ABS warps. You know this going in, and you still get caught by it, because the first layer goes down beautifully and you allow yourself a little hope. Then the corners curl up off the bed around the third centimetre of height, the part detaches with a noise like a small disappointment, and you are reminded why everyone tells you to just print in PETG.
The fix is not exotic. ABS wants a warm, draught-free environment so the lower layers do not cool and contract faster than the ones above them. That differential is the whole problem. Heat the bed all you like, the air above it is still ambient, and ambient in a garage in late December is not kind. What it wants is a chamber.
I did not want to spend money, so the first version was a cardboard box with a hole cut for the filament and a sheet of acrylic leaned against the front. It looked like something a child built and it worked far better than it had any right to. The internal temperature crept up to the high thirties just from the bed and hotend, the warping dropped off sharply, and a part that had failed three times came off the plate flat.
The temptation, of course, was to engineer it properly. I sketched out a sealed enclosure with a thermistor, a small heater, and a controller to hold the chamber at a set point, because if a cardboard box gets me to forty degrees then surely a real chamber held at fifty solves it entirely. I got as far as ordering parts before I stopped and looked at what I actually had.
What I had was working prints. The marginal gain from active heating, for the parts I print, did not justify a heater sitting inside a box full of plastic that softens when warm and electronics that prefer not to cook. There is a real fire-safety conversation to be had before you put a heating element in a sealed enclosure unattended, and "it was technically interesting" is a poor thing to tell the insurer.
So the eventually in the title is doing some work. The eventual setup is embarrassingly simple: an IKEA Lack table frame, acrylic panels held with magnets, a cheap thermometer poked through the side, and the printer's own waste heat doing the job. No active heating, no controller, no fire risk worth losing sleep over. It sits in the low forties on a long ABS print and that has been enough.
I will probably build the heated version one day, properly, with a thermal cutoff and a clear head about what can go wrong. But the lesson from this round is the usual one: solve the problem you have, not the project you fancied. The box was ugly. The parts came out flat. That was the brief.