Make Anything's Live CAD Session Shows Why Some Prints Just Suck

The Problem Starts Before the Printer

Every failed print gets blamed on the machine. Stringing, layer shifts, sagging bridges, and weak walls all look like printer problems. Devon from Make Anything ran a live CAD session this week that made the opposite point: most prints suck because the model was bad before the first layer went down. Working in Fusion 360 with the Fusion and Fika crew, he walked through designs that were almost guaranteed to print poorly and rebuilt them so they would actually work.

The session is worth watching for anyone who designs their own parts. It is one thing to know a printer can handle a 45 degree overhang. It is another thing to design a part that respects that limit from the start. Devon's live fixes covered the usual suspects: walls that were too thin, features that printed in the wrong orientation, and geometry that looked fine on screen but fell apart once it became plastic.

Wall Thickness Is Not a Suggestion

One of the most common mistakes in beginner CAD models is walls that are too thin. A 0.4 mm nozzle can print a single perimeter, but a single perimeter is not a structure. It will flex, crack, and look terrible. Devon showed how bumping walls to two or three perimeters, or adding a small rib, turns a flimsy shell into a usable part. The printer did not change. The CAD file did.

The same logic applies to holes, bosses, and mounting points. A screw boss needs enough plastic around it to hold torque. A hole printed horizontally will come out oval unless it is sized and oriented correctly. These are not advanced tricks. They are basic checks that many people skip because the slicer preview looks acceptable.

Orientation Changes Everything

Devon also stressed orientation. A part that looks perfect in one orientation can be a nightmare in another. Overhangs become unsupported loops. Layer lines run the wrong way and leave a weak axis. Curved surfaces step visibly because the layers cut across them at a bad angle. Rotating the model by 90 degrees often fixes all of this at once.

Orientation also affects strength. Prints are strongest along the layer lines, not across them. If a bracket needs to hold load, the load should run parallel to the layers. Designers who ignore this end up with parts that break along the weakest plane. The fix is sometimes as simple as laying the part on its side instead of printing it upright.

Mesh Cleanup Is Half the Battle

Imported meshes from Thingiverse or scan data are another source of bad prints. Non-manifold edges, flipped normals, and tiny internal gaps confuse slicers. Devon demonstrated how cleaning up a mesh in Fusion 360 before exporting can prevent weird slicing artifacts. The printer is not confused. It is just doing what the broken file told it to do.

Live sessions like this one are useful because they show the actual process, not just the finished part. You can see the thought process behind each change. That is the part most tutorials skip. They show a perfect model and say print it. They do not show the five bad versions that came first.

Prints That Suck Are Usually Fixable

The lesson from Devon's stream is that bad prints are usually bad designs in disguise. Fix the CAD and the print follows. It is tempting to blame the printer, the filament, or the slicer settings. Sometimes those are the issue. More often, the model needs a thicker wall, a better orientation, or a few support-friendly angles.

If your prints keep coming out ugly, load the model back into Fusion 360 and ask what the printer would have to do to build it. If the answer involves floating plastic, paper-thin walls, or stress along the layers, change the CAD. The printer will thank you, and your bench will have fewer half-failed parts on it.