I needed a bracket. A specific, slightly awkward bracket to hold a small fan against the side of a network switch in the cupboard under the stairs, at an angle, without drilling into anything. I spent twenty minutes looking for one to buy and found nothing that fit, which is the usual story. The thing that exists is always for the wrong size, the wrong angle, or sold in a pack of fifty from a warehouse three weeks away. So I made it instead, and the making took less time than the shopping had.
This is the quiet superpower of having a printer on the desk. The economics of small custom parts have completely inverted. It is no longer "can I find one", it is "can I be bothered to model one", and increasingly the answer is yes because the modelling is genuinely quick once the part is simple.
The part is dull, which is the point
This bracket is not clever. It is an L shape with two screw holes on one face and a flat pad on the other, plus a 15 degree tilt so the fan points where I want it. In FreeCAD that is a sketch, a pad, another sketch, another pad, and two holes. Five minutes of actual work and another ten of me second-guessing the wall thickness.
The discipline I have learned the hard way is to measure twice and design once. I had the fan in my hand and a pair of digital callipers on the bench, so the screw spacing and the body width went straight into the sketch as driven dimensions rather than guesses. The single biggest source of failed first prints, for me, has always been a number I eyeballed instead of measured. Callipers cost almost nothing and save you the slow misery of a print that is two millimetres too narrow.
A couple of things I now do by reflex on a bracket like this:
- Add a chamfer to any edge that meets a flat surface, so a stray blob of first-layer squish does not stop it sitting flush.
- Make screw holes a touch oversized, because filament swells and a hole modelled at exactly 3mm will not take an M3 screw.
- Print a tiny test tab of just the mounting face first if the fit matters, rather than committing to the whole part.
Worth it for a single bracket?
That is the honest question, and the answer is yes, but not for the reason you would think. The part itself probably saved me a few pounds and a wait. The real win is that the bracket fits the actual situation: the right angle, the right offset, holes exactly where the switch lets me put them. An off-the-shelf part would have been a compromise I then bodged around with cable ties. This one just bolted on and looked, frankly, deliberate.
The print took forty minutes in PLA at a layer height I did not even bother to optimise, because nobody is ever going to see it and it holds a 40mm fan, not a bridge. It came off the bed, the screws went in, the fan sits at its 15 degrees, and the switch in the cupboard runs a few degrees cooler. Total elapsed time from "I need a bracket" to "the bracket is installed" was under two hours, most of which was the printer doing the work while I did something else.
There is a version of this hobby that is all about the perfect benchy and the dialled-in retraction settings. I enjoy that too. But the part of it that has actually changed how I live is the boring stuff: the brackets, the spacers, the little clips that hold a cable out of the way. Things I would once have done without, or bodged, or bought in the wrong size and resented. Now I just make them, and the workshop is slowly filling up with small, ugly, perfectly-fitted solutions to problems nobody else has.