ABS is a lovely material right up until the corners lift off the bed and the whole print peels into a banana. I had been losing the bottom layers of tall parts for weeks. Hairspray helped, glue stick helped, brims helped, and none of it solved the actual problem, which is that ABS contracts as it cools and if different parts of the print cool at different rates, it tears itself apart. Warping is not an adhesion problem. It is a temperature gradient problem wearing an adhesion problem's clothes.
The real fix is to not let the part cool unevenly in the first place. Keep the whole thing warm, in still air, until it is done. That means an enclosure.
the cheap route first
I did not want to buy a commercial enclosure, partly on cost and partly on stubbornness. The well-trodden path in this hobby is an IKEA Lack table, or in my case a glass-fronted cabinet I already had, with the printer inside and the front mostly closed. No heater. The idea is that the heated bed at 100°C plus the hotend dumps enough waste heat into a small sealed volume to raise the ambient temperature on its own. Passive, free, no electronics.
It half worked. With the cabinet shut and the bed at 100, the internal air settled around 35 to 40°C after half an hour of printing. That was enough to stop the small parts warping. The tall ones still curled, just less. The problem was that the temperature was not stable: it climbed slowly during a print and dropped the moment the bed cycled, and ABS hates a moving target almost as much as a cold one.
measure before you heat
Before adding any active heating I put a cheap thermometer probe inside and logged the air temperature for a few prints. This is the step I would skip if I were being impatient, and it is the step that actually mattered. The log told me two things: the passive enclosure plateaued around 40°C, well short of the 50 to 60 you really want for big ABS, and most of the heat was escaping through the gap where the filament fed in and around the door seal.
Sealing the obvious leaks with some draught excluder bought me another five degrees for nothing. That is the lesson, really: an enclosure is an insulation problem first and a heating problem second. There is no point adding watts to a leaky box.
the eventually
The "eventually" in the title is because I have not actually fitted the active heater yet. I bought a small PTC heating element and a fan, the standard self-regulating type that cannot run away and set the cabinet on fire, and the plan is a simple thermostat holding the chamber at 50°C. But the sealed-and-insulated passive box, with the leaks taped up, is already getting me clean ABS prints on everything except the very tallest parts. So the heater has moved down the list, because the cheapest improvement, taping up the gaps, did most of the work.
If you are fighting warping: enclose it, seal it, and measure the air temperature before you spend money on heating it. You may find, as I did, that the box and the tape are nine tenths of the fix and the clever bit can wait for a rainy weekend.