im trying to remove that certain edge but it destroys the mesh… i tried deleting the edge and it deleted the whole face.
merging the 2 faces above and below the edge fixed it… but im curious why it did that though for future refernce
The first cropshot shows face indicator dots on the faces, which go to the average position of the vertices a face is constructed from. At the top there are face indicators on the edges, which means you have double geometry in form of 0 area faces.
The faces around the circle are all n-gons, faces with more than 4 sides, and they’re concave. That means you didn’t add the circle, you either connected the geometry or cut it out and while doing that, it required to add an edge to hold the shape of the face without automatic triangulation destroying it. The face indicators for a convex quadrilateral face is right in the middle of the four corner vertices. The n-gon faces have theirs close to the circle, averaged closer to the vertices that came from connecting or cutting the circle shape ot or from it.
The n-gon face in the middle of the circle doesn’t have a face indicator because it’s a severely non-planar polygon, and the indicator is on the inside of the cylinder. The indicators for the top and right side n-gon faces are under the ridge of the circle.
It wasn’t an extra edge you had and it wasn’t destroying your mesh. The way you’re modeling is already causing problems in form of concave, non-planar, and 0 area polygons. Also you’re connecting two mesh pieces with very different density which will cause problems if you ever want to change the cylinder.
How to model it properly? Don’t know without knowing what forms you’re modeling, what pipeline stages it will go through, and what the target use will be (still, animation, game engine, 3d printing, other). The little that is known or visible in the cropshots, correct way at that stage would be to not connect the circle, and remove doubles in the cyclinder.
what my final goal is something basic but challenging for a complete beginner…
a cylinder with a circular top and has an extruded circle and that extrude has a bit of thickness…
the approach i used that caused many problems was
-cylinder (top face deleted) + half uv sphere
-aplied boolean modifier and chose the operation value as : union(that was the easiest way i can think of but there are many more better ways of doing it, but currently unknown to me.)
next was the extruded circle near the left side of the cylinder, this is where i messed up big time.
i made another circle mesh and i used the knife project to mark the cylinder using that circle’s shape
deleted that 2nd circle mesh…
pressed F to get 1 faced circle and pressed I to inset to creat a 2nd circle and extruded it…
I would recommend trying to have a more uniform density of geometry. You can use booleans and a denser cylinder to cut a hole in the main cylinder with edge loops prepared at the right places to get the edge of it and the you can use Limited Dissolve(x -> l) with All Boundaries option checked (F6 just after the operation) to get rid of the unwanted vertices and extrude move and scale from there to form the ring.
You’ll save everyone’s time by not cropping screenshots and by preparing and uploading an example .blend for every question involving one
Describing a problem is rather easy, usually fits in the thread title, and with a screenshot to back it up it’s rather clear. But it’s not how you describe the problem that gives you answers, it’s how much information of what you’re doing and working with that needs to get across to get to the cause of the problem. What are those? With a modeling problem, the same information you see and use to model.
People with more experience can tell a lot from just one screenshot and can reverse engineer a model structure. That means two things: you don’t have to annotate everything in screenshots and only highlight non-obvious things in them, which saves you time, and since you don’t know to visualize and explain the cause of the problem you want to convey as much information as you can instead of hiding it. The most efficient way of doing that is using full interface screenshots and uploading an example .blend. The .blend also helps people to reply.
I would advice against using boolean operations for modeling, for now. Boolean operations would be one of the basic tools in CAD applications, especially ones using Solid Modeling paradigm. But this is not CAD, we’re using polygonal modeling where the structure is everything, and you need to learn how to read and prepare those structures for the forms you want to model.
Boolean operations have their use, but you’ll notice that those need a lot of clean up when you want clean structures. Which will be hard if you don’t know what clean means.
Quadrilateral faces have a direction. With those you have structure flows in the model, and that’s how you can read the structure
start by subdividing and rounding a cube so it gets an edge loop in the middle (equator) with enough vertices to form a cylinder, and a grid pattern for the cutout (W -> subdivide smooth)
deleted the lower half, extruded the cylinder, filled the bottom with an n-gon
added loop cuts with odd number of cuts (ctrl+R)
selected middle vertex, grew the selection (ctrl+numpad_+), inset (i)
Blender comes with Looptools addon, which you can activate in user preferences. I used its circle function to get the vertices ordered on a uniform circle
Grid fill to get a clean grid in the middle (ctrl+F -> grid fill)
Inset, extrude, scale to 0 along axis to flatten the form
Edge slide (gg = g twice) to average the face sizes around the circle
I also added edge loops with bevel modifier to support the subdivisions on hard edges (green ones)
It’s one example of modeling what you’re after. There are many other ways, but the structure at the end will be similar or same when it has to support the same requirements, subdivision surfaces in this case.