I want to create a mesh like this:
Cylinder with a circular Dent
using blender version 2.83.5 It should be easy to create it with a boolean difference between the cylinder and the torus, but it seems that in the defaukt position something goes wrong. I am new to blender forums and I never reported a bug so I want to avoid to make useless noise by first asking someone else if I am doing something wrong. I followed this sequence:

  • Open a new shene and delete the default cube
  • Create a cylinder and leave everything default
  • Create a torus and leave everything default
  • Add a booleand difference to the cylinder with the torus as a target
  • Nothing happens
  • Apply the boolean and hide the torus. The cylinder is still there intact
  • Go into edit mode and you'll see the traces of the difference as new edges and verteces where the torus was
  • I also noticed that if the torus is not perfeclty centered (i.e. moved x by 0.001) before applying the boolean, everything works as expected.

Is this a bug or I am doing something stupid? Thank you for any suggestion


The vertices of the torus and the edges of the cylinders are overlapping, overlappings prevent the boolean to work correctly, I think it's a problem that will be soon fixed, but for the moment you need for example to scale a little bit the torus and the boolean will work fine.


Not really a direct answer, but in this case it may be easier not to use a Boolean? Create a torus...

  1. Select desired loops,V rip those edges,
  2. Hover - L select unwanted faces and delete
  3. Select boundary edge loops, E extrude, right-click to drop in place
  4. SZ scale the extrusions up and down
  5. F fill the top and bottom faces

enter image description here

Assign a Bevel modifier by angle. AltN, if necessary, to recalculate normals to face outwards.

  • 1
    $\begingroup$ Thanks for the answer. Actually the mesh was a little more complicated than that, and I succesfully made it anyway using the boolean with a slightly (0.001) decentered torus, and moving back to center the interested verteces (those that were belonging to the torus before applying the boolean. The point was, not to find a way to make it, but if is was a bug preventing it to work as expected. Thanks for the suggestion anyway, to demonstrate once more that there's no single way to (skin a cat) get the result ;) $\endgroup$ – ziomau10 Aug 30 '20 at 19:31
  • $\begingroup$ @ziomau10 That's cool.. sometimes Booleans are the best way. I always look for the modelling approach first, and only reach for the Boolean when it does the job better, or well-enough in a much better time. $\endgroup$ – Robin Betts Aug 30 '20 at 19:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.