Join these two surfaces and one half vanishes!

Why does one half vanish with Join, but join ok with JoinEdge ?

ok the shape in the middle is not 100% smooth but its a long story ! and it doesnt need to be.

Its part of a larger object, I have even had it developing a gap of a few mm along the join during being joined with many other surfaces into one final surface.

SteveJoin and one half vanishes.3dm (240.6 KB)

This is JoinEdge coming back to bite you.

If you select each of your surfaces and type “what” you’ll see that one is invalid, because you used JoinEdge. Notice that the isocurves of one surface do not go all the way to the seam? That’s because JoinEdge was used to “fake” the distance between the two.

You need to untrim both surfaces, then use MatchSrf to make the edges of the two surfaces meet, then Join. As I said in the other post, you should really never use JoinEdge.


Level II, Page 51 (page 59 of the pdf)


Hi Skyg,
I will go on strike and read that section ! I must be allowed training.

After creating a patchwork quilt of sweep2 surfaces, should I just select them all and use Join then showedge and naked edges and tackle those with MatchSrf or just use MatchSrf from the outset working my way across them ?

For some reason I cant get MatchSrf to work at most times so reach for the next thing I think of and that is JoinEdge. The item I have just created after weeks of work/research is a patchwork quilt of many surfaces all sweep2. Given the complexity of the object it was THE ONLY WAY. I havent then trimmed any edges. MatchSrf should work on those ? p59 onwards and revisit the project !

As such I am puzzled that as the surfaces share rails, joinEdge declares 0.05 on one and 1.2 on another. I then think I can be sure of joints where it asks permission to just exceed tolerance by a fraction, for our job, no issue at all.

I will revisit the original and try again for MatchSrf but if MatchSrf wont work and I have naked edges I think back to a post by Pascal where naked edges then JoinEdge was used.

Faking distance is not good, something that takes my surface edges and extends them is good…MatchSrf.

Would I be right in saying that JoinEdge will cause further nasties when I come to offset 2mm the cowling inwards.

For the end result after offset2mm in ghost view I had a few very small dark triangles and odd shapes that I couldnt edit out. It also has caused odd junctions and other imperfections to my original complex patchwork quilt of surfaces. Using MatchSrf on just this shape that I have rebuilt with CurvatureGraph and NetworkSrf sees one part of it join and the other half go twisted on me.

see pics and file. The result after offsetSrf, note the oddment ‘E’ and distort to end of what was an elipse. Careful work ends up like this after offset. was that due to JoinEdge before offset ? I will have to experiment. The sweeps were ’ to point’ in this area.

I select the one side of this easter egg shape ‘A’, then select first segment ‘B’, then next ‘C’ and squiffo before I get to D, various settings make no difference. JoinEdge simply works so is easy to use unfortunately. 30 mins struggling on one shape, hate to think how long the entire cowling would take.

After OffsetSrf on my complex object I also run ShowEdge naked edge and see a few needing sealing (two points with line between- magenta in my case), but also a few single magenta squares which despite megazoom still show as such. All looked fine until offset.

I had a planar elipse joinedged to the surrounding shape (looked like a lake in a quarry and offsetSrf totally messed up on that part of the model. Its why I used this shape to replace it before offsetSrf redone. it was a perfect shape from elipses, doesnt get more simple, then after offset I get this shape acting up ! Simple shapes messed up with JoinEdge.

model attached.
MatchSrf_struggle.3dm (697.1 KB)


I’ve gotta get back to work so I’m not going to be able to answer all your questions, but see the attached screen shot. You’re never going to get surface A to meet up with B and C since the junction between B and C is not continuous. There is no command that can fix this, you have to go back and make the surrounding surfaces continuous to each other. Also, I can see you used, JoinEdge on C and the surface above it, which is just going to make things worse.

The surfaces around your egg shape “A” need to all be at a minimum of tangent continuous (using MatchSrf) in order for A to be able to be matched to them.

Yes. If I can give you one piece of advice it’s this - Forget that JoinEdge exists. Ditto for MergeSrf. Somewhere along the way, someone gave you the idea that those were appropriate commands for what you’re trying to do. They’re not. MatchSrf, Join, ShowEdges. Lather, rinse, repeat.

skyg gave you some very good advises but I’ll add that if your starting geometry, be it curves or surfaces, are good, you won’t need MatchSrf most of the time.

If Sweep2 is not giving you perfect results, either your geometry is flawed or your tolerances are not set right. And, as he said, JoinEdge is never the solution.

Another advice; you have a tendency to put very long posts here, it doesn’t invite other users to dig in all this dense text. You would get better response by restraining yourself to shorter and better focused posts.


Conciseness not my strongpoint :slight_smile:
Devil is in the detail though and one gets to see just what I did which helps sort things out, so I hope, and I do try and keep it minimal.

I find that minimal responses keep me half guessing whats what sometimes, but will endeavour to do so my end as it may well frighten folk off.

The A B C stuff is what Offset Srf did to my nicely crafted original. It hurts to see it make repairs necessary. That there is not my handiwork, originals were well crafted…but I used JoinEdge.

JoinEdge to be ditched, though I failed to find two naked edges with MatchSrf, not without creating further mess.

I did a micro level sweep2 to sort that A B sliver out, then join. It was 0.2mm at its widest tapering to a point but I was at crazy zoom to fix it. One felt that at that thinness sweep2 would fail. and it did on my first attempt. I am a very precise worker, not hamfisted.

Naked edges are not large easily edited affairs, they are down at some crazy zoom level, where things can start to go peculiar and one cannot trust so well the display. Makes for a nightmare trying to fix things. JoinEdge is two clicks and done, One cannot see quite what is going on with the results of offsetsrf as it creates curves I never created.


Hi Steve,

There’s always exceptions but if the surfaces are created from correct geometry
-JoinEdge shouldn’t be necessary
-MatchSrf shouldn’t be necessary
-0.2mm sweeps shouldn’t be necessary

If you create one out-of-form surface, you should try to find ways to build it better. Repairs will always bite you back further down the line, as with your OffsetSrf problems.

There are modeling operations that are bound to create problems when applied to complex though correctly constructed geometry.
Offsetsrf would be an example.