chamfers, solidworks can't make me a shell, which is logical.
I would have a solution which is not to merge the bosses between the other bodies in order to be able to make the shells of the largest parts. But the problem with doing this is that I don't know if 3D printing won't be a problem. So when importing in .stl the file will be understood as a block or in several?
I'm a draftsman in boilermaking so I know solidworks very well but I'm also working on a kickstarter project so I use functions that I don't usually use, especially the shells so there may be a solution with this function but I haven't been found.
Ps: I can't provide you with a file since it's a project and I don't have permission to disclose it.
The difficulty of the hull function is that it creates interior surfaces by shifting the thickness. If you have a convex radius of R-value on the outside, you'll have an R-thickness radius on the inside. So if you have a thickness greater than the mini convex radius, you get a negative radius. If the model does not have too many complex left-hand surfaces, this negative radius will be replaced by a zero radius. If, on the other hand, the model is mathematically complex, Solidworks will not make an effort and there will be an error. The minimum radius can be very small in case of sharp angles (zero radius). It is therefore necessary to look for the minimum radii (function check model/mini radius).
You have the solution of doing all your radii and convex fillets (or just some) after the hull function. The model is then lighter and the shell fits well. However, it requires a little time and the use of the "select other" tool by right click to select the interior surfaces.
You can also increase the convex radii so that they are at least equal to the thickness.
But your idea of not merging the bosses does not seem wise to me because you will have walls between bodies and many fillets/chamfers will no longer exist or will be made upside down.
In addition, when exporting, the STL asks you, in the case of multibody parts (unbound bosses in your case), the bodies you wish to process.
Thank you for your answers, I've never used the fingerprint function so I don't see how I could do with it.
I had indeed thought of Pierre's solution but I wanted to be sure that there are no other solutions.
So I'm in the process of reviewing my part and to avoid having to modify everything again, I modify the sketches so that they are all linked to planes at the beginning of the tree like this, I could move everything the functions as I see fit and thus insert the shell function where it is needed.
And thanks also for the info for the STL export in multibody.
The imprint function is very useful in the context of an assembly. It literally makes it possible to create the imprint of one room in another. Typically, you want to make the mold to make a part. You create an assembly from your part and then position a block with the assembly constraints so that the part is halfway in. After editing the block in the context of the assembly, you then use the imprint function by selecting the part. The block becomes the mold that can make the part. This amounts to a volume subtraction except that it does not remove the "tool" volume (here the part). Without using the notion of assembly, it is found with the intersection function that appeared in the 2014 version I think.
In your case, this would involve making a part corresponding to the inner volume of the shell beforehand and positioning it correctly before using either the imprint function or the intersection function.
By far, in your place, I would prefer to rework the spokes and their position in the shaft and then do a hull function.